Temperaments Between Men and Women Research Paper

Abstract

The focus of this research will be the examination of the effects of male and female temperaments on the extents of frustration both sexes experience while dealing with certain long-lasting activities like the waiting process at a busy intersection. The scholarly literature on the topic of temperament development and gender inequalities of modern society will be examined. The method of direct observation will be used in the research itself, as the researcher will place him/herself in a group of people at a busy intersection to observe and record the ways in which male and female frustration is manifested. The results of the research will be presented and discussed providing the implications for further study in the area of temperament differences.

Introduction

The study of differences in the male and female temperaments is rather significant as far as the modern society experiences the lack of understanding the inter-gender differences and displays such signs of gender discrimination as lower educational and employment opportunities for women, prejudiced social opinion against them, etc (Boyce & Herd, 2003). Accordingly, the current study will contribute to the inter-gender understanding and will provide for the further development of gender equality in the human society. Moreover, if properly carried out, this study will provide for the better understanding of ordinary human beings in their daily communication and cooperation.

Scholars take their time to examine this topic. In more detail, such authors as Bing (1999) inquire about the purely physiological and biological aspects of the socially adopted gender inequality. According to Bing (1999), the popular public believe that male brains are larger than female ones has led to the misperception of women as the ones to be excluded from doing such traditionally male jobs as politics, social work, etc (Bing, 1999). Further on, Walker, Berthelsen, & Irving (2001) study the topic of temperament formation in young children and influence of their temperaments on their peer acceptance. Gender perspective is examined by these authors who conclude that females are less active in the socialization, while males are more patient and persistent (Walker, Berthelsen, & Irving, 2001). Kasen, Cohen, Brook, & Hartmark (1996) argue about the huge role played by family environment and psychological traumas received by children upon their temperament formation and further socialization (Kasen, Cohen, Brook, & Hartmark, 1996). Boyce & Herd (2003), Mcglashan, Wright & Mccormick (1995), and Whisenant (2003) consider the social aspects of the study topic examining the ways in which women are less fortunate in employment, education, and even military service because of the public stereotyping of their temperament peculiarities and biological differences (Boyce & Herd, 2003; Mcglashan, Wright & Mccormick, 1995; Whisenant, 2003). However, the topic of frustration males and females experience during waiting has not yet been studied substantially enough. The purpose of this study will be to fill in this gap and provide an account of frustration rates in men and women busy with waiting. The hypothesis of the study is as follows:

Women experience less frustration while waiting due to the peculiarities of their psychology which allows them to be less focused on waiting and reduce the stress by this. On the other hand, men are more frustrated as their focus on current business makes them nervous about the circumstances preventing them from the action.

Method

This study will involve two major methods of research. The first method used for the conduct of this study will be standing on a busy intersection and observing the reactions of men and women to the necessity to wait for a certain amount of time. Accordingly, it is rather difficult to exactly identify the number and any other characteristics of the study participants. What can be stated more or less exactly is that the sample of the study will encompass males and females of the adult age groups that can be observed on the roads during their working days. Drawing from the nature of the study, it will be impossible to examine the participants individually (Mitchell and Jolley, 2006). However, the establishment of groups for their examination will also prove impossible in real-life situations. As a result, the study will be carried out with the help of the observation of the spontaneously formed group of people. The design of the study will also be determined by its real-life nature, as it will be impossible to successfully use any laboratory research during the observation of a group of people who, in addition, will not know that they are observed. As for the special apparatus for the study (Mitchell and Jolley, 2006), it will consist of only recording, either in the written form or with the help of audiotape, the reactions of male and female study participants to any challenging situations on the busy intersection chosen as the site of the study.

On the whole, the procedure anticipated for use in the study will begin with choosing the intersection with heavy people and car traffic. An important aspect in this choice will be the result of the prior observation according to which the intersection with the most balanced average rates of male and female participants will be chosen. The next step will be finding a proper position for the researcher (Mitchell and Jolley, 2006), who can either be in a car among the cars of the study participants, or a pavement from which the most relevant data can be obtained and timely recorded. Then, the very procedure of the study will be carried out during the pick hours of the early morning when the majority of people are heading for work or study. Finally, the results of the research will be analyzed and synthesized in order to provide an argumentative discussion and implications for the further study of this area (Mitchell and Jolley, 2006).

Finally, the second method used will be the integrated application of the Likert Rating Scales together with the consideration of a proposed study scenario as rated by the two female and two male student groups (30 persons each). The first step in applying this method will be recruiting 120 volunteers (60 males and 60 females to ensure the objective and equal representation of both genders) from the local universities or schools through internet and newspaper advertising. The next step in the research procedure will be presenting the study scenario to the volunteer assessors in order for them to grade it and express their own ideas, if any, on the information they can obtain from the study and the results of its assessment. The point here will be to offer one male and one female group to assess the male character of the study scenario, while the resting groups will be asked to assess the performance and frustration levels of the female character. After the groups are assigned their tasks, it will be necessary to explain to the volunteers the principle according to which Likert Rating Scales work.

Finally, the synthesis of the assessment results and making respective conclusions will be the closing stages of the research procedure. These stages will be necessary to see the differences, if any, in the assessment given by the male volunteers to the female study character, and vice versa. To make the research as complete as possible the assessments of the male volunteers of the actions and frustration level of the male character will also be considered, as well as the same factors relating to the female assessment of the female frustration level.

Results

As for the study results, they will be assessed according to the nature of participants’ responses and according to the approximate degree of indignation or calmness as observed in the recorded data. The lexical analysis of the responses will also be of great help as it will provide the study with the implications of the statements made by the participants as viewed from the nature of lexical units and intonations used. In accordance with these assessment criteria, the results will provide the study with certain information necessary to either prove or reject the hypothesis formulated and outline the directions for further study (Mitchell and Jolley, 2006).

Further on, the results of the Likert Rating Scales assessments will be considered rather specifically to ensure as small a sampling error as possible. First of all, the Likert Rating Scales are divided into two parts, one being the reflection of frustration rates and another being the display of the responsibility levels the study characters of two genders manifested while being in a frustrating situation. Both scales offer a range from 0 to 5. In the first scale, the smallest figure is equivalent to the absence of frustration, while the highest result evidences the extreme frustration level. The second scale uses the same range, with 0 showing no responsibility of the subject, while 5 assesses him or her as “very responsible”.

Thus, using the results of these assessment scales, it will be possible to calculate the average figure reflecting the male assessment of the female frustration and responsibility, as well as the female assessment of the male ones. These average figures will allow the researcher to have at least an approximate picture of the frustration levels typical of both genders, as well as of attitudes men and women have towards the psychological abilities of each other.

Discussion

Accordingly, if the study results are as predicted, it will be possible to confirm the research hypothesis and develop further research on this topic to the study of reasons for such temperament differences and test the same research design on different settings and using different variables besides frustration rates. However, the study design chosen is limited in its scope and in the credibility of its results as the certain number of participants examined will not allow making too generalized conclusions. In other words, what will turn to be true for this group of people will not necessarily be true for another or a larger one. That is why, further research will be necessary to provide adequate data eligible for generalization (Mitchell and Jolley, 2006).

References

Bing, J. (1999). Brain Sex: How the Media Report and Distort Brain Research [*]. Women and Language, 22(2), 4.

Boyce, L. A., & Herd, A. M. (2003). The Relationship between Gender Role Stereotypes and Requisite Military Leadership Characteristics. Sex Roles: A Journal of Research, 49(7-8), 365+.

Kasen, S., Cohen, P., Brook, J. S., & Hartmark, C. (1996). A Multiple-Risk Interaction Model: Effects of Temperament and Divorce on Psychiatric Disorders in Children. Journal of Abnormal Child Psychology, 24(2), 121+.

Mcglashan, K. E., Wright, P. M., & Mccormick, B. (1995). Preferential Selection and Stereotypes: Effects on Evaluation of Female Leader Performance, Subordinate Goal Commitment, and Task Performance. Sex Roles: A Journal of Research, 33(9-10), 669+.

Mitchell, M. L. and Jolley, J. M. (2006). Research and Design Explained (6th ed.), Wadsworth Publishing.

Walker, S., Berthelsen, D., & Irving, K. (2001). Temperament and Peer Acceptance in Early Childhood: Sex and Social Status Differences. Child Study Journal, 31(3), 177+.

Whisenant, W. A. (2003). How Women Have Fared as Interscholastic Athletic Administrators since the Passage of Title IX. Sex Roles: A Journal of Research, 49(3-4), 179+.

Preferential selection and stereotypes: effects on evaluation offemale leader performance, subordinate goal commitment, and task performance

Preferential treatment of minority group members continues to be a controversial yet sometimes necessary ingredient in equal opportunity employment. Managers recognize the increasing diversity of the workforce and the need to fully utilize its members. Furthermore, fulfillment of equal opportunity employment requirements and use of voluntary affirmative action plans have become essential, if not obligatory, elements of everyday human resource procedures. The growing number of women represented in the workplace, and in managerial positions in particular, have made it essential to understand possible hindrances to women’s organizational effectiveness in response to preferential employment practices. Such practices have been shown to have unintended negative effects, both on individuals’ perceptions (of self and others) and behaviors. These perceptions and behaviors can often translate into negative consequences for the organization, as will be explored in this paper.

The purpose of this study is to examine whether preferential treatment has effects on subordinate attitudes and behaviors in a task setting, e.g. in evaluations of their leaders, commitment toward leader-assigned goals, and in subordinate task performance. In addition, the effects of stereotypical views toward women as managers in leader evaluations, subordinate commitment and subordinate performance are explored.

Preferential Treatment

Previous preferential treatment studies have concentrated on either individual self-perceptions or perceptions of others who have been treated in a preferential manner. In many empirical studies, preferential selection has been shown to negatively affect individuals in self-reports of organizational commitment, role stress, satisfaction, self-perceptions and evaluations, and task selection (Chacko, 1982; Heilman and Herlihy, 1984; Heilman, Lucas, and Kaplow, 1990; Heilman, Rivero, and Brett, 1991; and Heilman, Simon, and Repper, 1987). Females selected preferentially have evidenced more negative self-views than in merit-based situations. Preferential treatment has resulted in women’s lower self-evaluations of performance, leadership ability, and future desire to lead (Heilman et al., 1987).

Individuals may also differ in their perceptions of others who have been treated preferentially and of the organization that acted preferentially. The amount of discrepancy in merit between candidates when a less qualified minority candidate is preferentially selected has been shown to affect the level of perceived injustice felt by the more qualified majority candidate (Singer and Singer, 1991). Perceived fairness of the selection process may also negatively influence beneficiaries of affirmative action in their attraction to certain organizations (Nacoste, 1987) and in reactions to individuals (e.g. job applicants) similar to themselves (Heilman, Kaplow, Amato, and Stathatos, 1993). These findings highlight the paradox of affirmative action programs – they may often harm the very individuals that they are designed to help.

Leader Evaluations

The evaluation of female leaders may also be detrimentally affected by their preferential appointment. Jacobson and Koch (1977) conducted a laboratory experiment involving the preferential, chance, or merit selection of a female confederate as leader in a dyad with a male subject. They found that the more equitable the method of leader selection, the more positively the subject rated the confederate’s performance and the more credit she received for success.

A leader’s source of authority, or legitimacy, may predict evaluation of his or her leadership effectiveness (Read, 1974; House and Baetz, 1979). A follower must not only perceive that a behavior is an influence attempt – he or she must also view the attempt as acceptable. If the influence attempt is not accepted by the perceiver, then the leader’s authority is not legitimized.

Goal Commitment and Subordinate Performance

One important component of leader legitimacy is real or perceived leader competence (Read, 1974). This component may prove perplexing in a preferential treatment situation. A preferentially selected leader may be viewed as less legitimate because competence is not perceived as a basis for selection. The leader’s task competence may truly be adequate, but this competence has not been validated in the selection process. Subordinates may perceive the leader as less qualified, and subordinate commitment to the leader may suffer.

Goal setting is most likely to improve task performance to the extent that subordinates accept leader assigned goals (Locke, Shaw, Saari, and Latham, 1981). Goal acceptance, however, is necessary but not sufficient for goal commitment. Goal commitment (i.e. the individual feeling bound to achieving the goal), is necessary for goal setting to be an effective motivational tool and is critical in predicting performance (Hollenbeck and Klein, 1987). We would expect goal commitment to be reduced in the case of preferential leader selection, as a leader’s legitimate authority is undermined. Lessened goal commitment by subordinates may translate into ensuing performance decrements. The extent to which subordinates perceive leaders to be legitimate will be positively related to the subordinates’ commitment to goals assigned by that leader. In turn, goals will most likely improve performance only to the extent that subordinates are committed to such goals.

From the preceding discussion on preferential treatment, we expect to find that preferential selection of a female into a leadership position will lessen subordinates’ feelings of commitment to both the female leader and to goals assigned by that leader, translating into a decrease in performance level. We have seen that preferential treatment negatively affects many individual perceptions, regardless of gender. We would therefore predict no significant difference in reaction based on gender of the subordinate. We thus expect to see a negative impact of preferential selection on subordinate evaluations of the female leaders, regardless of subordinate gender. Additionally, we expect to find that subordinate commitment to goals assigned by the female leader will be lessened in a preferential treatment situation. Last, we expect to observe lower task performance by subordinates because of lessened commitment to the leader-assigned goals. These proposed relationships are depicted in Fig. 1. Our formal hypotheses are as follows:

• Hypothesis one: Subjects will assign lower scores of leader effectiveness to a preferentially selected female leader than will subjects in a merit selection condition.
• Hypothesis two: Subjects will exhibit significantly lower levels of commitment to goals assigned by a preferentially selected female leader than will subjects in a merit selection condition.
• Hypothesis three: Subjects whose leaders have been selected preferentially will exhibit lower task performance than that of subjects in a merit selection condition.

Stereotypical Beliefs About Women as Managers

An additional complication in the affirmative action arena stems from gender-based views in the workplace and the effects of these views on evaluation of female employees. Most managerial jobs have been attributed masculine characteristics, i.e. they have been male “sex-typed” or gender-typed. Congruence of job and gender characteristics in gender-typed occupations stems from the differences in specific personality traits assigned to men and women and the judged appropriateness of those characteristics on the job. For example, assertiveness is viewed more as a male trait, while passivity is seen as a female trait (Locksley, Borgida, Brekke and Hepburn, 1980). These attribute differences have led to a widespread notion that men may possess more of those attributes viewed as necessary to succeed at work (Heilman, 1984).

It has been found in separate studies of male and female middle managers that successful managers are perceived to possess characteristics, attitudes and temperaments more commonly ascribed to men in general than to women in general (Heilman, Block, Martell and Simon, 1989; Schein, 1973, 1975). Men are perceived as more aggressive and independent. Women are viewed as more tactful, gentler and quieter. Congruence between gender-typed jobs and gender-typed tasks has been shown to affect performance ratings (Barnes-Farrell, L’Heureux-Barrett, and Conway, 1991; Eagly, Makhijani, and Klonsky, 1992; Stoppard and Kalin, 1983). In general, men have been found to receive more favorable evaluations than women (Deaux and Emswiller, 1974; Garland and Price, 1977; Nieva and Gutek, 1981). Deaux and Emswiller (1974) found that performance by a male on a “masculine” task was attributed to skill, while an equal performance by a female was attributed to luck. The reverse did not hold for a feminine-related task.

Garland and Price (1977) measured male undergraduates subjects’ attitudes toward women in management, using Peters’ et al.’s Women as Managers Scale (WAMS; Peters, Terborg and Taynor, 1974, as cited in Garland and Price, 1977). They found that the description of success for a female manager was attributed to external factors (i.e. luck) by those subjects with low WAMS scores (representing more traditional views of women). Likewise, those subjects with high WAMS scores (i.e. less traditional attitudes toward women managers) attributed success to the woman’s internal factors (ability and effort). The WAMS scale is essentially measuring an individual’s stereotypical views about women’s ability to be effective managers.

A second aim of this study was to explore the role of stereotypical beliefs on reactions to a female leader and the goals she assigns. Individuals with more traditional stereotypical beliefs about females typically score lower on the WAMS scale. We posit that higher levels of stereotypical beliefs regarding the female role in the workplace (i.e. lower WAMS scores) will negatively impact subordinates’ evaluation of female leaders and commitment toward goals assigned by these leaders, as well as negatively affecting subordinate performance. Proposed relationships are again depicted in Fig. 1. From the preceding discussion, our next hypotheses are:

• Hypothesis four: Subjects’ scores on the WAMS measure will be positively related to their evaluations of the leader’s effectiveness.
• Hypothesis five: Subjects’ scores on the WAMS measure will be positively related to commitment to leader-assigned goals.
• Hypothesis six: Subjects’ scores on the WAMS measure will be positively related to subordinates’ task performance.

In addition, it is possible that interactions exist between our preferential treatment variable and WAMS scores. A subject with more traditional stereotypical views of the female role, for instance, might feel more negatively toward the use of preferential treatment than would a subject with less traditional views. We would predict that lower WAMS scores interacting with preferential treatment would have an even more negative effect on goal commitment, leader evaluations, and performance. Therefore, we will also explore whether such interactions are significant.

Method

Subjects

Our sample consisted of 135 undergraduate students enrolled in a junior level management class during the Fall 1992 semester at Texas A&M University. The students participated in the laboratory experiment for extra credit points. Participants included 69 males and 66 females. Although race/ethnicity information was not collected, the class from which the sample was taken is typically homogeneous – approximately 98% white.

Task

Students worked on a class scheduling task in which they were to develop non-redundant schedules involving five courses offered at various times. They completed both a ten minute practice trial and a fifteen minute experimental trial. This type of task is both highly familiar and has high face validity to students (Earley and Kanfer, 1985) and is often used in goal-setting studies.

Variables

Manipulated Variable

Preferential Treatment. Preferential treatment was operationalized after administration of two tests purported to assess leadership ability. The experimental manipulation took place as the lab instructor collected the two tests for scoring. While the subjects completed an additional scale (a self-esteem measure), the lab instructor enacted one of two scenarios (determined at random by a coin toss prior to the laboratory session). In the merit selection condition, the instructor appeared to score the two measures and base leader selection on the highest test scores. The actual statement used by the instructor in this scenario was, “As stated before, the two tests you took were used to predict leadership effectiveness. Based upon the results of the tests, will be in the leadership position.”

In the preferential selection condition, the instructor did not attempt to score the tests. Instead, the instructor placed the tests aside and communicated that, “Although we wanted to place people in the leadership role based on ability we are experiencing a shortage of female leaders. Therefore, will be placed in the leadership role.” As noted, in both cases the female confederate was chosen as group leader.

The laboratory instructor was one of two graduate students, one male and one female, who conducted sessions on a relatively even basis. Four female undergraduate students alternated as the confederate. These students were familiarized with the lab scenario before the sessions took place, and they followed the same set script regardless of the session being preferential treatment or ability in nature. The operationalization of preferential treatment that was used is exactly like that used by Heilman and her colleagues (e.g. 1984, 1987, 1989, 1991, 1993) in which gender is used as a decision-making characteristic instead of ability.

Measured Variables

Goal: The leader-assigned goal was operationalized as twice the amount of schedules that each student completed in his or her ten-minute practice trial, to be completed in the fifteen minute experimental trial. The set goal level should have been realistic to students as it was specifically based on initial performance in the trial. This goal was chosen so as to be perceived as both specific and difficult to the subjects, factors which have been established as leading to higher motivation and subsequent performance (Locke et al., 1981.)

WAMS: The Women as Managers Scale devised by Peters, Terborg and Taynor, 1974 (Beere, 1990) is a twenty-one item scale that assigns scores on a numerical scale of one to seven, representing traditional to non-traditional stereotypes of women in business organizations held by the test subjects. The psychometric properties of WAMS have been demonstrated (Garland and Price, 1977). Coefficient alpha for the WAMS in our study was 0.92.

Leader Effectiveness: The scale used by the students to evaluate the female leader included five items with a seven-point Likert scale format. An example item is, “The leader was clear in providing instructions in how to perform the task.” Coefficient alpha for this scale in our study was 0.75.

Goal Commitment: The goal commitment scale used seven items from the scale developed by Hollenbeck, Williams and Klein (1989), measuring subjects’ commitment to the goal assigned by the female leader. The construct validity of this measure has been demonstrated (Hollenbeck, Klein, O’Leary, and Wright, 1989). Coefficient alpha in our study was 0.87.

Performance: Performance was operationalized as the number of schedules completed by the subjects in the fifteen minute experimental trial.

Procedure

Students took part in a laboratory experiment lasting approximately one and one half hours. Three to five students were involved in each session, along with a female confederate posing as a fellow student. Students had signed up for laboratory times on posted sign-up sheets, resulting in a varying mix of male and female subjects in each of the sessions. The session began with an introduction by the lab instructor. Bookkeeping duties such as registering attendance, obtaining subject consent, and obtaining demographic data were performed. Next, two tests were administered to the subjects and the confederate, purportedly to determine selection of a group leader to guide the group in a simple class scheduling task. The first test was the Wonderlic Personnel Test, which assesses general cognitive ability. The second test administered was the Leadership Opinion Questionnaire, which measures two facets of leadership (consideration and structure). Both of these tests have undergone extensive testing and are commercially available for research. Neither test was actually used to choose the leader as we utilized a confederate, although in the merit condition the lab instructor acted as though the tests were scored and used as the decision factor. In both the preferential selection and merit conditions, the confederate was then asked to lead the rest of the session.

The leader’s task was to guide the group through a class scheduling task. The task consisted of students being asked to write out sample schedules from a master schedule list, working independently. A ten minute trial scheduling task occurred, after which the leader checked on subjects’ progress and assigned individual goals for the actual fifteen minute class scheduling task. The goal was standard across laboratory sessions – twice the amount achieved on the ten minute trial run, to be accomplished in the fifteen minute actual run. Following goal assignment, the subjects were asked to complete measures of goal commitment.

Following the actual fifteen minute class scheduling task, the confederate was allowed to leave. Subjects were then asked to evaluate the effectiveness of the confederate in the leader role. The WAMS scale was also administered, followed by a manipulation check to observe whether the preferential selection condition was perceived as such by the subjects. Last, subjects were debriefed and were told that a full explanation of the study would be made available at its conclusion.

Results

Manipulation Check

Three items were included to check the preferential treatment manipulation (“The leader was chosen based on his/her performance on the leadership tests,” “The leader was chosen for non-job related reasons,” “The leader was chosen because of his/her sex”). The manipulation was checked by asking the subjects to rate these items from one (“strongly disagree”) to five (“strongly agree”). Correlations for the first and third items with the preferential treatment condition were highly significant (-.55, p [less than].01 and.54, p [less than].01) but insignificant for the second item (.15, n.s.). It is possible that the students were confused by the inclusion of the term “nonjob” when the task was only a laboratory simulation. We felt that the significance of the other two items lent sufficient evidence that the manipulation worked.

Tests of Hypotheses

The preferential/merit selection manipulation was dummy-coded “1” for preferential treatment and “0” for merit selection for purposes of the analysis. The leader-assigned goal level was coded as twice the number of trial schedules completed. The measure of goal commitment was coded as a value ranging from one to five. Performance was coded as the number of schedules completed by each student in the actual fifteen minute exercise. Each student’s evaluation of leader performance was coded using the one to seven scale, as was the level of stereotypical beliefs as measured by the WAMS scale. Correlations among the variables as well as means and standard deviations are shown in Table I.

Data were analyzed using hierarchical multiple regression.(1) Results of these regression analyses are summarized in Tables II, III and IV. Three different regression equations were completed. Equation one regressed leader effectiveness on WAMS scores in the first step, preferential treatment in the second, and a WAMS/preferential treatment interaction last. Equation two regressed goal commitment on leader-assigned goal level in the first step, WAMS scores in the second step, preferential treatment in the third, and a WAMS/preferential treatment interaction in the fourth step. Equation three regressed subordinate performance (number of schedules completed in the experimental trial) on leader-assigned goal level in the first step, goal commitment in the second, WAMS scores in the third step, preferential treatment in the fourth, and a WAMS/preferential treatment interaction last.

We did find significant two-tailed bivariate correlations between gender and the WAMS scale (-.51, p [less than].01), gender and goal commitment (-.20, p [less than].05), and gender and number of schedules completed (-.19, p [less than].05). We also found a significant one-tailed correlation between leader performance and gender (-.16, p [less than].05). The large association between gender and the WAMS scale appears to be driving these relationships, because WAMS scores are also significantly correlated with each of the above variables except for number of schedules completed. Controlling for WAMS, we ran partial correlations for gender, goal commitment, number of schedules, and leader effectiveness, and found that gender no longer had significant relationships with any of the other variables. We therefore felt comfortable with dropping gender from further analysis.

Hypothesis one stated that subjects would assign lower scores of leader effectiveness to a preferentially selected female leader than subjects in the merit selection condition. This was tested in step two of equation one, which showed that only 0.5% incremental variance in leader effectiveness scores was explained by the preferential treatment condition, which was non-significant. Thus, hypothesis one was rejected.
Hypothesis two stated that subjects would exhibit significantly lower levels of commitment to goals assigned by a preferentially selected female leader than subjects in the merit selection condition. This hypothesis was [TABULAR DATA FOR TABLE I OMITTED] tested in step three of equation two. The incremental variance explained was 0.04% (n.s.), thus rejecting hypothesis two.

Table II. Hierarchical Regression Analysis – Equation:

Hypothesis three stated that subjects whose leaders had been selected preferentially would exhibit lower task performance than subjects in the merit selection condition. This was tested in step four of equation three, with an incremental 1.4% of variance explained (p =.09). Thus, hypothesis three was marginally supported. Hypothesis four stated that subjects’ scores on the WAMS measure would be positively related to their evaluations of female leader effectiveness. This hypothesis was tested in step one of equation one, with variance explained equaling 6% (p [less than].01). Thus, hypothesis four was strongly supported.

Table III. Hierarchical Regression Analysis – Equation Two(a)

Table IV. Hierarchical Regression Analysis – Equation Three(a)

Hypothesis five stated that subjects’ scores on the WAMS measure would be positively related to commitment to leader-assigned goals. This was tested in step two of equation two. Incremental variance explained was 7.5% (p [less than].001), thus strongly supporting hypothesis five. Hypothesis six stated that subjects’ scores on the WAMS measure would be positively related to subordinates’ task performance. This hypothesis was tested in step three of equation three. The incremental variance explained was only 0.05% (n.s.), thus rejecting hypothesis six.

In addition, we examined in an exploratory fashion interactions between WAMS scores and preferential treatment. No interaction terms were significant in our regression equations.

Discussion

The present study examined the role of preferential treatment and stereotypical beliefs about women as managers in determining subordinate ratings of female leader effectiveness, commitment toward leader-assigned goals, and effects on subsequent subordinate performance. Previous evidence has established effects of preferential treatment on self-perceptions (e.g. Heilman et al., 1987) and some evidence exists that it may affect other organizational members’ perceptions and behaviors toward others (Heilman et al., 1993). In this study, however, only marginal effects of preferential treatment were demonstrated.

There are several possible explanations for our lack of preferential treatment effects. First, it may be that our manipulation was simply not strong enough to make a difference. In essence, the students’ instrumentality of being selected as the task leader may have been extremely low or even non-existent. Thus, a lack of reaction to the preferential treatment manipulation would not be surprising.

A second explanation involves the increasing trend toward “diversity” in both academic and business settings. It could be that the subjects viewed diversity as a legitimate goal, and thus preferential selection (e.g. using gender as a criterion) was seen as a legitimate means of selecting the leader. Given that students are consistently exposed to the diversity issue, it is possible that our subjects have internalized this goal as being completely legitimate. Obviously, this is a topic for future research.

The most interesting finding in this study is that WAMS scores were more influential than the presence of preferential selection on subordinate evaluations of female leaders as well as their commitment to the leader assigned goals. The actual relationships found are pictured in Fig. 2. This has intriguing implications for organizations. If stereotypical beliefs influence such activities as leader evaluations and subordinate goal commitment, what may the effects be on other, more fundamental attitudes toward female managers such as perceived credibility? If the same stereotypical biases are found to affect supervisors’ judgment of female subordinate capabilities, for example, it is possible that women’s opportunities could be significantly curtailed. As reflected in Fig. 2, future research may even find that stereotypes may work indirectly through subordinate evaluations of female leaders to influence goal commitment and subsequent task performance. Perhaps stereotypes present a much larger, more pervasive problem for females in the workplace than that of preferential treatment policies.
This stereotyping effect could be detrimental to women’s self-esteem, much like the effects of preferential treatment. Although we did not find any preferential treatment effects in this study, prior evidence suggests that preferential selection may undermine a woman’s feelings of self-efficacy because she does not receive positive ability feedback (Bandura, 1977, 1982; Heilman, et al., 1990). This can be especially detrimental in male-dominated careers, such as management, where women have been shown to exhibit less self-confidence and to attribute success to external rather than internal factors (Lenney, 1981; McMahan, 1982). The effects of negative stereotypes may be similar, in that women may not receive positive information or positive outcomes if they are thought to be “inappropriate” for managerial roles. Stereotypes may in fact be more detrimental to women because they tend to operate in all situations, rather than being limited to preferential treatment cases.

A danger may then exist, especially when stereotypes are used, of a spiral effect of sorts in which females’ self-efficacy perceptions are lessened, translating into their lower performance. When combined with possibilities of subordinates’ lessened commitment and lower leader evaluations, female managers’ self-worth could conceivably slide further. Thus, a downward spiral occurs, and it would become imperative to break the cycle so that women would not lose the accomplishments that they have strived so diligently for. Of course, all of this is conjecture, and future research would need to address whether such a process happens in actual organizational situations.
The results of this study also certainly suggest that negative stereotypes about females in the managerial role have a negative impact on leader performance ratings. This may be thought of as error, e.g. lower performance ratings are received because of nonjob factors. In the same line of thinking, it may be that positive stereotypes may also introduce error in the form of inflated performance ratings. This is an extremely intriguing point for future research to address.

We did not find that stereotypical views influenced subordinate task performance in a direct fashion. This is evidenced by a non-significant bivariate correlation of.16 between WAMS and schedules completed in Table I. It is possible, however, that an indirect relationship exists between the two. In our second regression model, we found that higher WAMS scores (e.g. less traditional stereotypical views) translated into higher commitment to leader-assigned goals (e.g. hypothesis five). Furthermore, in our third equation we found that higher commitment significantly led to higher subordinate task performance. Our hypotheses did not explicitly include this type of mediating relationship. It appears that stereotyping works through goal commitment to influence task performance. Further analysis of this relationship in the future is warranted.

A potential weakness of the experiment is its laboratory setting. Laboratory experiments are considered less realistic and less generalizable to the “real” world. However, such experiments allow for greater control over the experimental manipulation, thereby possibly resulting in stronger relational effects. The use of student confederates most likely weakened the impact of the preferential treatment manipulation upon subjects’ goal commitment and leader evaluations. Subjects may have perceived the confederates as less legitimate because of age proximity and other similarities to themselves. Also, the leader role in this experiment may not have been viewed as desirable, unlike actual leadership positions in organizations. Additionally, it may be that the strength of the stereotypical influences on our results stems from the more pervasive nature of those influences, i.e. in the absence of strong feelings about preferential treatment in the lab study, the students allowed their own personal attitudes toward women managers to play a larger role. Last, the short duration of the experiment does not represent well the on-going nature of organizational interactions.

Our study has implications for human resource policy. With increasing numbers of women in the workplace, the presence of glass ceilings, and the use of affirmative action, it is important to understand possible impediments that women may face in improving their positions in organizations. The possible hindrances resulting from stereotypes should not be taken lightly.

The necessity of affirmative action and equal employment opportunity ensures that women will have a place in organizational life. The demonstration of consequences to stereotypical beliefs, however, demands that we understand as much as possible the psychological processes underlying these consequences. Some evidence in our study combined with evidence of others (Heilman et al. 1987, 1993) points to some of the possible effects.

Future research should apply leader legitimization effects upon leader acceptance and evaluation, goal acceptance and goal commitment, and performance of subordinates in scenarios where women hold managerial roles. Since acceptance is critical to individuals’ commitment to goals, and since commitment theoretically leads to better performance, it is imperative that both stereotype and preferential selection effects be considered when selecting organizational leaders. This is not to say that these effects should prescribe elimination of minority promotion if such a promotion may be conceivably perceived as preferential treatment. As demonstrated by our study, this may not be the most important issue. Rather, the organization should understand individuals’ perceptions of others in these situations. The organization should seek to provide information to employees which neutralizes potential negative perceptions of women as organizational leaders. Such information may at least lessen the impact of stereotyping, and at most it may alter stereotypical views and result in a more positive environment for women in leadership positions.

Order entry of the WAMS and preferential treatment variables was arbitrary. The results do not change when variable order is changed.

References

Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84, 191-215.

Bandura, A. (1982). Self-efficacy mechanism in human agency. American Psychologist, 37, 122-147.

Barnes-Farrell, J. L., L’Heureux-Barrett, T. J., & Conway, J. M. (1991). Impact of gender-related job features on the accurate evaluation of performance information. Organizational Behavior and Human Decision Processes, 48, 23-25.

Beere, C. A. (1990). Gender roles: A handbook of tests and measures. Westport, CT: Greenwood Press.

Chacko, T. I (1982). Women and equal employment opportunity: Some unintended effects. Journal of Applied Psychology, 67, 119-123.

Deaux, K., & Emswiller, T. (1974). Explanations of successful performance on sex-linked tasks: What is skill for the male is luck for the female. Journal of Personality and Social Psychology, 29, 80-85.

Eagly, A. H., Makhijani, M. G., & Klonsky, B. G. (1992). Gender and the evaluation of leaders: A meta-analysis. Psychological Bulletin, 111, 3-22.

Earley, P. C., & Kanfer, R. (1985). The influence of component participation and role models on goal acceptance, goal satisfaction, and performance. Organizational Behavior and Human Decision Processes, 36, 378-390.

Garland, H., & Price, K. H. (1977). Attitudes toward women in management and attributions for their success and failure in a managerial position. Journal of Applied Psychology, 62, 29-33.

Heilman, M. E. (1984). Information as a deterrent against sex discrimination: The effects of applicant sex and information type on preliminary employment decisions. Organizational Behavior and Human Performance, 33, 174-186.

Heilman, M. E., Block, C. J., Martell, R. F., & Simon, M. C. (1989). Has anything changed? Current characterizations of men, women, and managers. Journal of Applied Psychology, 74, 935-942.

Heilman, M. E., & Herlihy, J. M. (1984). Affirmative action, negative reaction? Some moderating conditions. Organizational Behavior and Human Performance, 33, 204-213.

Heilman, M. E., Kaplow, S. R., Amato, M. A. G., & Stathatos, P. (1993). When similarity is a liability: Effects of sex-based preferential selection on reactions to like-sex and different-sex others. Journal of Applied Psychology, 78, 917-927.

Heilman, M. E., Lucas, J. A., & Kaplow, S. R. (1990). Self-derogating consequences of sex-based preferential selection: The moderating role of initial self-confidence. Organizational Behavior and Human Decision Processes, 46, 202-216.

Heilman, M. E., Rivero, J. C., & Brett, J. F. (1991). Skirting the competence issue: Effects of sex-based preferential selection on task choices of women and men. Journal of Applied Psychology, 76, 99-105.

Heilman, M. E., Simon, M. C., & Repper, D. P. (1987). Intentionally favored, unintentionally harmed? Impact of sex-based preferential selection on self-perceptions and self-evaluations. Journal of Applied Psychology, 72, 62-68.

Hollenbeck, J. R., & Klein, H. J. (1987). Goal commitment and the goal-setting process: Problems, prospects, and proposals for future research. Journal of Applied Psychology, 72, 212-220.

Hollenbeck, J., Klein, H., O’Leary, A., & Wright, P. (1989). An investigation of the construct validity of a self-report measure of goal commitment. Journal of Applied Psychology, 74, 951-956.

Hollenbeck, J., Williams, C., & Klein, H. (1989). An empirical examination of antecedents of commitment to difficult goals. Journal of Applied Psychology, 74, 18-23.

House, R. J., & Baetz, M. L. (1979). Leadership: Some empirical generalizations and new research directions. Research in Organizational Behavior, I, 341-423.

Jacobson, M. B., & Koch, W. (1977). Women as leaders: Performance evaluation as a function of method of leader selection. Organizational Behavior and Human Performance, 20, 149-157.

Lenney, E. (1981). What’s fine for the gander isn’t always good for the goose: Sex differences in self-confidence as a function of ability area and comparison with others. Sex Roles, 7, 905-924.

Locke, E. A., Shaw, K. N., Saari, L. M., & Latham, G. P. (1981). Goal setting and task performance: 1969-1980. Psychological Bulletin, 90, 125-152.

Locksley, A., Borgida, E., Brekke, N., & Hepburn, C. (1980). Sex stereotypes and social judgment. Journal of Personality and Social Psychology, 39, 821-831.

McMahan, I. D. (1982). Expectancy of success on sex-linked tasks. Sex Roles, 8, 949-958.

Nacoste, R. W. (1987). But do they care about fairness? The dynamics of preferential treatment and minority interest. Basic and Applied Social Psychology, 8, 177-191.

Nieva, B., & Gutek, V. (1981). Women and work: A psychological perspective. New York: Praeger.

Read, P. B. (1974). Source of authority and the legitimation of leadership in small groups. Sociometry, 37, 189-204.

Schein, V. E. (1973). The relationship between sex role stereotypes and requisite management characteristics. Journal of Applied Psychology, 57, 95-100.

Schein, V. E. (1975). Relationships between sex role stereotypes and requisite management characteristics among female managers. Journal of Applied Psychology, 60, 340-344.

Singer, M. E., & Singer, A. E. (1991). Justice in preferential hiring. Journal of Business Ethics, 10, 797-803

Stoppard, J. M., & Kalin, R. (1983). Gender typing and social desirability of personality in person evaluation. Psychology of Women Quarterly, 7, 209-218

How womenhave fared as Interscholastic Athletic Administrators since the passage of Title IX

Title VII and Title IX legislation were both intended to provide relief to individuals who were suffering from some form of exclusion or discrimination. Title VII of the Civil Rights Act of 1964 prohibited discrimination against any individual based on race, color, religion, sex, or national origin in any employment condition. In 1972, it was amended to expand its coverage to include government employees, educational institutions, and private employers of more than 15 persons. Title VII provided for a comprehensive prohibition on acts of employment discrimination (Anthony, Kacmar, & Perrewe, 2002), and Title IX of the Education Amendments of 1972 addressed the rights of all individuals, regardless of sex, to participate in educational programs or activities that receive federal financial assistance. In effect, Title IX required that both boys and girls have equal opportunities to participate in sports and derive the benefits of participation (National Coalition for Women and Girls in Education [NCWGE], 2002). With the full weight of the federal government behind these two cornerstone pieces of legislation, most individuals probably think that the inequity women had faced in athletics prior to 1972 has been eliminated. The results, more than 40 years after the passage of Title VII in 1964 and 30 years after the passage of Title IX, however, indicate that the successes girls have seen under Title IX have been eclipsed by the losses women have suffered despite their supposed protection under Title VII (Acosta & Carpenter, 2002b; Chesebro, 1985; Hart, Hasbrook, & Mathes, 1986; Heishman, Bunker, & Tutwiler, 1990; Lovett & Lowry, 1994; Oliphant, 1995; Pedersen & Whisenant, 2002). The complexities of expanding participation opportunities for girls in both interscholastic and collegiate athletics have resulted in the power and control of girls’ athletics being seized by men.

Gender and Athletics

College Athletics

To date, a great deal of research associated with Title IX issues has been guided by the works of Acosta and Carpenter (2002a), who focused on college athletics. They reported that since the passage of Title IX, participation opportunities have increased dramatically for female athletes. The number of female student athletes grew from 74,239 in 1981 to 150,916 in 2001. Women increased from 24.2 to 41.9% of the student athlete population (Bray, 2002). Female coaches and administrators, however, have not fared as well at their athletes. Acosta and Carpenter (2002b) noted that prior to the enactment of Title IX, 90% of the women’s teams had female head coaches. Thirty years later, in 2002, however, only 44% of the coaches were women. Women’s management of women’s programs has shown an even deeper decline. In 1972, women managed 90% of the women’s programs. By 2002, the number of programs being managed by a woman dropped under 20% (Whisenant, Pedersen, & Obenour, 2002). In addition, only 16.9% of the programs had a women serving as the head athletic director, despite the fact that 41.9% of the athletes (DeHass, 2002) participating in intercollegiate athletics were women.

High School Athletes

As with collegiate sports, the effects of Title IX on high school interscholastic athletics have been mixed for girls. The National Federation of State High School Associations (NFSHSA, 2002) reported that in 1971, prior to the passage of Title IX, the number of boys participating in high school athletics was approximately 3.7 million. At that same time, there were 294,015 girls participating in athletics. An immediate impact on girls’ participation rates was evident in the 1972-73 school year wherein the number of girls participating rose nearly threefold to 817,073. By the end of the 1977-78 school year, the number of girls participating in high school athletics had risen to over 2 million, a level that would not be seen again until the 1993-94 school year. Girls, who comprised only 7.4% of the student athletes prior to Title IX passage, represented 32.3% of the student athletes by the end of the 1977-78 school year. The total number of student athletes, boys and girls, participating in interscholastic athletics at that time had risen by 62.8% to over 6.4 million; girls accounted for 72% of that growth. During the following 15 years, 1978-79 to 1992-93, the number of girls participating in athletics fell below 2 million. Girls averaged 1.8 million (34.6%) of all high school student athletes. By the end of the 2000-01 academic school year, the NFHS reported that total athletic participation had increased for 12 consecutive years (“Athletic Participation,” 2001). Girls’ participation rates had increased for 18 of the previous 19 years to approximately 2.8 million, and girls accounted for 41.5% of all interscholastic student athletes by the end of the 2000-01 school year. Framed within the context of girls’ participation, Title IX appears to be a success.

Beyond the playing field, however, the effects of Title IX on the women who coached and managed girls’ interscholastic sports appear to have been unsettling. Studies that have focused on interscholastic athletics have demonstrated how men have dominated the coaching positions for girls’ sports (Chesebro, 1985; Hart, Hasbrook, & Mathes, 1986; Lovett & Lowry, 1994). A few studies have addressed the plight of female interscholastic athletic directors (ADs) and athletic administrators (Heishman, Bunker, & Tutwiler, 1990; Pedersen & Whisenant, 2002). Heishman, Bunker, and Tutwiler (1990), found that the number of female ADs in Virginia had declined in numbers from 187 to 60 over a 15-year period, 1971-72 to 1986-87. During that timeframe, the number of administrator positions declined from 458 to 382, a 17% decrease. The number of women in those positions declined at an accelerated rate of 68%, to 15.7% of the administrator positions. One positive finding was that the number of female ADs rose from.4% in 1971-72 to 6.5% in 1986-87. Overall, however, they found that the number of sports directors was declining, and there was a significant decline in the number of female administrators and coaches.

Other sources of data have indirectly shown the gender inequity in interscholastic athletics administration (Lovett & Lowry, 1994; Oliphant, 1995; Whisenant, 2001). Oliphant (1995) used a modified version of the collegiate gender-equity survey (38% response rate) to conduct a gender-equity study of high schools in the state of Iowa. The findings indicated that 98.1% of the ADs who responded were men. Lovett and Lowry’s study (Lovett & Lowry, 1994) of administrative structures in Texas schools indicated only 2% of the ADs in Texas were women. Raw data collected for a study of Florida and Texas ADs (Whisenant, 2001) indicated that 6% of the ADs in Texas and 13% in Florida were women. The National Interscholastic Athletic Administrators Association’s 2001-2002 NIAAA National Directory of State Athletic Director Associations (NIAAA, 2002) listed the officers and key personnel who were associated with each state’s high school athletic director association. An attempt to infer the sex of the individuals listed in the directory by their first names suggests that only 12% of the state association presidents were women, and 12% of the individuals serving as state liaisons for the NIAAA were women. The disparity is also evident at the national level, whereby the NIAAA’s own Board of Directors is dominated by men: 11 of the 13 members. As with the states’ leadership, women’s representation was below 15%. At every level of interscholastic athletics, men appear to dominate the key power positions.

The Importance of Equity

Equal access to the power and prestige of the athletic director position is mandated for two basic reasons. First, the law requires equal access under Title VII. Second, it is the right thing to do based upon the fundamental principles of integrity, fairness, and other maxims often used to validate the existence of school-sponsored athletics. The numbers to-date suggest that in reality, the law is not being followed. Regardless of the terminology (e.g., gender stratification, hegemony, or patriarchy) used to identify the root cause for the inequities that exist, the facts continue to support the propositions that men control athletics, and they have yet to share their power with women (Sage, 1990). The current environment is perpetuated by the systemic nature of the organizational culture by which athletics operates. Acosta and Carpenter (1992) found that hiring decisions are influenced by the gender of the hiring decision maker. They found that female ADs had a greater percentage of female coaches in their programs (50.8%) than did male ADs (46.4%). Lovett and Lowry (1994) came to the same conclusions in their study of head coaches in Texas: women hire women, and men hire men.

The issues that arise from the male domination of athletics carry on well beyond the playing field through a process associated with Bandura’s social-cognitive theory. Bandura (1997) noted that researchers have consistently found that cultural gender stereotyping influences career choice. “Such findings suggest that gender-related efficacy impediments arise less from the discrete skills themselves than from their linkage to stereotypically male occupations” (Bandura, p. 423). Hackett and Betz (as cited in Bandura, 1997) found that the beliefs women have with regard to their abilities and career goals were directly influenced by their families, the media, educational systems, and the culture in which they lived. Because girls see so few women in the power and leadership roles in interscholastic athletics as coaches or athletic directors, when they later make career choices, they may self-select themselves out of athletic careers. The results then, are a continuation of the perpetual cycle of male domination.

Again, this cycle is evident in data (National Center for Education Statistics [NCES], 2002) that indicate that 34.1% of all public school principals were at one time a coach or an AD. Over half (56.1%) of the high school principals in the survey had a background in athletics. These results could help explain the findings of de Mello and Broughman (1996), who assessed the gender of high school principals across the United States. They reported that only 13.8% of the principals were women, despite the fact that 53.1% of the high school teachers were women. Snyder and Hoffman (2002) reported that 42.9% of all high school seniors participated in athletics, and 30.3% of all senior girls participated in athletics. With high school athletics playing such a high impact role in the lives of so many young girls, the absence of female role models in the power positions may be sending girls the message that athletics is not a viable career choice for women. The purpose of this study was to assess the magnitude of the male control over interscholastic athletics, which would provide insight into how well women have faired under Title IX at the interscholastic level, not as participants or coaches, but beyond the playing field into the offices of athletic administrators.

Research Questions

The research questions addressed in this study were:

1. To what extent did men dominate the administration of interscholastic athletics as evidenced by the gender composition of interscholastic athletic administrators within each state’s athletic director associations?
2. To what extent did the gender composition of state athletic director associations differ regionally across the United States?

Method

Each state’s interscholastic athletic administrators association was contacted by e-mail. The associations were asked if their state’s AD association had demographic data on its members, and, if so, whether they would provide the researcher with the number of male and female members. Whenever the e-mail was returned undelivered, the following predetermined contact sequence was followed: (a) the President of the state athletic administrators association as noted in the NIAAA National Directory of State Leadership, (b) the NIAAA Liaison for each state, (c) the President of each state’s coaches association as listed on the website of the National High School Coaches Association. To determine if gender differences existed regionally, the states were grouped into five regions: West, Midwest, Great Lakes, Northeast, or Southeast. The regional groupings matched the state groupings used by the NFHS on its website.

Results

Twenty-six of the 50 state associations contacted responded to the request for the gender composition of their membership. Of those that responded, 22 state associations provided the requested information. Four states indicated the data were not available. All five of the regions had representative data for analysis. The 22 state associations reported a membership level of 7041 interscholastic athletic administrators or athletic directors. Of those members, 6142 were men (87%) and 899 were women (13%). The first research question sought to determine the extent to which men dominated the administration of interscholastic athletics as evidenced by the gender composition of interscholastic athletic administrators within each state’s athletic director association. A chi-square test was used to compare group frequencies between men and women. The difference was significant; men outnumbered women, [chi square](1, N = 7041) = 3904, p <.001.

The second question sought to determine if the gender composition of state athletic director associations differed regionally across the United States. The men/women composition of the regions was as follows: West 83/17%, Midwest 93/7%, Great Lakes 88/12%, Northeast 84/16%, and Southeast 88/12%. Gender comparisons were made utilizing an analysis of variance. The findings indicated that there were no significant differences in the percentage of women in each state’s association when the memberships were compared on a regional basis, F(4, 17) = 1.858, p =.164.

Discussion

The results of this study indicate that although more girls are playing more athletics, they are playing in programs controlled by men. Since the passage of Title IX, girls can play the games, but women are not allowed to control the games. As girls become women, they relinquish control of their sports to men. As with intercollegiate athletics, men have gained control and maintain the power over interscholastic athletics. The pattern of male domination of athletics is evident in the fact that 87% of the athletic administrators at the local and state levels were men. Furthermore, the inequity is not limited geographically within the United States. In all regions, women are apparently subjected to job discrimination in interscholastic athletics administration.

These findings are disturbing when we consider the gains women have made in the private sector. In the business community, the U.S. Department of Labor reported (as cited in Bajdo & Dickson, 2001) that in 1997 women held 40% of the management positions within American businesses. However, they only held 3-5% of the top executive positions. Within the public school structure, the AD’s position can be aligned with that of a manager, in that the school district’s superintendent can be considered the top executive. Given that structure, women in interscholastic athletics lag far behind their peers in other management fields.

Several remedies recommended by Kandel (1995) to address inequities in the private sector are well suited for use in interscholastic athletics. Women involved with the administration of interscholastic athletics should not be pigeonholed into positions that do not have professional development opportunities or do not logically lead to advancement with greater responsibilities and visibility. Job openings should be effectively communicated in a manner that reaches women. For example, the Texas High School Coaches Association website posts jobs for coaches and ADs. Although the Texas Girls’ Coaches Association website listed 25 jobs on September 19, 2002, none of them were for administrative positions. Professional development and training opportunities should be exploited by women to enhance their credentials. Programs such as the NIAAA’s certification program enables current administrators and women who seek to move into the field of interscholastic athletics administration to make up ground lost to men who have already acquired AD positions; the program provides the opportunity to be certified as a Registered Athletic Administrator (RAA), Certified Athletic Administrator (CAA), and Certified Master Athletic Administrator (CMAA). Involvement in organizations and associations such as The National Association for Girls and Women in Sport and the Womens Sports Foundation can also assist women in building both professional and social networks which can lead to advanced career opportunities in sport administration. Although several suggested remedies have been mentioned, further research is need to explore fully any systemic barriers that may be in place that deny women access to these positions. One key question that needs to be addressed is: how many women apply for positions that are awarded to men?

A byproduct of the continued inequity in interscholastic athletics is the perpetuation of myths associated with gender differences in management. The absence of highly visible women in key leadership roles will negatively influence how girls and boys perceive women in the work place. Boys will grow up with the perception that successful managers are required to have characteristics, attitudes, and temperaments more associated with men than women (Schein, Mueller, Litucy, & Liu, 1996). These boys who will become men in the workplace may also not look upon women with favor when making judgments of them as managers (Deal & Stevenson, 1998). In addition, many girls may abandon career pursuits associated with athletics because of the absence of a network of women to serve as role models and mentors. The ultimate outcome of this pervasive discrimination will be the perpetual cycle of male domination in interscholastic athletics.

As more women gain entry into the AD ranks, they will be able to form gender friendly networks and act as mentors to those women who will follow in their footsteps. Women in the leadership roles will also be able to serve as role models for both boys and girls, thereby demonstrating that sex does not predetermine managerial success. Title IX has benefited girls by increasing their participation opportunities, however it is evident that those gains were made at the expense of women’s rights to employment under Title VII. With 14,505 school districts and over 20,000 high schools (NCES, 2002) in the public school systems across the nation, opportunity to advance abounds.

References

Acosta, V., & Carpenter, L. (1992). Women in intercollegiate sport. A longitudinal study–Fifteen year update, 1977-1990. Unpublished manuscript. Brooklyn College, New York.

Acosta, V., & Carpenter, L. (2002a). Women in intercollegiate sport: A longitudinal study – Twenty five year update 1977-2002. Web.

Acosta, V., & Carpenter, L. (2002b). Status of women as administrators. Web.

Anthony, W., Kacmar, K., & Perrewe, P. (2002). Human resource management: A strategic approach (4th sed.). Fort Worth, TX: Harcourt College Publishers.

Athletic participation increases for 12th straight year. (2001). NFHS News, 19(3), p. 4.

Bajdo, L., & Dickson, M. (2001). Perceptions of organizational culture and women’s advancement in organizations: A cross-cultural examination. Sex Roles, 45, 399-414.

Bandura, A. (1997). Self-efficacy: The exercise of control. New York: Freeman.

Bray, C. (2002). NCAA year-by-year sports participation 1982-2001. Indianapolis, IN: National Collegiate Athletic Association.

Chesebro, P. (1985). High school women coaches and athletic directors: Diminishing numbers. Illinois Journal of Health, Physical Education, and Recreation, 20, 7-9.

Deal, J., & Stevenson, M. (1998). Perceptions of female and male managers in the 1990s: Plus ca change… Sex Roles, 38, 287-300.

DeHass, D. (2002). 1999-00 NCAA gender equity report. Indianapolis, IN: National Collegiate Athletic Association.

de Mello, V., & Broughman, S. (1996). SAS by state 1993-94 schools and staffing survey: Selected results (NCES 96-312). Washington, DC: National Center for Education Statistics.

Hart, B., Hasbrook, C., & Mathes, S. (1986). An examination of the reduction in the number of female interscholastic coaches. Research Quarterly for Exercise and Sport, 57, 68-77.

Heishman, M., Bunker, L., & Tutwiler, R. (1990). The decline of women leaders (coaches and athletic directors) in girls’ interscholastic sport programs in Virginia from 1972 to 1987. Research Quarterly for Exercise and Sport, 61, 103-107.

Kandel, W. (1995). Affirmative action and the glass ceiling: Contract compliance and litigation avoidance. Employee Relations Law Journal, 21, 109-118.

Lovett, D., & Lowry, C. (1994). “Good old boys” and “good old girls” clubs: Myth or reality? Journal of Sport Management, 8, 27-35.

National Center for Education Statistics (NCES). Schools and staffing survey, 1999-2000: Overview of the data for public, private, public charter, and bureau of Indian affairs elementary and secondary schools. (NCES 2002-313). Washington, DC: National Center for Education Statistics.

National Coalition for Women and Girls in Education (NCGWE). (2002). Title IX athletic policies: Issues and data for education decision makers. Washington, DC.

National Federation of State High School Associations (NFSHSA). (2002). National federation of state high school associations participation study 1971-01. Indianapolis.

National Interscholastic Athletic Administrators Association (NIAAA). (2002). NIAAA national directory of state athletic director associations.

Oliphant, J. (1995). Iowa high schools athletic gender-equity study. Web.

Pedersen, P., & Whisenant, W. (2002). Successful when given the opportunity: Investigating gender representation and success rates of interscholastic athletic directors. Manuscript submitted for publication.

Sage. G. (1990). Power and ideology in American sport: A critical perspective. Champaign, IL: Human Kinetics Books.

Schein, V., Mueller, R., Lituchy, T., & Liu, J. (1996). Think manager-Think male: A global phenomenon? Journal of Organizational Behavior, 17, 33-41.

Snyder, T, & Hoffman, C. (2002). Digest of education statistics, 2001. Washington, DC: National Center for Education Statistics.

Whisenant, W. (2001). [Athletic director demographics]. Unpublished raw data.

Whisenant, W., Pedersen, E, & Obenour, B. (2002). Success and gender: Determining the rate of advancement of intercollegiate athletic directors. Sex Roles, 47, 485-491.

Warren A. Whisenant, To whom correspondence should be addressed at Department of Health and Human Performance, College of Education, University of Houston, Houston Texas 77204-6321; e-mail: [email protected].

The relationship between gender role stereotypes and requisite military leadership characteristics

During the 1970s, the U.S. military academies participated in one of the largest changes that helped to recast the role of women in the American military. In 1975, President Gerald R. Ford signed Public Law 94-106, which authorized admission of women to the Army, Navy, and Air Force (Devilbiss, 1990), and, in the following year, the first class of women entered these service academies (Colaw, 1996). Now, more than 20 years later, over 6,000 women have graduated and received commissions from the military academies. Although senior officers at the military academies indicate that their institutions have attempted to create an environment where leadership by both female and male students is encouraged and recognized, traditional military leadership models have been male-oriented and have posed challenges for women in the military (Versaw, 1997; Youngman, 2001). Military responsibilities are often perceived as contrary to societal expectations about appropriate roles for women (Arkins & Dobrofsky, 1978; Eagly, Karau, & Makhijani, 1995; Versaw, 1997). The purpose of this study is to examine perceptions among military academy students regarding gender roles associated with military leadership positions. An understanding of current perceptions among military academy students regarding the relationship between gender and leadership success characteristics is critical for the success of continued initiatives geared toward gender integration and leadership development at military academies and other male-dominated organizations.

Gender Stereotyping

Gender stereotyping has been defined as “the belief that a set of traits and abilities is more likely to be found among one sex than the other” (Schein, 1978, p. 259). This tendency to attribute characteristics to gender can be extended to occupations that are more likely to be held by primarily men or women (Eagly & Johnson, 1990; Norris & Wylie, 1995). For example, the military has traditionally been regarded as a masculine occupation (Arkins & Dobrofsky, 1978; Youngman, 2001); thus, military leaders, when selecting or promoting other soldiers to be leaders, may look for personal attributes thought to be more characteristic of men than of women. More men than women may be perceived as having leadership potential and thus be given more opportunities to exhibit leadership. These perceptual dynamics may then strengthen stereotypes that women are less qualified than men for military leadership positions, and this can have far-reaching implications for recruitment, selection, placement, evaluations, and promotions, as well as for military readiness and performance (Norris & Wylie, 1995; Youngman, 2001).

Over the past two decades there has been much research on gender stereotyping. Early studies, which focused on the differences in perceived stereotypical expectations of women and men, generally indicated that men are perceived as better suited than women for leadership roles (Nieva & Gutek, 1981). For example, a classic early survey conducted by Bowman, Worthy, and Greyser (1965, p. 28) indicated that women were perceived to be “temperamentally unfit” for managerial positions. Bass, Krusell, and Alexander (1971) also reported similar negative perceptions, whereas early studies conducted by Broverman and colleagues (Broverman, Broverman, Clarkson, Rosenkrantz, & Vogel, 1970; Broverman, Vogel, Broverman, Clarkson, & Rosenkrantz, 1972) identified traits predominantly attributed to men as more positive, such as competency and rationality. Traits attributed to women were less positive and tended to be restricted to affective characteristics such as warmth and expressiveness. Other studies, such as those conducted by Rosen and Jerdee (1973, 1974a, 1974b) and Bartol and Butterfield (1976), showed that gender stereotypes were related to expectations regarding the appropriateness of specific supervisory behaviors, which could influence supervisory performance evaluations and personnel decisions.

Job-Relevant Stereotypes

Whereas the above studies investigated the general nature of gender role stereotypes, early research by Schein (1973, 1975) concerned the relationship between gender role stereotypes and perceptions of effective management characteristics. Schein (1973) initially investigated managers’ perceptions of the relationship between gender stereotypes and characteristics perceived as required by successful managers. Specifically, 300 male middle managers rated women, men, or successful middle managers on the Schein Descriptive Index, a 92-item scale that elicits perceptions of gender role stereotypes. It was found that successful middle managers were perceived to possess characteristics, attitudes, and temperaments more commonly ascribed to men than to women. There were significant similarities between ratings of men and successful managers, whereas there was little resemblance in attributes assigned to women and successful managers.

An analogous study of female middle managers suggested that they held similar perceptions of a male managerial model (Schein, 1975). Female managers accepted stereotypical masculine characteristics as a model for success in management implying that female managers in the study were as likely as male managers in the previous study to make selection, promotion, and placement decisions in favor of men (Schein, 1975).

In a later replication of these earlier studies, Brenner, Tomkiewicz, and Schein (1989) found that male middle managers adhered to a masculine managerial stereotype in their perceptions of managerial success. However, female middle managers in the study rated successful middle managers as having both masculine and feminine characteristics and attributes, which indicates that female middle managers held more expanded gender role perceptions of managerial success.

Additional recent studies of undergraduate and graduate students (Dodge, Gilroy, & Fenzel, 1995; Norris & Wylie, 1995; Schein, Mueller, & Jacobson, 1989) have shown similar patterns. Dodge et al. (1995) and Schein et al. (1989) reported that female management students did not gender-type the managerial job, but male management students characterized the successful manager with stereotypically masculine traits. Norris and Wylie (1995), who used the Personal Attributes Questionnaire instead of Schein’s Descriptive Index, also found that male college students stereotyped the successful managers in masculine terms but women students did not.

Role Congruity Theory

Role congruity theory offers an explanation for the gender stereotyping of leadership positions by maintaining that perceived gender roles may conflict with expectations regarding leadership roles, especially when an occupation is held predominantly by one sex (Eagly et al., 1995). Meta-analyses of gender effects in the evaluation and effectiveness of leaders indicate support for role congruity explanations. In the analyses of 61 empirical studies, it was found that women leaders tended to be devalued to a greater extent when they held leadership positions in male-dominated areas or fields and when they exhibited stereotypically masculine leadership styles, such as autocratic or directive styles (Eagly, Makhijani, & Klonsky, 1992).

Academy Cadet Gender Role Stereotypes

The primary purpose of this study was to determine the extent to which associations between gender role stereotypes and requisite leadership characteristics exist among military officer candidates at the United States Air Force Academy (USAFA) and to compare the outcomes to the previous studies conducted by Schein (1973, 1975) and Schein et al. (1989). The Academy population is unique, in that the mission of USAFA is specifically focused on developing leaders by exposing cadets to extensive leadership training and providing opportunities and experiences in a variety of leadership challenges.

Despite the efforts by the institution to expose cadets to a variety of leadership models and despite recent research results (Norris & Wylie, 1995; Schein et al., 1989) that suggest that current college students may hold less gender-specific managerial stereotypes than students three decades ago, cadets at USAFA are expected to hold gender role stereotypes. Several factors weighted the hypotheses toward mirroring Schein’s early results as opposed to her more recent findings. These factors include the masculine tradition of the military and its academies, the nature of applicants who apply to military institutions, and research that specifically assessed gender stereotypes in military populations (Boldry, Wood, & Kashy, 2001).

Foremost, the military is still considered a masculine occupation. With less than 16% of the entire military officer corps composed of women and even less, slightly over 13%, of the service academies composed of women (Manning & Wright, 2000), some argue that “the military is the most male dominated of all social institutions” (Johnson, 1997, pp. 3-4). Because there are few women at the service academies, there are also few cadet women leaders as role models. The traditions associated with service academies are traditions associated with men, such as short hair and performing push-ups as punishment. In addition, cadets within the academies receive less exposure to the world outside their academy than do civilian college students. Thus, cadets are strongly steeped in the attitudes, norms, and traditions within the academy walls during their 4 years at the academy. Recent research with military college trainees by Boldry et al. (2001) demonstrated that both male and female military trainees evaluated women as less suitable for military work. The college military trainees in the study perceived that men, more than women, possessed the motivation and leadership skills necessary for effective military performance. The authors concluded that gender stereotypes most likely accounted for the evaluative differences because the men and women in the study did not differ on objective performance indices.

The combination of the strong male-dominated population, the traditions associated with service academies, and recent research showing gender bias among both male and female military trainees suggest that cadets’ perceptions of gender stereotypes may be more pronounced than those of their civilian college student counterparts. Thus, the first hypothesis was as follows:

• Hypothesis 1: Male and female cadets will perceive successful officers as possessing attitudes, characteristics, and temperaments more commonly ascribed to men in general than to women in general.

Work Experience With Women Leaders and Gender Role Stereotypes

Bowman et al. (1965) found that men and women who reported actual experience working with women managers were more likely than those without such experience to be favorable toward women in managerial positions. Further, results of studies by Baron (1984) and Heilman, Block, Simon, and Martell (1989) suggested that employees’ gender role stereotypes about female managers vanished after employees worked for them. Powell’s extensive review of subordinates’ responses to managers (Powel, 1990), as well as research and reviews by Eagly and associates (Eagly, 1987; Karau & Eagly, 1999), corroborates these earlier findings. Laboratory studies indicate that once subordinates have worked for both female and male managers, the effects of stereotypes disappear, and managers are treated as individuals rather than as representatives of their sex (Powell, 1990). In alignment with these research findings, it is expected that cadets with more exposure to female cadet commanders will have a less masculine gender role stereotype of successful officers than will cadets who have had less exposure to female commanders. Thus, the following hypothesis was advanced:

• Hypothesis 2: Cadets with more exposure to female commanders will perceive successful officers as possessing attitudes, characteristics, and temperaments more commonly ascribed to both men and women than will cadets with less exposure.

Successful Performance and Gender Role Stereotypes

Research studies on gender and leadership effectiveness generally reveal equal effectiveness of male and female leaders in the aggregate, when generalized across a variety of studies in a variety of settings. However, consistent with role congruity theory, leadership behaviors exhibited by male and female leaders may differ and may be evaluated differently depending on the extent to which the particular role is defined in masculine terms (Eagly et al., 1995; Thompson, 2000). Women in male-dominated areas or fields tend to be seen as less effective than their male counterparts. Likewise women may be evaluated negatively when they violate gender role expectations by failing to exhibit consideration or affective leadership behaviors, whereas there may be less expectation that male managers will exhibit these types of leadership behaviors (Dobbins & Platz, 1986; Eagly & Johnson, 1990; Nieva & Gutek, 1981; Russell, Rush, & Herd, 1988).

Other studies on leadership prototypes and cognitive categorization processes suggest that individuals’ ratings of traits associated with successful leadership vary in self-serving patterns. For example, when rating the prototypicality of traits associated with the social category of leadership, individuals tend to rate positive traits that they believe themselves to possess as more highly prototypical of leadership than traits they do not believe themselves to possess (Dunning, Perie, & Story, 1991).

At the USAFA, each cadet’s military performance is evaluated and documented by officer and cadet leaders through a Military Performance Average (MPA), which is similar to a grade point average. A cadet earns an aggregated rating (on a 4.00 scale) from faculty, peers, and military leaders based on his or her performance in the areas of duty performance, initiative, followership and teamwork, character, and leadership and supervision. Cadets who score high (> 3.00) have been rated as excelling in each of these areas, whereas cadets who score low ( <2.00) are on military probation, are closely monitored, and are provided additional leadership development training to gain the skills essential to become an officer.

Initial empirical research suggests a relationship between perceived successful military leadership characteristics and gender role stereotypes, however, the literature is currently too immature to offer support for a rigorous hypothesis regarding the relationship between performance level and gender-role stereotypes. Therefore, exploratory hypotheses were proposed. It was expected that a cadet’s MPA would be related to his or her perceptions of successful officer stereotypes. These perceptions were expected to differ for male and female cadets, because of the increased salience of their own characteristics (e.g. gender) when making attributions about their own successful leadership performance (Dunning et al., 1991). The following exploratory hypothesis was formulated:

• Hypothesis 3: Perceptions, of successful officers as possessing attitudes, characteristics, and temperaments more commonly ascribed to men in general or women in general, are moderated by cadet performance level and gender.

Seniority and Gender Role Stereotypes

Tenure or length of service was found by Schein (1975) to moderate the relationship between gender role stereotypes and perceived requisite management characteristics. In Schein’s study, managers with greater than 5 years of service perceived “women” and “managers” as more similar than did their junior counterparts (Schein, 1975). Likewise, results of recent studies and reviews (Eagly & Johnson, 1990; Karau & Eagly, 1999; Kolb, 1999) have suggested that experience and time diminish the perceived importance of masculinity in leadership stereotypes.

As indicated earlier, the USAFA provides structured opportunities for cadets to learn and develop as leaders throughout their 4-year education. Cadets take academic courses on leadership from USAFA civilian and military faculty, and they participate in leadership development workshops. In addition, cadet careers span a leadership hierarchy, and they take on actual work roles in this hierarchy: freshmen are considered subordinates, sophomores are administrators, juniors are first line supervisors, and seniors are commanders or upper management. Because senior cadets have had ample opportunities to study and experience leadership, increased seniority may be related to weaker masculine gender role stereotypes for officer leadership. On the other hand, if the institutional culture strongly links masculine gender role stereotypes with leadership, increased seniority may be related to more strongly held masculine gender role stereotypes for the officer role. Thus, seniority was investigated as an exploratory variable that serves as a moderator in gender role stereotypes for leadership.

• Hypothesis 4: Perceptions, of successful officers as possessing attitudes, characteristics, and temperaments more commonly ascribed to men in general or women in general, are moderated by cadet seniority.

Method

Participants

Participants were drawn from the USAFA, a 4-year undergraduate military service college. Student participants were identified to represent 15 of the 40 total cadet squadrons. The 15 squadrons that participated in the study were chosen by a random numbers generator. Each cadet squadron is composed of approximately 120 randomly assigned cadets across all four class years. Only 10-16 women reside in each squadron. To increase response rates from women, three female athletic teams (softball, rugby, and swimming) were also surveyed. Seven hundred ninety surveys were administered, and 755 (96%) were usable. Surveys were eliminated if:

1. demographic data, specifically sex, was not reported;
2. over 10% of the items were not rated;
3. no variability was demonstrated in item ratings.

Of the 755 volunteers who completed usable surveys, 635 were men and 140 were women. Responses by class year included 149 seniors, 170 juniors, 224 sophomores, and 212 freshmen (20 surveys did not include class year). Cadets ranged in age from 17 to 24 years; the mean student age was 20.08 years. Nearly one-third of the cadet respondents indicated that they had not been directly supervised by a female cadet commander; nearly one-fourth of the respondents reported having had at least two female cadet commanders. Respondents’ MPA ranged between 1.00 and 4.00; the mean cumulative MPA was 2.84.

Materials

The Schein 92-item Descriptive Index (Brenner et al., 1989; Schein, 1973, 1975; Schein et al., 1989) was used to identify both gender role stereotypes and characteristics of successful officers. Three forms of the questionnaire were used. All forms contained the same descriptive terms and instructions. However, one form asked for a description of women in general, one for a description of men in general, and one for a description of successful officers. Each cadet received only one form of the questionnaire.

The instructions on the three forms of the descriptive index were as follows:

• On the following pages you will find a series of descriptive terms commonly used to characterize people in general. Some of these terms are positive in connotation, others are negative, and some are neither very positive nor very negative.
• We would like you to use this list to tell us what you think (women in general, men in general, successful officers) are like. In making your judgments, it may be helpful to imagine that you are about to meet a person for the first time and the only thing you know in advance is that the person is (an adult female, an adult male, or a successful officer). Please rate each word or phrase in terms of how characteristic it is of (women in general, men in general, successful officers).

Respondents rated the descriptive terms on a 5-point scale: 5 indicated “characteristic,” 3 indicated “neither characteristic nor uncharacteristic,” and 1 indicated “not characteristic.” Procedure The three versions of the modified Descriptive Index were distributed in an alternating manner so that a similar number of each form was completed by each class year. Trained data collectors informed the cadets that their participation was completely voluntary and confidential. The cadets were not told that there were different versions of the form.

Of the 775 usable questionnaires collected, 267 (34.5%) were completed on “men in general,” 274 (35.4%) on “women in general,” and 234 (30.2%) on “successful officers.” The large sample size facilitated the segmentation of the respondents for secondary analyses. However, even a sample of this size has limitations. Each analysis required first a split into thirds (by questionnaire form), then a split by gender, which yielded cell sizes that ranged from 180 to 231 for male participants and from 36 to 54 for female participants. Secondary analyses, such as those on the variable MPA, required a further split that identified only a small group of women who had earned low MPAs (n = 34). Further gradations of variables might shed further light on women’s perceptions, but female sample sizes became inadequate to generate confidence in the patterns. Therefore, the secondary analyses are reported for the male participants only.

Results

In a replication of Schein and colleagues’ analysis procedures (Schein, 1973, 1975; Schein et al., 1989), the degree of resemblance between the descriptions of “men” and “officers” and between “women” and “officers” was determined by computing intraclass correlation coefficients from two randomized-groups analyses of variance (see Hays, 1963). The classes (or groups) were the 92 descriptive items. In the first analysis, the scores within each class were the mean item ratings of “men” and “officers,” whereas in the second analysis they were the mean item ratings of “women” and “officers.” According to Hays, the larger the value of r’, the more similar the observations in the class tend to be relative to observations in different classes. Thus, the smaller the within-item variability, relative to the between-item variability, the greater the similarity between the mean item ratings of either “men” and “officers” or “women” and “officers.” These analyses were performed separately for the male and female samples.

Intraclass correlation coefficients and the results of the analyses of variance mean item ratings are shown in Table I. For comparison, intraclass correlation coefficients from the Schein et al. (1989), Brenner et al. (1989), and Schein (1973, 1975) studies are included.
There was a significant resemblance between the ratings by male cadets of “men” and “officers” (r’ =.41, p <.01), whereas there was no significant resemblance between their ratings of “women” and “officers” (r’ = -.11, ns). These results confirm the hypothesis that the men would perceive military leaders as possessing characteristics more commonly ascribed to men in general than to women in general.
As shown in Table I, these results are similar to the results of previous studies.

There was a significant resemblance between the ratings by female cadets of “women” and “officers” (r’ =.30, p <.01), but not between “men” and “officers” (r’ =.25, ns). However, these intraclass coefficients were not significantly different from each other (z = 0.36, ns). Therefore, the hypothesis that women would perceive successful military leaders as possessing characteristics more commonly ascribed to men than to women was not confirmed. These results are similar to results of previous research, which showed that women managers’ perceptions of the resemblance between men and managers were not significantly greater than between women and managers (Brenner et al., 1989). Thus, Hypothesis 1 was supported only for male cadets.

Comparison of Responses by Work Experience With Women Leaders

The effect of exposure to women leaders on the relationship between gender role stereotypes and requisite officer characteristics was examined. The male cadets were divided into two experience levels:

1. those who had not worked for female commanders
2. those who had worked for two or more female commanders.

Cadets who had experience with only a single female commander were not included as their ratings would conceivably be biased by the influence of the style and effectiveness of the particular woman. Included in Table II are the intraclass correlation coefficents computed between the mean ratings of the descriptions of “men” and “officers” and between the ratings of “women” and “officers” within the two experience levels.

For men with no experience of women leaders and men with several experiences of women leaders, there was a significant resemblance between the ratings of “men” and “officers” (r’ =.34, p <.01, and r’ =.45, p <.01, respectively) and a nonsignificant resemblance between the ratings of “women” and “officers” (r’ = -.09, ns, and r’ = -.13, ns, respectively). Further, there was no significant difference between the intraclass correlation coefficients for the different experience levels. Therefore, the second hypothesis that respondents who have more experience and exposure to women leaders would perceive successful military leaders as possessing traits typically associated with both men and women was not supported.

Comparison of Responses by Successful Leadership Performance

To determine whether successful leadership performance moderates the demonstrated relationship between gender, gender role stereotypes, and requisite officer characteristics, the total sample was divided into two levels of performance, high and low. Because the numbers of low-performing female respondents were small, an interrater reliability analysis was performed to determine the consistency of ratings. Cronbach’s alphas ranged from.83 to.92 for female high- and low-performing respondents, respectively, whereas the Cronbach’s alphas ranged from.88 to.93 for low- and high-performing male respondents, respectively. As all reliabilities were relatively high, we proceeded with the data analysis.

As shown in Table II, the results indicate that level of performance did not moderate the male respondents’ perceptions regarding the similarities between “men” and “officers” and between “women” and “officers.” Statistically the same pattern of intraclass correlation coefficients was found for both performance levels, in that both low- and high-performing men rated “men” and “officers” as similar (r’ =.30, p <.01, and r’ =.45, p <.01, respectively) and “women” and “officers” as dissimilar (r’ =.00, ns, and r’ = -.16, ns, respectively). Performance level did moderate the gender role stereotypes exhibited by female respondents. Low-performing women did not show a significant correlation between the ratings of “men” and “officers” (r’ = -.01, ns), but did show a moderate correlation between ratings of “women” and “officers” (r’=.34, p <.01). Higher performing women demonstrated a moderate and significant correlation between the ratings of “men” and “officers,” and between “women” and “officers” (r’ =.37, p <.01, and r’ =.29, p <.01, respectively). This difference in perceptions regarding similarities between “men” and “successful officers” by low- and high-performing female respondents (r’ = -.01, ns, and r’ =.37, p <.01), was statistically significant (z = 2.66, p <.01). Further, the difference in perceptions regarding the similarities between “women” and “successful officers” by high-performing male and female cadets (r’ = -.16, ns, and r’ =.29, p <.010, respectively) was also statistically significant (z = 3.07, p <.01). Therefore, partial support was provided for the exploratory hypotheses, in that successful female cadets’ perceptions of a successful military leader include traits commonly ascribed to women, whereas successful male cadets maintained a predominantly masculine perception of a successful leader. However, contrary to the hypothesis explored, female successful performers expanded their perceptions of successful military leaders to include traits commonly ascribed to men as well as women, whereas their lower performing counterparts described military leaders in feminine terms.

Comparison of Responses by Seniority

The effect of seniority upon the relationships between gender role stereotypes and requisite officer characteristics was also examined. Table II shows the intraclass correlation coefficients computed between the mean ratings of “men” and “officers” and between the ratings of “women” and “officers” within the four class years for male respondents only.

The results reveal that length of service did not have an effect upon the relationships. Among male respondents with more than 1 year at the military academy, the intraclass correlation coefficients between ratings of “men” and “officers” were.49,.30, and.49, p <.01, for the senior, junior, and sophomore cadets respectively. Responses from men with less than 1 year of service did not demonstrate a significant resemblance between the mean ratings of “men” and “officers” (r’ =.27, p <.01). However, the intraclass correlation was not significantly higher for sophomore, junior, and senior cadets compared to freshman cadets (z = 1.73, ns). The intraclass coefficients between ratings of “women” and “officers” were not significantly different by seniority. These results are similar to the Brenner et al. (1989) study, but counter to Schein’s (1975) study; they do not support the prediction that seniority would moderate cadets’ perceptions of gender role stereotypes.

Descriptive Items

Although an understanding of the resemblance between the mean descriptive ratings of men and officers and women and officers was our primary interest, an exploratory examination of the specific descriptive items on which “men” or “women” were perceived as different from “officers” was made by performing multivariate analyses of variance with post hoc Tamhane’s T2 analyses. Tamhane’s T2 is a conservative post hoc pairwise comparisons test for unequal sample sizes. An alpha level of.001 was used as the criterion of significance. The analyses incorporated the three groups (“officer,” “men,” and “women” surveys) and the two genders (men and women cadets) for the 92 items. The 24 items where “successful officers” were rated differently from “women in general,” or where “successful officers” were rated differently from “men in general,” are shown in Table III.

The results reveal that men rated 19 items as more similar between “officers” and “men” than between “officers” and “women.” As shown in Table III, both male and female respondents rated “officers” as significantly different from “women” and similar to “men” on five items (authoritative, feelings not easily hurt, frank, self-reliant, and submissive). Male respondents rated “officers” as significantly different from “women” and similar to “men” on 14 additional items (e.g., talkative, aggressive, desire for friendship, forceful, reserved). For example, the mean male cadets ratings of aggressive and forceful in describing “successful officers” (Ms = 3.76 and 3.31, respectively) were statistically different from the mean male and female cadets ratings of these adjectives in describing “women in general” (Ms = 2.86 and 2.76, respectively). In comparison, the female cadets’ ratings of “successful officers” (Ms = 3.54 and 3.22) were not statistically different from the male and female cadets’ ratings of “women in general” on these two adjectives. Both men and women rated “successful officers” as similar to “women in general” and significantly different from “men in general” on four items (cheerful, kind, neat, and sympathetic). Women also rated “successful officers” and “women in general” as similar on one additional item, need for autonomy.

Table IV depicts items where women in particular rated “successful officers” as similar to “men in general” or to “women in general.” As shown in Table IV, women rated “successful officers” as more direct, firm, and industrious, and less demure, fearful, passive, and shy than “women in general.” In contrast, men rated “successful officers” as significantly different from both “men in general” and “women in general” on these eight characteristics. Both men and women rated “successful officers” as similar to “women in general” on two items (generous and creative). Women rated “successful officers” and “women in general” as similar on four additional items (humane, helpful, exhibitionist, understanding, generous). These results suggest that female cadets are more likely to use the same terms in describing women, men, and successful officers.

Discussion

This study produced several interesting outcomes including:

1. men’s perceptions of the lack of similarities between women and leaders,
2. support of previous findings that indicated that women recognize similarities between women and leaders,
3. greater experience with being led by female commanders did not change the men’s masculine stereotype of successful leaders,
4. the surprising finding that senior-level male cadets possess stronger masculine stereotypes of successful officers than do freshman male cadets,
5. men, regardless of their level of performance, maintained a masculine stereotype of leaders.

Successful female leaders, however, perceived successful officers as having characteristics commonly associated with both men and women, whereas their less successful counterparts perceived successful military leaders as possessing characteristics, attitudes, and temperaments more commonly ascribed to women in general than to men in general.

Despite the changes in the military to include admission of women into military academies and some combat roles, the underlying stereotypical perceptions of leadership in the military are masculine in nature. Results of the study indicate that men at the USAFA share attitudes similar to those of civilian managers 25 years ago (Schein, 1973, 1975).

Results of this study also suggest that increased experience and seniority in the military environment may increase, rather than decrease, the pattern of masculine stereotyping of the officer/leader role. Unlike previous civilian findings that indicated that managerial experience modified stereotypical perceptions, this cadet military sample’s responses indicate that experience with leadership and seniority at the military academy maintained or increased masculine stereotyping. Researchers (Karau & Eagly, 1999; Powell, 1993; Schein, 1973) have suggested that increased exposure to women as managers helps modify perceptions of masculine gender role stereotypes of the leadership role. However, examination of male cadets’ perceptions by exposure to or experience with women leaders indicated that experience was not related to the degree of masculine stereotyping. In this study, men who had worked for two or more female commanders perceived requisite officer characteristics to be stereotypically male-oriented, which was the same view reported by men who had never worked for a female commander. These results, as well as those from studies of other military populations (Boldry et al., 2001; Youngman, 2001), may indicate that the culture of the military academy is so strongly masculine that senior cadets who have been involved in this culture perceive successful leadership characteristics to be masculine to a greater extent than do junior cadets who have not been as long exposed to the academy culture. On the other hand, the cross-sectional study design poses a limitation in interpreting the results related to the seniority variable, because it is possible that the perceptions of the junior and senior cadet classes were somehow inherently different. A longitudinal study design would better answer questions regarding the effect of seniority on cadets’ perceptions of gender role and leadership.

Results of this study did not confirm the hypothesis that more successful female cadets would gender-type the officer job as requiring primarily feminine characteristics, in light of the self-serving biases demonstrated in previous studies of leadership prototyping (Dunning et al., 1991). Rather, similar to the Schein et al. (1989) study of management students, successful female cadets perceived women and men both to possess characteristics necessary for military leadership success. If this gender-free view of requisite officer characteristics continues as these women advance as officers, we might expect them to treat men and women more equitably, on the basis of performance criteria rather than gender role expectations, in their selection and promotion decision processes.

In contrast to the more successful female cadets in this study, female cadets with lower military performance averages rated the successful officer as possessing traits associated with women. These results may be due to the female cadets’ own self-descriptive attributes; that is, if the less successful female cadets describe themselves in more feminine terms than do their more successful counterparts, they may be weighting these feminine traits more strongly when describing the successful officer, as predicted by the self-serving biases found in leadership prototype studies (Dunning et al., 1991). It is also possible that the perceptions of the less successful female cadets (that successful officers possess feminine traits) are partly responsible for lower performance ratings in an environment where most of their peers have different perceptions. Along these lines, it is interesting to note that in the study of military students cited earlier by Boldry et al. (2001), female military trainees rated themselves as more feminine and less masculine than male trainees rated themselves, yet there were no differences between the men’s and women’s self-ratings of their motivation, leadership, and character. Thus, further research is needed to clarify cadets’ self-assessments of their own leadership potential and gender roles in relation to their gender role stereotypes of military officers.

Limitations and Future Directions

In addition to the cross-sectional nature of the study, limitations of this study include the use of a questionnaire to measure abstract perceptions of leadership rather than actual leader behavior and performance outcomes among men and women leaders. Laboratory studies, compared with field studies of actual behavior, typically exaggerate the salience of gender differences in perceptions of leadership (Eagly & Johnson, 1990; Klenke, 1996).

Current research on gender issues in leadership as well as the study’s design limitations suggest a variety of important avenues for future research. One set of variables for future research concerns individual difference variables such as gender role orientation and attitudes toward leadership. Studies of gender role identity in relation to managerial aspirations have suggested that both women and men with managerial aspirations score high on scales that measure masculinity (Marshall & Wijting, 1980). These scales include many of the traits that are perceived as necessary for leadership, such as dominance, responsibility, achievement, and self-assurance (Kolb, 1996, 1999; Powell, 1988; Steinberg & Shapiro, 1982). In addition, it has been suggested that “people choose careers that are consistent with their self-image” (Korman, 1970, p. 32). Because the military is traditionally recognized as a masculine institution, it is likely that the military academies would attract women and men with more traditional views, including expectations that effective leaders are masculine. Feminine applicants or those who do not expect “masculine” leadership styles may be less likely to apply to the USAFA. Consistent with this interpretation is the finding that military trainees in the study by Boldry et al. (2001) rated the ideal cadet as low in femininity.

Another question for future study is whether expectations regarding masculine leadership characteristics have positive or negative, consistent or differential effects for male and female cadets. Role congruency theory suggests that the “male managerial model” poses barriers for women who aspire to leadership positions, because if women engage in expected feminine behaviors they may be seen as not able to behave in a way congruent with appropriate leadership behaviors (Eagly et al., 1995). However, some studies of the evaluation of leader behaviors have suggested that women who are successful in stereotypically masculine environments may be evaluated more favorably than their male counterparts (Eagly & Johnson, 1990; Eagly, Mladinic, & Otto, 1991; Nieva & Gutek, 1981; Thompson, 2000). Because masculine characteristics are not directly linked to sex (Kolb, 1999), it is possible that women who apply, are accepted, and who opt to stay in the military academy are perceived favorably by their peers and leaders because they are able to perform successfully in roles that differ from those considered traditionally feminine. Likewise, motivation to control prejudice has been found to be an important variable in the evaluation of women’s leadership (Ziegert & Hanges, 2002). Thus, studies of actual gender differences in cadet leader behavior, and differences in the evaluations of these behaviors, are warranted. Similarly, studies in which perceptions of the extent of, and conditions correlated with, military women’s leadership effectiveness are measured directly would help to clarify the extent to which gender role stereotypes serve as barriers against women’s entry into and success in military leadership positions.

A more difficult question to research, but one that is potentially quite important, is the underlying validity of assumptions regarding the masculine nature of successful leadership in the military organization to achieve the organization’s objectives. It is widely recognized that the military has had an enduring culture that is primarily “masculine” in nature. It could be argued that the American military has been largely successful in achieving national and international military objectives throughout America’s history. On the other hand, just as civilian organizations are characterized by unprecedented change in their environments and required operations, so is the military of the twenty-first century characterized by unprecedented change in its range of missions and operations.

The current war on terrorism, as well as a variety of peacekeeping missions in which the U.S. military is currently engaged across the globe, have required the unprecedented increases in use of U.S. special operations forces (Burns, 2002). Unlike the conventional military, these special forces are organized into much smaller, more flexible team units, trained to conduct the gamut of unconventional missions including humanitarian, special reconnaissance, foreign internal defense, and guerilla warfare missions (Burns, 2002). It seems likely that the increased demands placed on the military by the global war on terrorism, as well as increased number and range of unconventional missions, will also lead to increased demands on military leadership. A host of current researchers and management practitioners have posited the need for less gender-typed approaches to leadership (e.g., Rosener, 1990), and the benefits of “balanced,” more wide-ranging, and flexible leadership orientations to meet changing demands on organizations (Thompson, 2000). These exhortations seem to apply to both military and civilian organizations in today’s fast-changing world. Transformational and charismatic approaches to leadership, which are seen as less incongruent with women’s roles than traditional masculine leadership approaches, might be especially appropriate during times of change and crisis, such as the war on terrorism (Carless, 1998; Conger & Kanungo, 1998; Rosener, 1990; Thompson, 2000; Yoder, 2001).

Practical Implications

Suggestions for overcoming possible leadership barriers posed by gender role stereotypes in masculine organizational cultures include individual, organizational, and contextual strategies (Bajdo & Dickson, 2002; Heilman, 2001; Heilman et al., 1989; Loden, 1985; Ragins, Townsend, & Matthis, 1998; Yoder, 2001). At the individual level, women are encouraged to use communal gender-congruent leadership styles, build trust before attempting to influence their group, and increase the salience of their task-related competencies to decrease status differences between themselves and their followers and also to increase their perceived status. At the military academy, academic as well as military leadership courses could specifically include research on the gendered context of leadership and the equifinality of various leadership styles.

Organizational strategies for enhancing women’s leadership opportunities include making clear the measurement of and criteria for leadership performance and promotion decisions (Ragins et al., 1998: Ragins & Sundstrom, 1989; Yoder, 2001). Legitimating women selected as leaders has also been shown to increase their effectiveness (Yoder, Schleicher, & McDonald, 1998). At the military academy, explicit publication of the skills and competency criteria used to select cadet leaders would benefit women and men who are selected for leadership positions.

Contextual strategies for enhancing women’s leadership effectiveness focus primarily on minimizing the negative effects of tokenism (Kanter, 1977: Yoder, 2001). To this end, several researchers have suggested ensuring that work groups are composed of at least 35% women to make them congenial for women’s leadership and to reduce gender role stereotyping (Tolbert, Simons, Andrews, & Rhee, 1995; Yoder, 2001). At the military academy, the small percentage of female cadets is evenly distributed among the cadet squadrons so that female cadets usually have token status (less than 15%) in all the squadrons. In light of research on the negative effects of tokenism, the costs and benefits of increasing the percentage of female cadets in some squadrons (and thus decreasing the percentages in others) should be assessed.

Table I. Aalyses of Variance of Mean Item Ratings and Intraclass Coefficients

Table II. Intraclass Correlation Coefficients by Gender, Seniority/Class Year, (a) Female Leadership, (a) and Performance (b)

Table III. Item Mean Scores Rated Significantly Different Between Surveys by Gender

Table IV: Item Mean Scores Rated Significantly Different Between Surveys by Gender

Acknowledgments

This research was supported by the United States Air Force Academy (USAFA), Department of Behavioral Sciences (DFBL) and the USAFA Corbin Counsel. The authors thank two anonymous reviewers for their comments on an earlier draft of this article.

• Portions of this paper were presented at the 108th Annual Convention of the American Psychological Association, Washington, DC, August 2000.
• Schein’s original Descriptive Index, which requested ratings of “successful middle managers,” was modified, with permission, to “successful officers” to meet the needs of this study.
• The authors thank the Editor for this alternative explanation.

References

Arkins, W., & Dobrofsky, L. R. (1978). Military socialization and masculinity. Journal of Social Issues, 34, 151-168.

Bajdo, L. M., & Dickson, M. W. (2002). Leadership preferences, organizational culture, and women’s advancement in organizations. Paper presented at the meeting of the Society for Industrial and Organizational Psychology, Toronto, Canada.

Baron, A. (1984). The achieving woman manager: So where are the rewards? Business Quarterly, 49, 70-73.

Bartol, K. M., & Butterfield, D. A. (1976). Sex effects in evaluating leaders. Journal of Applied Psychology, 61, 446-454.

Bass, B. M., Krusell, J., & Alexander, R. A. (1971). Mate managers’ attitudes toward working women. American Behavioral Scientist, 15, 221-236.

Boldry, J., Wood, W., & Kashy, D. A. (2001). Gender stereotypes and the evaluation of men and women in military training. Journal of Social Issues, 57, 689-705.

Bowman, G. W., Worthy, N. B., & Greyser, S. A. (1965). Are women executives people? Harvard Business Review, 43, 15-28, 164-178.

Brenner, O. C., Tomkiewicz, J., & Schein, V. E. (1989). The relationship between sex role stereotypes and requisite management characteristics revisited. Academy of Management Journal, 32, 662-669.

Broverman, I. K., Broverman, D. M., Clarkson, F. E., Rosenkrantz, P., & Vogel, S. R. (1970). Sex-role stereotypes and clinical judgments of mental health. Journal of Consulting and Clinical Psychology, 34, 1-7.

Broverman, I. K., Vogel, S. R., Broverman, D. M., Clarkson, F. E., & Rosenkrantz, P. S. (1972). Sex role stereotypes: A current appraisal. Journal of Social Issues, 28, 59-78.

Burns, R. (2002). Special operations role grows. The Fayetteville Observer, p. 11A.

Carless, S. A. (1998). Gender differences in transformational leadership: An examination of superior, leader, and subordinate perspectives. Sex Roles, 39, 887-902.

Colaw, R. S. (Ed.). (1996). Women in motion. Available from HQ/USAFA/PA, 2304 Cadet Drive. Suite 320, USAF Academy, CO 80840-5016.

Conger, J. A., & Kanungo, R. N. (1998). Charismatic leadership in organizations. Thousand Oaks, CA: Sage.

Devilbiss, M. S. (1990). Women and military service: A history, analysis, and overview of key issues. Maxwell Air Force Base, AL: Air University Press.

Dobbins, G. H., & Platz, S. J. (1986). Sex differences in leadership: How real are they? Academy of Management Review, 11, 118-127.

Dodge, K. A., Gilroy, F. D., & Fenzel, L. M. (1995). Requisite management characteristics revisited: Two decades later. Journal of Social Behavior and Personality, 10, 253-264.

Dunning, D., Perie, M., & Story, A. L. (1991). Self-serving prototypes of social categories. Journal of Personality and Social Psychology, 61, 957-968.

Eagly, A. H. (1987). Sex differences in social behavior: A social-role interpretation. Hillsdale, NJ: Erlbaum.

Eagly, A. H., & Johnson, B. T. (1990). Gender and leadership style: A meta-analysis. Psychological Bulletin, 108, 233-256.

Eagly, A. H., Karau, S. J., & Makhijani, M. G. (1995). Gender and the effectiveness of leaders: A meta-analysis. Psychological Bulletin, 117, 125-145.

Eagly, A. H., Makhijani, M. G., & Klonsky, B. G. (1992). Gender and the evalution of leaders: A meta-analysis. Psychological Bulletin, 111, 3-22.

Eagly, A. H., Mladinic, A., & Otto, S. (1991). Are women evaluated more favorably than men? An analysis of attitudes, beliefs, and emotions. Psychology of Women Quarterly, 15, 203-216.

Hays, W. L. (1963). Statistics for psychologists. New York: Holt, Rinehart, & Winston.

Heilman, M. E. (2001). Description and prescription: How gender stereotypes prevent women’s ascent up the organizational ladder. Journal of Social Issues, 57(4), 657-674.

Heilman, M., Block, C., Simon, M. C., & Martell, R. F. (1989). Has anything changed? Current characterizations of men, women, and managers. Journal of Applied Psychology, 74, 935-942.

Johnson, M. M. (1997). Women’s status in military examined. Coast Guard Academy News, pp. A3, A4.

Kanter, R. M. (1977). Men and women of the corporation. New York: Basic Books.

Karau, S. J., & Eagly, A. H. (1999). Invited reaction: Gender, social roles, and the emergence of leaders. Human Resource Development Quarterly, 10, 321-327.

Klenke, K. (1996). Women and leadership: A contextual perspective. New York: Springer-Verlag.

Kolb, J. A. (1996). A comparison of leadership behaviors and competencies in high- and average-performance teams. Communications Reports, 9, 173-183.

Kolb, J. A. (1999). The effect of gender role, attitude toward leadership, and self-confidence on leader emergence: Implications for leadership development. Human Resource Development Quarterly, 10, 305-320.

Korman, A. K. (1970) Toward an hypothesis of work behavior. Journal of Applied Psychology, 54, 31-41.

Loden, M. (1985). Feminine leadership, or how to succeed in business without being one of the boys. New York: Times Books.
Manning, L., & Wright, V. R. (2000). Women in the military: Where they stand.

Washington, DC: Women’s Research and Education Institute.

Marshall, S. J. & Wijting, J. P. (1980). Relationships of achievement motivation and sex-role identity to college women’s career orientation. Journal of Vocational Behavior, 16, 299-311.

Nieva, V. F., & Gutek, B. A. (1981). Women and work. New York: Praeger.

Norris, J. M., & Wylie, A. M. (1995). Gender stereotyping of the managerial role among students in Canada and the United States. Group and Organization Management, 20, 167-181.

Powell, G. N. (1990). One more time: Do female and male managers differ? Academy of Management Executive, 4, 68-75.

Powell, G. N. (1993). Women and men in management (2nd ed.). Newbury Park, CA: Sage.

Ragins, B. R., & Sundstrom, E. (1989). Gender and power in organizations: A longitudinal perspective. Psychological Bulletin, 105, 51-88.

Ragins, B. R., Townsend, B., & Matthis, M. (1998). Gender gap in the executive suite: CEOs and female executives report on breaking the glass ceiling. Academy of Management Executive, 12, 28-42.

Rosen, B., & Jerdee, T. H. (1973). The influence of sex-role stereotypes of evaluations of male and female supervisory behavior. Journal of Applied Psychology, 57, 44-48.

Rosen, B., & Jerdee, T. H. (1974a). Effects of applicant’s sex and difficulty of job on evaluations of candidates for managerial positions. Journal of Applied Psychology, 59, 511-512.

Rosen, B., & Jerdee, T. H. (1974b). The influence of sex-role stereotypes in personnel decisions. Journal of Applied Psychology, 59, 9-14.

Rosen, B., & Jerdee, T. H. (1978). Perceived sex differences in managerially relevant characteristics. Sex Roles, 4, 837-843.

Rosener, J. B. (1990). Ways women lead. Harvard Business Review, 68, 119-126.

Rudman, L. A., & Kilianski, S. E. (2000). Implicit and explicit attitudes toward female authority. Personality and Social Psychology Bulletin, 26, 1315-1328.

Russell, J. E. A., Rush, M. C., & Herd, A. M. (1988). An exploration of women’s expectations of effective male and female leadership. Sex Roles, 18, 279-287.

Schein, V. E. (1973). The relationship between sex role stereotypes and requisite management characteristics. Journal of Applied Psychology, 57, 95-100.

Schein, V. E. (1975). The relationship between sex role stereotypes and requisite management characteristics among female managers. Journal of Applied Psychology, 60, 340-344.

Schein, V. E. (1978). Sex role stereotyping, ability, and performance: Prior research and new directions. Personnel Psychology, 31, 259-268.

Schein, V. E., Mueller, R., & Jacobson, C. (1989). The relationship between sex role stereotypes and requisite management characteristics among college students. Sex Roles, 20, 103-110.

Steinberg, R., & Shapiro, S. (1982). Sex differences in personality traits of female and male master of business administration students. Journal of Applied Psychology, 67, 306-310.

Thompson, M. D. (2000). Gender, leadership orientation, and effectiveness: Testing the theoretical models of Bowman and Deal and Quinn. Sex Roles, 42, 969-992.

Tolbert, P. S., Simons, T. Andrews, A., & Rhee, J. (1995). The effects of gender composition in academic departments on faculty turnover. Industrial and Labor Relations Review, 48, 562-579.

Versaw, P. E. (1997, March). Greetings from the Superintendent of the U.S. Coast Guard Academy. Leadership in a Gender-Diverse Military. Available from the U.S. Coast Guard Academy, Public Affairs Office, New London, CT 06320-4195.

Yoder, J. D. (2001). Making leadership work more effectively for women. Journal of Social Issues, 57, 815-828.

Yoder, J. D., Schleicher, T. L., & McDonald, T. W. (1998). Empowering token women leaders: The importance of organizationally legitimated credibility. Psychology of Women Quarterly, 22, 209-222.

Youngman, J. (2001). Women in the military: The struggle to lead. In C. B. Costello & A. J. Stone (Eds.), The American woman 2001-2002: Getting to the top (pp. 139-168). New York: Norton.

Ziegert, J. C., & Hanges, P. J. (2002, April). Evaluation of female leaders: The role of attitudes and motivation. Paper presented at the meeting of the Society for Industrial and Organizational Psychology, Toronto, Canada.

Lisa A. Boyce (2,4) and Ann M. Herd (3) (2) Department of Psychology, George Mason University, Fairfax, Virginia. (3) Gettysburg College. (4) To whom correspondence should be addressed at Department of Psychology, MSN 3F5, Geroge Mason University, Fairfax, Virginia 22030-4444; e-mail: [email protected].

Brain Sex: How the Media Report and Distort Brain Research

Testosterone gives men a particular advantage in that it is focusing and galvanizing a brain that is already, by its very structure, more focused than the female. Remember that the male brain is a tidier affair, each function in its special place… Biology, then, every bit as much as social conditioning, militates against a strongly feminine role in areas traditionally regarded as male preserves.

In this quote from Brain Sex: The Real Difference Between Men & Women, Moir and Jessel (1989, P. 96) explain that women should be excluded from certain “male preserves” because their brains are different. Brain Sex illustrates how reports of scientific research first show differences between groups and then, with no explanation, equate difference with deficit or inferiority, a ready-made explanation for limiting the opportunities of girls and women. This is a successful strategy because when one biological myth about biological essentialism becomes discredited, a new one can be quickly found to replace it. Current arguments for essential differences between the sexes (and, by implication, the inferiority of females) can be found in reports of brain research. What is troubling about current discoveries of differences between male and female brains is not the research itself, but the way this information is reported, distorted, and widely disseminated by the media and then used as a justification for discrimination.

In the late 1990’s, newspapers and magazines informed the public that men have four billion more brain cells than women (Beagle, 1997; Vazsonyi, 1997), that men are more intelligent because they have larger brains (Siddiqui,1996), that women’s brains shrink during pregnancy (Abraham, 1997; Reuter, 1997), that women talk more because the areas of the brain that control language are larger in women than in men (Hall, 1997), and that women become depressed because of their brains (Leutwyler, 1995, News Service Reports, 1997). Other articles tell us that because of their brains, the sexes are at war (Beagle, 1997), can’t communicate, and are “worlds apart” (Bower, 1995).

In this paper, I explore three closely related issues related to how the media influence public attitudes. I first examine how newspapers, magazines, and trade books simplify, exaggerate, and sometimes misrepresent research findings about the brain in ways that suggest that differences between males and females are inherent, categorical, and unchangeable. I then show how some writers begin by discussing difference, but then quickly redefine any difference as deficiency and as an excuse for limiting vocational, political and educational opportunities for women. Finally, I argue that feminists should avoid the “difference debate” altogether and suggest that the important issue is equal opportunity, not equal ability.

Gender Polarization, Biological Determinism, and Androcentrism

In her book, The Lenses of Gender, Sandra Bem (1993) shows how biological essentialism has long been used to explain why women cannot perform certain tasks and activities as well as men. Citing Hippocrates, early scientists accounted for sex differences on the basis of complexions, the balance of the qualities hot, cold, moist, and dry. At that time men were granted more rights because they were judged to be superior: men had greater “heat” than women, which allowed them to purify their souls (Cadden, 1993: 171). The science has changed, but the arguments have not. In 1970, in response to a plea by Representative Patsy Mink that women’s rights deserved the highest priority at the Democratic party’s Committee on National Priorities, physician and committee member Edgar Berman responded that “raging hormonal influences” caused by the menstrual cycle and menopause should exclude women from executive responsibility (Jamieson, 1995:53). Arguments from biological essentialism are apparently still taken seriously ( as in Rushton, 1995).

Those who assume biological essentialism often use gender polarization to divide humanity into two mutually exclusive classes. As Bem (p.2) defines it, gender polarization establishes “a cultural connection… between sex and virtually every other aspect of human experience,” including those that have nothing to do with sex. Scott (1988:45) describes how gender polarization works:

In effect, the duality this opposition creates draws one line of difference, invests it with biological explanations, and then treats each side of the opposition as a unitary phenomenon. Everything in each category (male/female) is assumed to be the same; hence, differences within either category are suppressed.

Underlying both biological essentialism and gender polarization, androcentrism, stipulates males to be the norm, defining females and female behavior in reference to that norm.

Although the details have differed in different eras, biological determinism, gender polarization and androcentrism have a long history of excluding women and racial minorities from education and positions of public responsibility. For example, in 1873, Edward Clark’s Sex in Education used the concept of “vital force” to argue against the education of women. Since the human nervous system was believed to have a fixed amount of vital force, any energy spent in the development of a woman’s brain would be diverted from her reproductive organs, endangering her primary role as a mother. (Bem, 1993: 10).

The Brain and Biological Essentialism

Most people no longer believe that women’s wombs will suffer when they use their brains, but women are still being informed by researchers and the media that men’s and women’s brains are essentially different. The media continue to seek new ways to ask “How are men and women different?”

Social claims based on brain research have long been used to distance privileged groups from those judged to be less worthy. As Gould (1980:153, 1981: 52-69) shows, nineteenth-century scientists such as Morton and Broca looked for and found “objective” evidence to prove that whites had larger brains than Indians and Blacks. Morton, for example, believed he was being objective and scientific, but his unconscious biases were so powerful that they influenced his results (Gould, 1981:69). Broca accepted the “scientific truth” that “[w]omen, like it or not, had smaller brains than men and therefore, could not equal them in intelligence.” (Gould, 1980:153) Not only were women’s brains smaller (since smaller people usually have smaller brains), they were assumed to be deficient.

Despite many scientific counterarguments, problematic research about the relative size of male, female, and African-American brains is still being done. In his 1995 book, Race, Evolution and Behavior, J. Philippe Rushton, a psychologist at the University of Western Ontario, claims that different races have different brain sizes. On the basis of statistical evidence he concludes (p. 190) that Blacks and women have smaller brains and lower mental abilities than white males. He claims that “there is a small but robust correlation between brain size and intelligence” (p. 260). It is surprising that this research continues to be funded and published, in spite of what Fausto-Sterling (1985:37) calls the “elephant problem”:

If size were the determinant of intelligence, then elephants and whales ought to be in command. Attempts to remedy this by claiming special importance for the number obtained by dividing brain size by body weight were abandoned when it was discovered that females came out “ahead” in such measurements.

The brain research by Rushton has gotten little sympathetic notice in the media, possibly because his claims are racist as well as sexist (Siddiqui, 1996). However, other research exploring differences between male and female brains, such as that of Bennett and Sally Shaywitz and their colleagues (1995: 609) at Yale University, was widely reported and discussed in the media. The Yale researchers claimed to have found “clear evidence for a sex difference in the functional organization of the brain for language.”

In one of three parts of their study, female and male subjects were asked to perform rhyming tasks. For one task, magnetic resonance imaging scans showed left lateralization of brain activity for all of the men (19 subjects) and 42% of the women ( 8 subjects), but showed bilateral activity for 58% of the women (11 subjects). The results were based on a limited number of subjects. The task indicating such “remarkable” differences was that of judging whether nonsense syllables rhymed; this task activated areas of the brain believed to be involved in phonological processing.

The other tasks in the experiment, orthographic and semantic processing, revealed no differences by sex for the 38 subjects. In other words, claims of “remarkable” and “essential” differences by sex were based on the bilateral processing of 11 women for a single rhyming task. Patterns of 42% of the women and results of the two other tasks, which showed no variation between the sexes, were minimized or ignored, both in the authors’ report in Nature and in subsequent reports in popular periodicals.

For example, Science News announced the results of the Yale experiments with the headline “Brain scan tags sexes as worlds apart” and introduced the story with the lead, “More often than they would like, men and women have trouble talking to one another” (Bower, 1995:101). Jet’s cover story “Why men and women cannot be like each other” reported that “Researchers at Yale University recently discovered that men and women will never be like each other because they use their brains differently” and concluded, “Many researchers agree that it’s no wonder the battle between the sexes rages on.” (“Why men and women cannot be like each other” 1995:15-16). Omni magazine (Phillips, 1990) reported the results this way: “Male and female brains aren’t the same. Does this mean that sexual differences are biologically determined?” Reading these articles, it is easy to forget the basis for the reported differences. One wonders how the processing of a simple rhyming task can possibly be characterized as a “battle between the sexes.”

Some of the sex and brain research reported in newspapers and magazines concerns the effects of sex hormones on the brain, primarily on the brains of cats and rats. There is little doubt that sex hormones do affect the brain, just as they do other parts of the body. The majority of researchers who investigate the hormonal effects are careful to note that the human brain is extremely complicated, and most researchers caution that, on the whole, very little is understood about how the brain works.

However, others, such as Doreen Kimura, claim that “the effects of sex hormones on brain organization result in differently wired brains in girls and boys” (1992: 119) and that different hormones result in different potentials in almost every aspect of experience. Kimura (1992:121) says, “The hormonal effects are not limited to sexual or reproductive behaviors: they appear to extend to all known behaviors in which males and females differ.” Having discovered differences in the brain, Kimura (1992:125) speculates about social roles:

The finding of consistent and, in some cases, quite substantial sex differences suggests that men and women may have different occupational interests and capabilities, independent of societal influences. I would not expect, for example, that men and women would necessarily be equally represented in activities or professions that emphasize spatial or math skills, such as engineering or physics.

Kimura reveals her assumptions about biological essentialism both in her choice of the metaphor “differently wired” and in her reservations about societal influences. Kimura looks for and finds differences between males and females, and her work is published for popular as well as scientific audiences (Holloway, 1990; Kimura, 1992; Kimura, 1985).

Research showing difference is generally published because academic books and journals have a bias towards reporting differences and against reporting negative results (Crawford, 1995, chapter 1). In discussing differences in brain laterality, Springer and Deutsch (1993:211-12) note:

Investigators are much more willing to report differences between groups (and journal editors are much more eager to accept such studies) than they are to publish negative or “no-difference” results. Critics have suggested that journals contain only the tip of the sex-differences-inlaterality-research iceberg and that the majority of studies with negative results are never published.

As Kimura (1992: 121) admits, “We strain to look for differences and, of course, tend to emphasize the few we find.”

However, to her credit, Kimura, like most researchers, acknowledges both the complexity of the problems she investigates, the emphasis on differences, and the enormous overlap found within groups:

First, biological sex itself has turned out to be much more variable and dynamic than we ever imagined. And brain-organization patterns are even more variable from person to person, and probably even within the same person at different times. Further, on most tests of cognitive ability there is enormous overlap of men and women. (Kimura, 1992: 121)
What is this “enormous overlap”? Many researchers have claimed that, on average, males have better spatial skills and females have better verbal skills. In discussing the data from Maccoby and Jacklin (1974) about male and female spatial and verbal skills, Fausto-Sterling (1985:33) explains:

Maccoby and Jacklin point out that, as with differences in verbal skills, differences in spatial skills are quite small–accounting for no more than 5 percent of the variance. Expressed another way, if one looks at the variation (from lowest to highest performance) of spatial ability in a mixed population of males and females, 5 percent of it at most can be accounted for on the basis of sex. The other 95 percent of the variation is due to individual differences that have nothing to do with being male or female (emphasis added).

Effect sizes below 5 percent are not statistically significant and are rarely emphasized in either scientific reports or news articles. In contrast to the widely reported results found by Shaywitz and his Yale colleagues, Kimura’ s (1992:120) findings were not considered news:
Effect sizes below 0.5 are generally considered small. Based on my data, for instance, there are typically no differences between the sexes on tests of vocabulary (effect size 0.02), nonverbal reasoning (0.03) and verbal reasoning (0.17).

Unfortunately, books for general audiences, such as those of Moir and Jessel (1991) and Goldberg (1977, 1993) generally ignore any similarities and the extent of overlap. For example, Moir and Jessel (1989:55) claim:

The world, in one sense, means different things to each sex. This phenomenon can be observed at such an early stage and is sometimes so obvious–that it further undermines the argument that society, rather than sexuality, conditions these inherent biases.

This is exactly the emphasis on differences that Cynthia Epstein (1988:37) cautions against:

Reports of sex differences tend to gloss over the size of difference. The titles of articles that report findings convey the impression of mutually exclusive categories rather than overlap. Thus, results tend to be perceived as based on attributes that are innate or set early in life.

Authors such as Moir and Jessel (1991:162) and Goldberg (1993, 1997) cite the research of scientists like Kimura and Shaywitz et al., distort their results, and then use them to argue that men consistently end up in more powerful positions and have higher salaries because of their biological makeup. They conclude that women, unlike men, don’t have a biological need to pursue power (Moir and Jessel, pp. 156-157). Goldberg (1993:68) summarizes:

Whatever the terminology used, the important point, and the central argument presented in this book, is that differences in the male and female neuro-endocrinological systems are such that the environmental stimulus of hierarchy, status, or a member of the other sex elicits from the male a stronger tendency to give up whatever must be given up–time, pleasure, health, physical safety, affection, relaxation–for the attainment of a higher hierarchical position, for a social role which is rewarded by greater status, and for dominance in male-female relationships.

A number of the claims that Goldberg, Moir, and Jessel make about brain difference are clearly questionable. For example, Moir and Jessel claim (p. 150) that women are biologically more suited than men to do housework because men’s brains do not do not predispose them to notice dirt. They also claim (p. 105) that because of their brains, “men are born to be more promiscuous.”
What is wrong with emphasizing biological difference?

Although some readers have no difficulty recognizing the biases in books such as Brain Sex (Moir and Jessel, 1992), The Inevitability of Patriarchy (Goldberg, 1977), and Why Men Rule (Goldberg, 1993), others seem eager to accept “scientific” arguments that males and females are essentially different. [1] Subsequently, it is a small step to the conclusion that males and females should have different responsibilities and opportunities, as writers such as Kimura, Joseph, Goldberg, and Moir and Jessel all propose. The strategy is usually to note average differences between large groups of males and females, ignore large areas of overlap, assume that all members of each category are the same, find biological or evolutionary explanations, and then generalize to social issues, urging women to avoid traditional male preserves and privileges.

Ritchie (1975) reports how biological essentialism and gender polarization have historically been used to exclude women from voting rights and public office. For example, she discusses the Towns Improvement (Ireland) Act (1854), in which the question was whether to allow women to vote. In overturning a decision that would have allowed women to hold public office, the judge based his negative decision partly on biology. He admitted that there had been cases in which women had been capable of holding office, but stated:
Having regard to every one of the reasons of the Common Law, the subordination of sex, the inferiority of bodily ability, and the mental inferiority, in the sense explained [less education and mental training], as well as to decency and decorum, I am not sorry that I am able, on the best consideration I have been able to give the case, to come to the conclusion that this judgement ought to be reversed (Ritchie, 1975: 693).

Reasoning based on biological essentialism and fundamental differences is apparently still an issue for some judges. In 1966, Mississippi’s Supreme Court justified the exclusion of women from juries “so that they may continue their service as mothers, wives, and homemakers, and also to protect them.. from the filth, obscenity and noxious atmosphere that so often pervades a courtroom during a jury trial” (State v. Hall 187 So. 2nd 861,863, Miss., quoted in Jamieson, 1995:101).

More recent lawsuits in which male and female differences have been issues include the 1973 EEOC v. Sears Roebuck & Co. (Jamieson, 1995:112), a 1992 class action lawsuit against Lucky Stores (Jamieson, 1995:114), and the 1993 Harris v. Forklift Systems Inc. lawsuit (Jamieson, 1995:119). The issue of “fundamental differences” is still being debated in court cases in the U.S.

Presumably, new studies showing how the brains of males and females differ could be used as evidence of fundamental differences for future court cases. The emergence of new “facts” has historically signaled new approaches to gender polarization. As Hess (1990: 81) reminds us,
For two millennia, ‘impartial experts’ have given us such trenchant insights as the fact that women lack sufficient heat to boil the blood and purify the soul, that their heads are too small, their wombs too big, their hormones too debilitating, that they think with their hearts or the wrong side of the brain.

How is the distortion of research in the media related to what happens in the courts, board rooms, or educational institutions? Most of the population hear about research findings only when the results are reported in newspapers and magazines and if judges, juries, legislators, and policymakers read articles reporting scientific “proofs” of biological differences, such information is likely to influence their judgments. Good science is, after all, based on facts.

The Perception of Facts

The “facts” that the media present on any subject, however, are often those that reinforce prevailing ideologies. Romero (1986:72) discusses how journalists chose and present certain facts in news stories, but ignore others:

Regardless of the events he attends, his flexibility in characterizing them remains. From the exercise of this freedom, and from such influences on his decisions as habit, ideology, and his understanding of what readers want, come the “facts” that we conventionally accept as such in the press.

In addition, even when facts contradictory to general beliefs are reported, readers may ignore them. People often ignore any information that conflicts with their beliefs or preconceived ideas. Schaff (1984) calls this tendency to disregard any facts and opinions that conflict with prevailing beliefs cognitive dissonance (Festinger, 1957) and offers a summary of some of the research about it. Schaff (1984:96) explains that:

Schaff claims that facts inconsistent with previous beliefs and biases can be intellectually acknowledged, but “emotionally blocked,” overlooked, and not assimilated. According to Schaff (p. 94):

In conflict situations, if the opinions and attitudes (in the sense of readiness to act) of a human being concerning certain issues, primarily social ones, are at variance with the realities of life and if neither those realities can be brought into agreement with the said opinions nor those opinions modified without ruining the ideology of their carrier, then a psychological defense mechanism is put into operation to make one’s mind immune against inconvenient information. [s]uch situations are in a sense schizophrenic because a given person at the same time knows something and does not know it, which, while it must appear strange, often does occur in practice. This in turn breeds specific forms of dogmatism and the phenomenon of ‘closed mind,’ deaf to all arguments.

If the media repeatedly tell the public that males and females are essentially different because they have different types of brains, and if the public seems eager to hear this message, how is it possible to introduce ideas that would encourage people to reexamine their beliefs that males and females are essentially different? Even if a large amount of data can be collected to show that the claims of books such as Brain Sex are not accurate, how does one effectively get this message across?

Refraining the Issues

Nobody denies that there are biological differences between the majority of males and the majority of females. Given the obvious differences, why should feminists be concerned with the question of whether or not male and female brains are essentially different? How does one admit some differences, yet challenge the questionable and exaggerated claims of books and articles with titles such as The Inevitability of Patriarchy (Goldberg, 1977) and “Why men and women cannot be like each other”? If many people accept biological essentialism, and if they also believe that women’s opportunities should be limited because of biological differences, how can these issues be reframed so that reasonable people will actually listen?

A scholarly approach to challenging books such as Brain Sex is to cite researchers who remind us of the wide range of variability (Kimura 1992) and caution that “we do not at present understand the cognitive function of any brain area”( Efron, 1990:27). Even a cursory view of the scientific literature makes it clear that the types of social claims made by Moir and Jessel (1992) are unjustified and not supported by the evidence they cite.

Strong counterarguments against biological essentialism have been made by a number of well-known writers, including Fausto-Sterling (1985, 1993) and Gould (1980, 1981). These scholarly arguments are well-reasoned and supported by extensive factual evidence. Unfortunately, those who find such facts convincing are probably those already biased to share the authors’ viewpoints. Advocates of separate spheres for males and females will probably not find them compelling.

Another scholarly approach for questioning popular beliefs about essential differences between male and female brains is a consideration of the many variables and explanations for reported brain differences other than sex. The fact that left-handed and right-handed people process information in different parts of the brain (Bryden, 1982a, 1982b) is widely accepted, although as Harshman and Hampson (1987:83) note, even these differences have been “surprisingly difficult to demonstrate” and “previous findings have been inconsistent and often contradictory.”

Some researchers have associated different types of cognitive functions with different brain hemispheres and have identified left brain activity with linear analytic thinking and right brain activity with the unconscious and with creativity (Corballis 1983, Ehrenwald 1984, Springer and Deutsch 1993). Other researchers have shown that different temperaments extroverts and introverts or optimists and pessimists have different patterns of brain activity (Robinson, 1996, Chapter 6).

The list can be extended. Ehrenwald (1984:10) discusses research on professional groups predicted to have different cognitive styles and reports that “alpha measurements show that the people from business, law, and accounting professions differ from individuals in creative professions in the way their hemispheres process cognitive tasks.” Gannett (n.d.), explores cognitive distinctions for writers finding significant differences between the lateralization of critics and writers of fiction and poetry. Obler (1981) reports different hemisphere participation and dominance for second-language learners of languages, depending on whether they read from left to right (as with English) or right to left (as with Hebrew). Tadanobu Tsunoda discovered that Japanese speakers and Westerners usually process vowels differently (Merrill, 1981: 74).

To summarize, differences in hemisphere processing have been discovered in response to the following questions, all of which presuppose (as well as seek) differences:

• How do the patterns of brain activity of creative and analytical thinkers differ?
• How do the patterns of brain activity of left-handed and right handed people differ?
• How do the patterns of brain activity of optimistic and pessimistic people differ?
• How do the patterns of brain activity of people from different countries differ?
• How do the patterns of brain activity of people who speak and read different languages differ?
• How do the patterns of brain activity of males and females differ?

Not surprisingly, when researchers ask these questions, they discover answers showing differences. When those answers reinforce popular ideas in a culture, they are widely reported and discussed in the media. However, when experimenters find few or no differences, the results are rarely published, and variation within groups tends to be ignored or minimized.

For example, when originally reported, Tsunoda’s studies showing differences between the ways that Americans and Japanese process vowels were far more interesting to the Japanese than to Americans. Tsunoda’s research was reported in Science Digest by Merrill (1981), who skeptically observed, “Intriguing as Tsunoda’s findings are, they have not yet been replicated…. Some scientists find his theories astonishingly elegant, while others are waiting for more data to come in” (Merrill 1981: 75). Tsunoda’s work was of such interest to many Japanese, however, that “[d]espite its highly technical language, the book became a best-seller in Japan” (Merrill 1981: 74), possibly because it reinforced the belief of many Japanese that they are a unique people.

In the U.S., there is a widespread belief that males and females are essentially different, and the current use of the mind as a “computer” reveals how widespread that belief is. This metaphor, particularly the use of references to “hard wired” and “hardware/software”, also reinforces the belief that physical differences in the brain are the result of genetic rather than environmental influences, despite proven environmental factors on brain development such as touch, nutrition, and stimulation (Levine, 1973; Merzenich, 1990; Blakeslee, 1995; Perry et. al, 1995; Shore, 1997). The computer metaphor of hardware/software is misleading because it conceals the importance of socialization on brain development. Unlike computers, brains are organic, and, like other organic entities, brains interact with and respond to their environments.

Stephen Pinker (1998: Weekend Page T022) is only one of many linguists who use the metaphor of the mind as computer. He writes, “The computational theory of mind expressed by the mathematician Alan Turing, among others, is one of the great ideas in intellectual history.” Recently I heard a radio talk show in which a scientist doing brain research repeatedly used the term “hard wired.” This particular scholar, like Pinker and Kimura, seems to assume that male-female differences are biologically determined because they are part of the “hardware” rather than part of the “software.”

However, many studies have established a clear link between environmental enrichment or impoverishment and brain development, including the work on monkeys (Greenough, 1984), rats (Diamond, 1984) and traumatized children (Perry, 1995; Perry et al, 1995; Perry and Pollard, 1997). Orphans who are deprived of human touch later have alarmingly smaller brains and lower intelligence than comparable infants who had been given “normal” nurturing (Blakeslee, 1995). As researchers such as Perry and his colleagues have noted, the effect of the environment on young children is profound, both in the emotional and cognitive domains.

These findings strongly suggest that when early life neglect is characterized by decreased sensory input (e.g., relative poverty of words, touch and social interactions) it will have a similar effect on humans as it does in other mammalian species. Sensory deprivation has been demonstrated to alter the physical growth and organization of the brain in animals (Perry and Pollard, 1997).

Well-documented studies of orphans and violent children establishes a clear relationship between early childhood abuse and atypical brain development. It is impossible to attribute the atypical characteristics of the brains of neglected and abused children to “hard wiring.” Therefore, why are the environmental effects of proven differences in behaviors towards male and female children not acknowledged?

Many people are unaware of the fact that sex differences in the brain believed to be “hard wired” vary from culture to culture. For example, there are no sex-related differences in spatial abilities in Eskimo males and females, possibly because Eskimo girls are allowed considerable autonomy. In contrast, there are marked differences between Temne males and females in a culture where the roles of females are circumscribed (Fausto-Sterling, 35). However, in spite of the evidence demonstrating that males and females are treated differently from birth in most cultures, and in spite of extensive evidence that information processing differs from culture to culture, the belief in essential “hard-wired” male-female differences persists.

An awareness that the metaphor “the brain as computer” can encourage biological essentialism may make some people more cautious about using brain research as a basis for social policy, but a more serious problem is the original flawed question, “What is the difference between males and females (and their brains)?” The question, “How do male and female brains differ?” presupposes that they do, and it is important not to automatically accept this presupposition. The question itself is part of the problem.

An even more serious problem is the rhetorical strategy used by those who advocate limiting opportunities for girls and women. This strategy is simply to show difference, redefine difference as deficit or inferiority, and then argue that females should be excluded because of their biological makeup.

The best strategy for feminists, in my opinion, is simply to ignore the question, “How do male and female brains differ,” and to insist on a clear distinction between equal ability (absence of difference) and equal opportunity (absence of discrimination). Those who argue for biological essentialism often conflate the two.

For example, Nicholas Wade (1994:32) notes that in higher math, “the topmost ranks are thronged with male minds,” He says, “Some feminist ideologues assert that all minds are created equal and women would be just as good at math if they weren’t discouraged in school.” He discounts the effect of social bias and cites an expert who “concludes that boys’ superiority at math is mostly innate.” On the basis of statistical differences, Wade proposes that boys and girls should be educated differently; specifically, girls should be discouraged from entering fields for which they are not biologically suited such as physics and engineering.

Androcentric books like Brain Sex also confuse equal opportunity and identical ability. Moir and Jessel (1991:6) also urge women to accept their “biological limits”:

If women have reason to rage, it is not because science has set at naught their hard-won struggle towards equality; their wrath should rather be directed at those who have sought to misdirect and deny them their very essence. Many women in the last thirty or forty years have been brought up to believe that they are, or should be, “as good as the next man,” and in the process they have endured acute and unnecessary pain, frustration, and disappointment. They were led to believe that once they had shaken off the shackles of male prejudice and oppression–the supposed source of their second-class status-the gates of the promised land of equal achievement would be thrown open. [emphasis added]

It is important to note that Moir and Jessel do not simply claim that males and females are different. They claim difference but assume inferiority; to them, women are not “as good as the next man.”

By replacing the issue of equal ability (“How are man and women different”) with the issue of equal opportunity (“Why shouldn’t every individual have the same opportunity?”), it is possible for feminists to change the discourse about essential differences. Fortunately, a belief in equal opportunity is fundamental to the value system of the United States, as are beliefs in fair play and individual autonomy. Because of this, it is possible to acknowledge differences between groups, but still insist on equal opportunities for individuals. When the central question is, “Should everyone have an opportunity regardless of background?” arguments from biology or questions of how males and females are different become irrelevant. When we ignore the issue of difference, it is possible to say, “Perhaps male and female brains are, on average, different. So what?” Even when one group does on average show a statistical biological superiority in some skills, there is no justification for exclusion as long as every individua l has equal access to education, employment and political power.

When pundits such as Wade (1994) argue “that boys’ superiority at math is mostly innate,” this argument is problematic only when used as an excuse for biased behavior in the classroom. As long as equality of opportunity remains the central issue, schools still have the responsibility of educating “atypical” people, such as girls who just happen to excel at math and women who wish to become scientists.

There are always potential risks when insisting on equal opportunity and individual autonomy. Equal opportunity has been used as an argument against affirmative action. However, when this argument arises, new issues can be raised, issues such as the historical lack of opportunity and the unearned privileges of particular groups (McIntosh, 1988). Anyone who succeeds in moving the discourse from questions of whether males and females are different to issues of equal opportunity has a better chance of being heard. Changing the issue from “How are men’s and women’s brains different?” to “Why shouldn’t every individual have equal opportunity?” switches the focus from an irrelevant issue to one to one that has important social consequences, and not just for women.

The strongest barriers to opportunities for traditionally excluded groups are often not official policies, but social pressures and the constant messages that the excluded group is “different,” and, by implication, inferior. Although educational opportunities for females are expanding, the latest AAUW report shows (1998:91) that high school boys and girls still self-select into fields conventional for their gender; males continue to disproportionately enter fields dominated by technology. It is difficult to show a causal relationship between media messages about difference and the achievement of particular groups, since socialization is a result of so many different influences. Just as no single drop of water nor even a large river can create a canyon overnight, an occasional remark or news article will not discourage a young woman. Just as no single chocolate chip cookie or even an occasional indulgence cannot make someone fat, so no single teacher or advisor can steer a female away from a lucrative career in science or technology. No single media message that men’s brains are different from women’s brains will convince women that they are not biologically suited for certain roles. However, repeated messages from the media, parents, teachers, and friends can eventually convince women to limit their options, just as months of overeating can result in obesity and centuries of running water can create changes in a landscape.

Issues of biological essentialism will probably continue to be raised. However, feminists should avoid public debates about the differences between males and females (and their brains). A more relevant issue is: “How can a society that values individual achievement deny equal opportunity to any individual?” It is hard to fault Gould (1985:197) who says, “I can only view equality of opportunity as inalienable, universal, and unrelated to the biological status of individuals.”

Janet Bing teaches linguistics and women’s studies at Old Dominion University in Norfolk, Virginia, U.S.A. Her publications include Aspects of English Prosody, Grammar Guide, Rethinking Language and Gender Research (with Victoria Bergvall and Alice Freed) and articles about phonology, discourse, African languages, and language and gender.

References

Abraham, C. (1997) Pregnant women’s brains shrink, British researchers say: Report angers many women who fear job discrimination, The Ottawa Citizen: A4.

American Association of University Women (1998) Gender gaps: Where schools still fail our children. Washington, D.C.: AAUW Educational Foundation.

Beagle, B. (1997) Danes add ammo to war between sexes, Roanoke Times & World News.

Bem S.(1993) The lenses of gender: transforming the debate on sexual inequality. New Haven: Yale University Press.

Bergvall, V. (1996) Divided minds: Gender polarization in brain and language research. In Warner, N. et al. (Eds.) Gender and belief systems, Proceedings of the fourth Berkeley Women and Language Conference (pp. 11-23). Berkeley, CA: University of California.

Bing, J. M. and Bergvall, V.L. (1996) The question of questions: beyond binary thinking. In Bergvall, V.L, Bing, J.M. and Freed, A.F. Rethinking language and gender research: Theory and Practice (pp.1-30). Harlow, Essex: Addison Wesley Longman.

Blakeslee, S. (1995) In brain’s early growth, timetable may be crucial, New York Times.

Bower B. (1995) Brain scans tag sexes as worlds apart. Science News.

Bryden M.P. (1982a) Laterality: functional asymmetry in the intact brain. New York: Academic Press.

(1982b) Handedness and cerebral organization: Data from clinical and normal populations. In Ottoson, D. (ed.) Duality and unity of the brain. Wenner-Gren International Symposium Series, Vol. 47 (pp. 55-70). New York: Plenum Press.

Cadden J. (1993) Meanings of sex difference in the middle ages: medicine, science, and culture. Cambridge: Cambridge University Press.

Corballis, M.C. (1983) Human laterality. New York: Academic Press.

Crawford, M. (1995) Talking difference. London: Sage.

Efron R. (1990) The decline and fall of hemispheric specialization. Hillsdale NJ: Lawrence Erlbaum Associates.

Ehrenwald, J. (1984) Anatomy of genius. New York: Human Sciences Press.

Epstein, C. F. (1988) Deceptive distinctions: sex, gender and the social order. New Haven:Yale University Press.

Fausto-Sterling, A. (1985) Myths of gender. New York: Basic Books, Inc.

Festinger, L. (1957) A theory of cognitive dissonance. Stanford, CA: Stanford University Press. (1993.) Sex, race, brains, and calipers. Discover: 32, 35-37.

Gannett, C. L. (n.d.) Writers and appositionality. Durham, N.H.: University of New Hampshire, unpublished paper.

Goldberg, S. (1977) The inevitability of patriarchy. London: Temple Smith. (1993) Why men rule: A theory of male dominance. Chicago: Open Court Publishing.

Gould, S.J. (1980) Women’s brains. In Gould, S.J, The panda’s thumb: More reflections in natural history, pp. 152-159. New York: W.W. Norton & Co. (1981) The Mismeasure of Man. W.W. New York: Norton & Co. (1985) The flamingo’s smile: Reflections in natural history. W.W. Norton & Co., New York.

Hall, C. (1997) Women’s brains make for better communication, new study shows. The Ottawa Citizen, D10.

Harshman, R.A. and Hampson, E. (1987) Normal variation in human brain organization: Relation to handedness, sex and cognitive abilities. In D. Ottoson (ed.) Duality and Unity of the Brain. Wenner-Gren International Symposium Series, Vol. 47 (pp. 55-85). New York: Plenum Press.

Hess B. (1990) Beyond dichotomy: drawing distinctions and embracing differences. Sociological Forum, 5, 1:75-93.

Hoagland, S. L. (1980) Androcentric rhetoric in sociobiology. Women’s Studies International Quarterly 3:285-293.

Holloway, M. (1990) Profile [Doreen Kimura]: vive la difference. Scientific American (Oct.) 263:40, 42.

Jamieson, K. H. (1995) Beyond the double bind. New York: Oxford University Press.

Joseph, R. (1992) The right brain and the unconscious: Discovering the Stranger within. New York: Plenum Press.

Kimura, D. (1985) Male brain, female brain: The hidden difference, Psychology Today (Nov.): 51-58. (1992) Sex differences in the brain. Scientific American.

Leutwyler, K. (1995) Depression’s double standard: Clues emerge as to why women have higher rates of depression. Scientific American.

Levine, S. (1973) Stimulation in infancy. In Greenough, W.T. (Ed.) The nature and nurture of behavior (pp.55-61). San Francisco: W.H. Freeman and Co.

McIntosh, P. (1988) White privilege and male privilege: A personal account of coming to see correspondences through work in women’s studies. Working Paper No. 189 (Wellesley, MA: Wellesley College Center for Research on Women.)

Maccoby, E.E. and Jacklin, C.N. (1974) The psychology of sex differences. Stanford, CA: Stanford University Press.

Merzenich, M.M. (1990) Development and maintenance of cortical somatosensory representations: Functional ‘maps’ and neuroanatomical repertoires. In Barnard, K.E. and Brazelton, T.B., Touch, the foundation of experience (pp. 47-71). Madison: International Universities Press, Inc.

Merrill, S. (1981) The Japanese brain. Science Digest: 74-75.

Moir, A. and Jessel, D. (1992) Brain Sex: The real difference between men & women. New York: Delta.

News Service Reports (1997) Study shows women produce less serotonin, The Record (14 May): A12.

Obler, L.(1981) Right hemisphere participation in second language acquisition. In Diller, K. (Ed.) Individual differences & universals in language learning aptitude (pp. 53-64). Rowley, MA: Newbury House.

Perry, B.D. (in press) Incubated in terror: Neurodevelopmental factors in the ‘cycle of violence.’ In Osofshy, J.D. (Ed.) Children, Youth and Violence. New York: the Guilford Press. Web.

Perry, ED. et al. (1995) Childhood trauma, the neurobiology of adaptation and use-dependent development of the brain: How “states” become “traits.” Infant Mental Health J, 16(4): 271.291.

Perry, B.D and Pollard, D. (1997) Altered brain development following global neglect in early childhood. Society For Neuroscience: Proceedings from Annual Meeting, New Orleans. Web.

Phillips, K. (1990) Why can’t a man be more like a woman…and vice versa. Omni: 42-44, 68.

Pinker, S. (1998) A mind to love, The Guardian, Weekend Page T022 (Available through Dow Jones Interactive).

Reuter (1997) Brains can shrink during pregnancy, Montreal Gazette: B5.

Ritchie, M.E. (1975) Alice through the statutes. McGill Law Journal, 25:685-707.

Robinson, D. L. (1996). Brain, mind, and behavior. A new perspective on human nature. Westport, Conn.: Praeger.

Romero, C. (1986) The grisly truth about facts. In. Manoff, R.K and Schudson, M. (Eds.), Reading the news (pp. 38-78). New York: Pantheon.

Rushton, J. Philippe (1995) Race, evolution, and behavior. New Brunswick, NJ: Transaction Publishers.

Sanders, B., Wilson, J. R. and Vandenberg, S.G. (1982) Handedness and spatial ability. Cortex, 18., 79-90. (Quoted in Harshman and Hampson, 1987, p. 85).

Schaff, A. (1984) The pragmatic function of stereotypes. International Journal of the Sociology of Language 45: 89-100.

Scott, J. (1988) Deconstructing equality-versus-difference: or, the uses of poststructuralist theory for feminism. Feminist Studies 14,1:33-50.

Shaywitz, B. et al. (1995) Sex differences in the functional organization of the brain for language.

Shore, R. (1997) Rethinking the brain: New insights into early development. New York: Families and Work Institute.

Siddiqui, S. (1996) Rushton draws crowd but little support: Smirks, chuckles greet psychologist’s race-based theories. Montreal Gazette.

Springer, S. and Deutsch, G. (1993) Left brain, right brain. New York: W.H. Freeman and Co.

Vazsonyi, B. (1997) Mysteries of the brain, The Washington Times.

Yen, W. M. (1975) Independence of hand preference and sex-linked genetic effects on spatial performance. Perceptual and Motor Skills, 41, 311-318. (Quoted in Harshman and Hampson, 1987).

Wade, N. (1994) Method and madness: how men and women think. New York Times Magazine. Why men and women cannot be like each other (1995) Jet 87: 15-16.

Temperamentand peer acceptance in early childhood: Sex and social status differences

The purpose of this study was to investigate the relation between sex, social status and temperament in a sample of preschool-aged children. Sociometric interviews were conducted with 182 children (92 boys and 90 girls). Status groups of popular, rejected, neglected, controversial and average children were identified according to previously established criteria. Teachers rated children’s temperament Results indicated that rejected children displayed a more difficult temperament than popular children in terms of higher activity levels, higher distractibility and lower persistence. Both rejected and neglected children were rated as displaying lower adaptability and more negative mood than popular children. Boys were also rated as more active, more distractible and less persistent than girls. Results are discussed in terms of the relevance of particular temperament dimensions to successful social functioning for boys and for girls.

Consideration of the contribution that temperament might make towards children’s social competence and social status within the peer group has stemmed from the recognition of particular individual differences that appear, at least in part, to be constitutionally based and that reflect stylistic patterns of behavior. While adaptive or maladjustive outcomes are clearly not dependent on the contribution of temperament alone, there is evidence that individual differences in temperament qualities, such as activity level or approach/withdrawal, may be related to children’s social functioning and adjustment within the peer group (Farver & Branstetter, 1994; Mobley & Pullis, 1991; Stocker & Dunn, 1990).

Research with preschoolers has indicated that individual differences in temperamental characteristics may influence the adjustment children make to the preschool setting, the responses they make to their peers and the quality of their relationships with other children (Farver & Branstetter, 1994; Keogh & Burstein, 1988; Mobley & Pullis, 1991; Stocker & Dunn, 1990). In general, children with easy temperaments, defined as approachful, adaptive and positive in mood (Thomas & Chess, 1977) have been found to respond prosocially to peer distress (Farver & Branstetter, 1994), have more positive and interactive relationships with friends and peers (Keogh & Burstein, 1988; Stocker & Dunn, 1990) and be rated as behaviorally adjusted to the preschool environment in terms of cooperation and persistence (Mobley & Pullis, 1991). In contrast, children with difficult temperaments appear to have relationships that are more problematic with their peers and are more likely to exhibit socialization and behavioral problems (Fabes , Shepard, Guthrie, & Martin, 1997; Mobley & Pullis, 1991).

Although there is evidence to suggest that individual temperamental characteristics may be linked to social adjustment and to frequency of socialization problems, the relationship between temperament, sex and social status has not been explored fully with respect to preschool-aged children. While some previous research has revealed clear sex differences in the display of temperamental characteristics identified as difficult (e.g., Farver & Branstetter, 1994; Mobley & Pullis, 1991; Sanson, Prior, Smart, & Oberklaid, 1993), whether temperamental characteristics are differentially related to social status for boys and for girls is unclear. For example, there is some evidence that the temperament dimension of arousability may be negatively related to peer status for girls, whose play tends to be more sedentary than boys, yet positively related to peer status for boys, at least in early adolescence (Bukowski, Gauze, Hoza, & Newcomb, 1993). Thus, the aim of the present study was to examine sex and social status dif ferences in temperamental characteristics for preschool-aged children.

On the basis of present research it was expected that, in contrast to rejected children, popular children would exhibit fewer of those temperamental characteristics identified as difficult such as high activity levels, high distractibility, and negative mood. Although a difficult temperament may be predictive of low peer status for both boys and girls, it was expected that not only may boys be more likely to display difficult temperaments than girls, but that contextual features such as the differing interactional styles and norms for behavior that exist within boys’ groups and within girls’ groups may mediate the relationship between temperamental characteristics and social status. Given the importance of positive peer relationships for children’s concurrent and future adjustment, examination of the linkages between temperamental characteristics, sex and social status appears to be important in understanding the influence and functional significance that temperament may have with respect to behavioral indivi duality and social adjustment.

Method

Participants

The sample consisted of 182 preschool children (mean age 62.4 months, SD = 4.22) from eleven suburban, community-based preschools serving lower to upper middle class families in Queensland, Australia. The sample included 92 boys and 90 girls. The children were predominately Caucasian with only three children of Asian origin and one Aboriginal child.

Procedure

Sociometric Status Classification. In the present study, sociometric data were collected through a combination of positive nominations and a rating scale. This procedure, developed by Asher and Dodge (1986), involves the substitution of a lowest play rating score for a disliked score which is obtained if a negative nomination method is used. Prior to commencing sociometric testing, photographs were taken of all children for whom parental permission had been given to participate in the research. The use of photographs increases the reliability of the sociometric measure for preschool-aged children.

Sociometric interviews were conducted individually during the second term of the school year. Children were first asked to select photographs of the three children with whom they most liked to play (positive nomination). Selected children were given a score of 1 for each time they were nominated. Next the participants were asked to rate all the children on a 3-point scale according to how much they liked to play with them by posting their photographs into one of three boxes. Depicted on the boxes were a happy face, a neutral face, and a sad face. Children were advised that the happy face meant they liked to play with that child a lot, the neutral face that they liked to play with that child a little bit or sometimes, and the sad face that they did not like to play with that child. Children whose photographs were placed in the box with the happy face received a rating of three; in the box with the neutral face, a rating of two; and in the box with the sad face, a rating of one.

For each child the following scores were computed: (a) number of positive nominations (L score), (b) number of low play ratings (LPR score), (c) a social preference score (SP) based on subtracting the number of low play ratings (LPR) from the number of positive nominations (L), and (d) a social impact score (SI) computed by combining the number of low play ratings (LPR) and the number of positive nominations (L). These scores were converted into standardized scores for each sex within each preschool class. Using the procedure outlined by Asher and Dodge (1987), children were classified into sociometric groups as follows: popular (L score greater than 0, LPR score less than 0 and SP score greater than 0), rejected (L score less than 0, LPR score greater than 0 and SP score less than -1.0), neglected (L score less than 0, LPR score less than 0 and SI score less than -1.0), controversial (L score greater than 0, LPR score greater than 0 and SI score greater than 1.0) and average (SP score between -.05 and.05 an d SI score between -.05 and.05). Classification resulted in 26 popular children (12 boys, 14 girls), 22 rejected children (12 boys, 10 girls), 24 neglected children (13 boys, 11 girls), 11 controversial children (7 boys, 4 girls) and 33 average children (13 boys, 20 girls). All remaining children (35 boys and 30 girls) were classified as other.

Child Temperament. Teachers completed the 23 item Teacher Temperament Questionnaire (TTQ) developed by Keogh, Pulls and Cadwell (1982) which is a short form of the 64 item Thomas and Chess (1977) Teacher Temperament Questionnaire. The TTQ incorporates the eight temperamental dimensions used in the Thomas and Chess scale and describes behaviors related to specific dimensions of temperament which might typically be observed by teachers (e.g., Child sits still when a story is being told or read; When with others this child seems to be having a good time). Teachers were asked to rate on a 6-point scale, ranging from 1 (rarely) to 6 (almost always) the degree to which children displayed specific behaviors.

The authors of the TTQ identified three factors in the measure, which they labeled Task Orientation, Personal-Social Flexibility and Reactivity. Keogh and colleagues (1982) reported that the three factors demonstrated high internal consistency (Cronbach’s alpha =.94,.88 and.62, respectively). However, as the same temperament structure may not fit the data for all ages covered by the questionnaire (3 to 7 years) or for different populations of children, the use of a factor structure derived from the data of the population in question is preferable for research purposes. Therefore, a factor analysis of the items was conducted to identify the relevant temperament dimensions for this sample.

Results

Principal Axis Factor Analysis

The subscale structure of this measure was confirmed with a principal axis factor analysis of the 23 items on the TTQ. A four-factor solution with orthogonal (VARIMAX) rotation afforded the simplest, interpretable structure and explained 67% of the variance. Given the sample size, a cut-off level for factor loadings was set at.40 (Stevens, 1996). All items had a factor loading which met this criterion with no cross loading of items across factors. The factor loadings for each item, item communalities and percentage of variance accounted for by each factor are presented in Table 1.

Factor One contained nine items, accounting for 26% of the variance, and was similar to Keogh et al.’s (1982) factor of Task Orientation. Items loading on this factor reflected the dimensions of activity, persistence, and distractibility. Factor Two contained eight items, accounting for 20% of the variance, and was similar to Keogh et al.’s (1982) factor of Personal-Social Flexibility. Items loading on this factor reflected the dimensions of approach/withdrawal, positive mood, and adaptability. Factor Three contained four items, accounting for 12% of the variance, and was similar to Keogh et al.’s (1982) factor of Reactivity. Items loading on this factor reflected the dimensions of intensity and negative mood. The fourth factor, accounting for 9% of the variance consisted solely of two threshold of response items, which had loaded on Keogh et al.’s (1982) Reactivity factor. As interpretation of factors defined by only two variables is questionable, the items on the factor Threshold of Response were omitted fr om further analyses.

Although differing slightly from Keogh et al.’s (1982) three-factor structure, the items in this analysis clustered in logical and interpretable ways into three factors of Task Orientation, Personal-Social Flexibility and Reactivity. High mean scores on each factor represent a less desirable rating in terms of temperament. Thus, high scores on Task Orientation indicate ratings of low persistence, high activity and high distractibility; high scores on Personal-Social Flexibility indicate ratings of negative mood, low adaptability and low levels of approach; and high scores on Reactivity indicate ratings of negative mood and high intensity. As high scores on Task Orientation and Personal-Social Flexibility represented low levels of Task Oriented behavior and low levels of Personal-Social Flexibility respectively, these factors were renamed Task Distractibility and Personal-Social Inflexibility to more clearly reflect the direction of the ratings. Thus, low scores on all factors denoted more positive attributes and high scores more negative behaviors.

The appropriateness of this factor structure for boys and girls was assessed by separate factor analysis for each sex. The resultant factor solutions were virtually identical and reflective of the solution for the total sample. Factor scores were calculated by summing the ratings for the item defining each factor. Reliability (Cronbach’s Alpha) of the factor scores for Task Distractibility, Personal-Social Inflexibility and Reactivity was.94,.88, and.77, respectively.

Sex and Social Status Differences

A MANOVA was conducted in which sex and social status served as independent variables. Dependent measures were the factor scores for Task Distractibility, Personal-Social Inflexibility and Reactivity. Means and standard deviations related to the dependent measures are presented in Table 2. Using Wilks’ lambda statistic, significant mean effects were found for sex, F (1,170) = 8.29, p <.0005, and social status, F (5,170) = 2.38, p =.003, but not for the sex by social status interaction, F (5,170) = 1.43, p =.127. Univariate tests revealed significant sex differences along the dimensions of Task Distractibility, F (1,170) = 19.70, p <.0005, and Reactivity, F (1,170) = 9.14, p =.003. Results indicated that boys exhibited higher Task Distractibility and Reactivity than girls.

Univariate tests with respect to social status (see Table 3 for means and standard deviations) revealed significant social status differences along the dimensions of Task Distractibility, F (5,170) = 4.24, p =.001, while the dimension of Personal-Social Inflexibility approached significance, F (5,170) = 2.04, p =.077. Post hoc analyses using Duncan’s multiple range test revealed significant differences between both popular children and children classified as other and rejected and average children on the dimension of Task Distractibility. Specifically, compared to popular children and other children, rejected children and average status children were rated as exhibiting high Task Distractibility. Popular and other children were rated similarly on this dimension. There were no other significant differences between social status groups on Task Distractibility.

Post hoc comparisons with respect to Personal-Social inflexibility revealed a significant difference between popular children and both rejected and neglected children on this dimension indicating that, compared to popular children, rejected and neglected children were rated as exhibiting higher Personal-Social Inflexibility. There were no other significant differences between status groups on this dimension.

Sex and Social Status Differences on Dimensions within Factors

Although use of factors consisting of clusters of similar items reflecting more global characteristics or traits is common when analyzing questionnaire data, caution must be used when examining group differences using constellations of characteristics such as Task Distractibility. Several authors (e.g., Coie, Dodge & Kupersmidt, 1990; Hymel, Bowker, & Woody, 1993) have pointed out the difficulty of interpreting the meaning of aggregate scores comprised of distinct variables. To the extent that different characteristics within each large factor may be differently associated with sex or social status, conclusions based on analysis of factors on their own might be erroneous. Given the potential value of both approaches (see Hymel, Bowker, & Woody, 1993) initial analyses examined subgroup differences across aggregated scores based on the results of a factor analysis. Further analyses presented here considered subgroup differences in the specific dimensions underlying each of the more global factors. The purpose o f these analyses was to ascertain whether the identified sex and/or social status differences between groups were present for all aspects of the relevant dimensions. The dimensions included in these analyses were those related to factors on which significant differences were found. Specifically, Task Distractibility and Reactivity with respect to sex differences and Task Distractibility and Personal-Social Flexibility with respect to social status differences. Two-tailed t tests were used to test for differences within these dimensions.

With respect to the sex differences on the dimensions within the factors of Task Distractibility and Reactivity, two-tailed t tests revealed girls were more likely to exhibit lower Task Distractibility in terms of lower activity levels, t (180) = 3.42, p =.001, lower distractibility, t (180) = 3.26, p =.001, and higher persistence, t (180) = 2.59, p.010, while boys were more likely to show high Reactivity in terms of higher intensity, t (180) = 3.18, p =.002, but not more negative mood. Means and standard deviations with respect to these dimensions are presented in Table 4.

With respect to social status differences on the dimensions within the factor of Task Distractibility, two-tailed t tests revealed that, compared to popular children, rejected children were rated as exhibiting high Task Distractibility in terms of higher activity rates, t (47) 2.48, p =.017, higher distractibility, t (47) = 2.96, p =.005, and lower persistence, t (47) 2.42, p =.019. Average status children also differed significantly from popular children on this dimension in terms of higher activity, t (57) = 2.45, p =.017, and higher distractibility, t (57) = 3.86, p <.0005, but not lower persistence, t (57) = 1.82, p =.075. Rejected and average children also differed significantly from children classified as other. Specifically, children classified as other were rated as exhibiting lower activity rates than both rejected, t (86) = 2.00, p =.049, and average children, t (96) = 2.21, p =.030, and lower distractibility than rejected, t (86) 2.36, p =.020, and average children t (96) = 3.48, p =.001. Means and standard deviations with respect to the dimensions within Task Distractibility and Personal-Social Inflexibility are presented in Table 5.

Two-tailed t tests with respect to the dimensions within the factor of Personal-Social Inflexibility revealed that, compared to popular children, rejected children were rated as exhibiting high Personal-Social Inflexibility in terms of lower adaptability, t (47) = 2.43, p =.019, and more negative mood, t (47) = 3.10, p =.003. Neglected children also differed significantly from popular children on the dimensions within this factor in terms of lower adaptability, t (48) = 2.05, p =.046, and more negative mood, t (48) = 2.31, p =.025. Popular children did not differ significantly in their rates of approach or withdrawal from either rejected children, t (47) = 1.31, p =.198, or neglected children, t (48) =.68, p =.502. However, controversial children were rated as significantly different from rejected children, t (32) = 4.12, p <.0005, neglected children, t (33) 2.97, p =.005, and other children, t (74) = 2.41, p =.018, with respect to levels of approach and withdrawal. Controversial children also differed from popular children on this aspect of Personal-Social Inflexibility, t (35) = 2.19, p =.035, although not on ratings of adaptability or mood. Specifically, controversial children were rated as more likely to approach than rejected, neglected, other, or popular children. There were no other significant differences between status groups on the dimensions within this factor.

Discussion

Previous research has indicated that temperamental patterns such as high activity level and patterns of approach and withdrawal may play a part in social competence and adjustment within the peer group (Farver & Branstetter, 1994; Mobley & Pullis, 1991; Stocker & Dunn, 1990). The present results support and extend these previous findings by indicating that temperamental characteristics appear to be directly related to children’s social status within the peer group. In contrast to expectations, there were no significant interactions between sex and social status with respect to temperamental characteristics. Findings with respect to sex differences will be discussed first.

Sex Differences

Past research has produced mixed results with respect to sex differences in temperamental characteristics with several studies reporting no significant differences between preschool-aged boys and girls (e.g., McDevitt & Carey, 1978; Mobley & Pullis, 1991; Sanson et al., 1993; Simpson & Stevenson-Hinde, 1985; Wright Guerin & Gottfried, 1994). However, the results of the present study confirm intuitive notions regarding sex differences in expression of temperamental characteristics for preschool-aged children. Specifically, boys were rated by teachers as higher on Task Distractibility (more distractible, more active and less persistent) than girls. Boys were also rated as more intense than girls although not more likely to display negative mood (Reactivity). While the dimension of Task Distractibility appears to be particularly important for behavioral adjustment to a preschool or school environment (Sanson et al., 1993), previous research has indicated that behavioral problems such as aggression or noncomplian ce, exhibited more often by boys, may be more closely linked to characteristics such as high intensity (Mobley & Pullis, 1991). Thus, the present findings provide support for the proposition that there are clear sex differences in temperamental characteristics at preschool age that may place boys at greater risk for social difficulties.

If, as results reported by Sanson et al., 1993; Sanson, Smart, Prior, Oberkiald, & Pedlow, 1994) suggest, temperament differences between boys and girls are negligible during infancy yet increase with age, the sex differences in temperament identified in the present study may be as much, if not more, a result of socialization into gender roles within both family and peer interactions as biological differences between boys and girls. Given that social role stereotypes appear quite early on, and that socialization processes appear to be heavily implicated in the development of aggressive or antisocial behavior (Perry, Perry, & Rasmussen, 1986; Sanson et al., 1993), it is important to recognize that early social learning experiences may play a large role in the development of childhood temperament.

Social Status Differences

With respect to social status differences, there were significant differences between popular children and both rejected children and average status children along the temperament dimension of Task Distractibility. Specifically, children identified as rejected by their peers were rated by teachers as showing higher activity rates, higher distractibility and lower persistence than popular children while children classified as average status were rated as showing higher activity rates and higher distractibility but not lower persistence than popular children. Thus, while rejected children do not appear to differ from average children with respect to Task Distractibility, low activity rates, low distractibility and high persistence appear to be linked to popular social status for preschool-aged children. Interestingly, children classified as other were rated similarly to popular children on the dimension of Task Distractibility and as significantly different from both rejected and average children in terms of lo wer activity rates and lower distractibility.

The dimension of Personal-Social Inflexibility also emerged as an important discriminator between social status groups with both rejected and neglected children being rated as displaying significantly higher Personal-Social Inflexibility in terms of lower adaptability and more negative mood than popular children. These results are in line with previous research that has identified Personal-Social Inflexibility (or flexibility) as related to children’s success in peer socialization (Mobley & Pullis, 1991; Stocker & Dunn, 1990) and friendship status (Farver & Branstetter, 1994). Although, interestingly, rejected and neglected children did not differ significantly from popular children with respect to rates of approach or withdrawal, rejected, neglected, popular and other children were all rated as significantly different from controversial children on this aspect of Personal-Social Inflexibility. Specifically, controversial children were rated as displaying higher levels of approach and lower levels of withdraw al than children in rejected, neglected, popular or other groups.

Previous research has indicated that controversial children display behavior that represents a combination of the behavioral patterns shown by rejected and popular children. For example, in a meta-analytic review of the behavioral correlates of sociometric status, Newcomb, Bukowski, and Pattee (1993) concluded that while controversial children displayed levels of aggression similar to those exhibited by rejected children, they had other prosocial qualities which protected them from exclusion from the peer group. The present results support the proposition that the elevated levels of approach with respect to social interaction displayed by controversial children may ameliorate their more negative qualities enabling them to maintain a level of social acceptance denied to rejected children. Although research into the behavioral characteristics of controversial children has been hampered in the past due to the low numbers of children classified into this group, further research into the reasons controversial chil dren are not rejected, despite their aggressive behavior, may improve our understanding of the phenomenon of peer rejection.

Conclusion

The strong links between temperamental characteristics and social status evident in the present findings suggest, as Brownell and Hazen (1999) propose, that individual temperamental characteristics might be more directly translated into individual differences in styles of social interaction during early childhood than in later school years. While individual differences in peer competence among older children may be more a function of complex interactions between temperamental characteristics and social experiences, during the early childhood years, temperament may make a large contribution towards both the quantity and the quality of children’s interactions with their peers. Thus, a child initially rejected by his or her peers due to temperamental characteristics may fail to develop the effective interpersonal skills necessary for mature and competent social behavior. The present results therefore indicate that intervention programs for children at preschool age need to take into account the particular temper amental styles, which appear to be associated with rejection in early childhood.

As models of social competence emphasize the ongoing interactions between individuals and the environment, competent behavior must include the ability to generate behavioral responses, which match the situational requirements of the social environment (Wine & Smye, 1981). The temperament differences in Task Distractibility between boys and girls identified in the present study appear to be relevant to adjustment both within the peer group and, more broadly within the preschool environment (Mobley & Pullis, 1991). Thus, future research into the social learning experiences of young boys which appear to interact with early temperamental characteristics to place boys at greater risk for the development of social difficulties would appear to be well worthwhile.

Footnotes

This study was conducted as part of a program of doctoral research by the first author. The research was supported in part by a scholarship from Queensland University Technology. Portions of this research were presented at the Australasian Human Development Association Biennial Meeting, June 1999.

References

Asher, S. R., & Dodge, K. A. (1986). Identifying children who are rejected by their peers. Developmental Psychology, 22(4), 444-449.

Brownell, C. A., & Hazen, N. (1999). Early peer interaction: A research agenda. Early Education and Development, 10(3), 403-413.

Bukowski, W. M., Gauze, C., Hoza, B., & Newcomb, A. F. (1993). Differences and consistency between same-sex and other-sex peer relationships during early adolescence. Developmental Psychology, 29(2), 255-263.

Coie, J. D., Dodge, K. A., & Kupersmidt, J. B. (1990). Peer group behaviour and peer group social status. In S. R. Asher & J. D. Coie (Eds.), Peer rejection in childhood (pp. 17-59). Cambridge, England: Cambridge University Press.

Fabes, R. A., Shepard, S. A., Guthrie, I. K., & Martin, C. L. (1997). Roles of temperamental arousal and gender-segregated play in young children’s social adjustment. Developmental Psychology, 33(4), 693-702.

Farver, J. A. M., & Branstetter, W. H. (1994). Preschoolers’ prosocial responses to their peers’ distress. Developmental Psychology, 30(3), 334-341.

Hymel, S., Bowker, A., & Woody, E. (1993). Aggressive versus withdrawn unpopular children: Variations in peer and self-perceptions in multiple domains. Child Development, 64(3), 879-896.

Keogh, B. K., & Burstein, N. D. (1988). Relationship of temperament to preschoolers’ interactions with peers and teachers. Exceptional Children, 54(5), 456-461.

Keogh, B. K., Pullis, M. E., & Cadwell, J. (1982). A short form of the teacher temperament questionnaire. Journal of Educational Measurement, 19(4), 323-329.

McDevitt, S. C., & Carey, W. B. (1978). The measurement of temperament in 3-7 year old children. Journal of Child Psychology and Psychiatry, 19, 245-253.

Mobley, C. E., & Pullis, M. E. (1991). Temperament and behavior adjustment in preschool children. Early Childhood Research Quarterly, 6, 577-586.

Newcomb, A. F., Bukowski, W. M., & Pattee, L. (1993). Children’s peer relations: A meta-analytic review of popular, rejected, neglected, controversial, and average sociometric status. Psychological Bulletin, 113 (1), 99-128.

Perry, D. G., Perry, L. C., & Rasmussen, P. (1986). Cognitive social learning mediators of aggression. Child Development, 57, 700-711.
Sanson, A., Prior, M., Smart, D., & Oberklaid, F. (1993). Gender differences in aggression in childhood: Implications for a peaceful world. Australian Psychologist, 28(2), 86-92.

Sanson, A. V., Smart, D. F., Prior, M., Oberklaid, F., & Pedlow, R. (1994). The structure of temperament from age 3 to 7 years: Age, sex and sociodemographic influences. Merrill Palmer Quarterly, 40 (2), 233-252.

Simpson, A. E., & Stevenson-Hinde, J. (1985). Temperamental characteristics of three to four-year-old boys and girls and child-family interactions. Journal of Child Psychology and Psychiatry, 26 (1), 43-53.

Stevens, J. (1996). Applied multivariate statistics for the social sciences. New Jersey: Lawrence Erlbaum.

Stocker, C.,& Dunn, J. (1990). Sibling relationships in childhood: Links with friendships and peer relationships. British Journal of Developmental Psychology, 8, 227-244.

Thomas, A., & Chess, S. (1977). Temperament and development. New York: Brunner/Masel.

Wine, J. D., & Smye, M. D. (1981). Social competence. New York: Guilford.

Wright Guerin, D., & Gottfried, A. W. (1994). Developmental stability and change in parent reports of temperament: A ten-year longitudinal investigation from infancy through preadolescence. Merrill-Palmer Quarterly, 40 (3), 334-355.

A multiple-risk interaction model: effects of temperamentand divorce on psychiatric disorders in children

Exposure to highly stressful major life change events may overwhelm the coping capacity of children, and thus compromise favorable adjustment (e.g., Garmezy, Masten, & Tellegen, 1984; Gersten, Langner, Eisenberg, & Simcha-Fagan, 1977; Rutter, 1983). Research has indicated that this is particularly true for children in the circumstances surrounding parental divorce, and in the immediate aftermath (see reviews by Emery, 1982, 1988; Hetherington & Camara, 1984). Compared to children of intact families, many children of recently divorced families are reported to demonstrate less social competence, more behavioral problems, more psychological distress, and more learning deficits (e.g., Amato & Keith, 1991a; Hetherington, 1972; Hetherington, Cox, & Cox, 1979, 1982; Peterson & Zill, 1983, 1986; Wallerstein & Kelly, 1980), and are overrepresented in referrals to clinical services (Guidubaldi, Perry, & Cleminshaw, 1984; Kalter, 1977). Further, an accumulating body of evidence from longitudinal studies of divorce supports continuity of negative effects beyond the 2-year postdivorce crisis period in a substantial minority of children and adolescents (Guidubaldi & Perry, 1984, 1985; Hetherington & Anderson, 1987; Hetherington & Clingempeel, 1992; Hetherington, Cox, & Cox, 1985; Hetherington, Stanley-Hagan, & Anderson, 1989; Wallerstein, 1985, 1987, 1991), as well as the reemergence or emergence of problematic behavior in adolescents who had previously recovered from or adjusted well to parental divorce (Hetherington, 1991a). Moreover, reports of long-term negative outcomes in offspring beyond the adolescent period suggest that the ramifications of parental divorce on adult behavior may be even more deleterious than those on child behavior (Amato & Keith, 1991b; Zill, Morrison, & Coiro, 1993). The evidence appears to be quite convincing that dissolution of two-parent families, though it may benefit spouses in some respects (Hetherington, 1993), may have farreaching adverse effects for many children.

The divorce and family systems literatures indicate that negative family processes may be more important predictors of poor adjustment in children than family structure (Baumrind, 1991a, 1991b; Kelly, 1988; O’Leary & Emery, 1984). Interparental conflict, for example, is associated with adjustment disturbances in children in both divorced and nondivorced families (Camara & Resnick, 1988; Johnston, Campbell, & Mayes, 1985; Peterson & Zill, 1986; Reid & Crisafulli, 1990), and is considered to be a critical mediator of divorce effects in children and adolescents (Atkeson, Forehand, & Rickard, 1982; Emery, 1982; Forehand, Long, & Brody, 1988; Kelly, 1988; Luepnitz, 1979). In addition, the stress associated with shifting family roles and relationships in newly divorced families contributes to a breakdown in effective parenting practices, which in turn influences adjustment outcomes in children. Decreased levels of warmth, support, tolerance, control, and monitoring, and increased levels of punitive erratic discipline among recently divorced mothers have been related to problematic adjustment in children (Bray, 1990; Brody & Forehand, 1988; Maccoby, Buchanan, Mnookin, & Dornbusch, 1993). Furthermore, long-term studies of divorce suggest that negative family processes and concomitant stressors may be in operation well after the divorce has occurred, and may become exacerbated when offspring enter adolescence (Hetherington, 1993; Hetherington & Anderson, 1987; Hetherington & Clingempeel, 1992).

Coping with family stressors of such a demanding nature, particularly over an extended time period, may easily tax or exceed the cognitive and behavioral resources that are available to children (Lazarus & Folkman, 1984). Therefore, it is reasonable to assume that the capacity to cope with and adjust to a stressful life circumstance such as divorce may be even further undermined among those children with difficult temperaments or histories of behavioral or emotional problems (Caspi, Elder, & Herbener, 1990; Hetherington, 1991b; Rutter, 1987). To test this supposition, the role played by preexisting (i.e., predivorce) individual characteristics such as temperament on children’s responses to divorce needs to be examined, thus advancing a multiple-risk interaction model of adjustment outcomes in children. This same risk model may also be applied to investigate whether divorce affects the course of psychopathology already present in children from a developmental trajectory perspective.

One of the pathways to later disturbance has been linked to earlier temperament difficulties and/or problematic adjustments in childhood (Cohen & Brook, 1987; Farrington, Loeber, & Van Kammen, 1990; Loeber & Dishion, 1983; Mannuzza et al., 1991; Olweus, 1980), suggesting a vulnerability for future disorders among already troubled children. For example, Mannuzza et al. found that a childhood diagnosis of attention deficit disorder persisted in 40% of probands at age 18, and increased the risk of antisocial and conduct disorder diagnosis by almost five times. However, while childhood disorder was clearly a risk for later disturbance, the alternative perspective is that stability of diagnosis was not observed in over half the probands. Further, findings from other studies investigating the persistence of childhood hyperactivity and deficits in attention span and impulse control are more equivocal, and indicate more modest relationships in adolescence and in adulthood (see review by Klein & Mannuzza, 1991). Thomas and Chess (1980) found that the continuity of childhood behavioral adjustment difficulties may be altered by the family’s response to the developing child, and suggested that earlier problems per se are not sufficient to predict later maladjustment. Such observations reflect an interactionalist model of human development whereby individual characteristics interact with psychosocial features of the environment, and as a result are modified (Plomin, 1993; Thomas & Chess, 1980). From a developmental trajectory framework, it is the moderating role played by divorce on the risk of disturbance among already vulnerable children that is at issue. It is this perspective that we consider here.

The purpose of the current study was twofold. The first was to test an adjusted risk model of the effect of family status on outcomes of psychopathology in children by examining whether postdivorce difficulties of children merely reflect stability in preexisting temperament-adjustment difficulties or if divorce has an independent effect. The second was to test a multiple-risk interaction model of temperament-adjustment and family status effects by examining whether parental divorce alters the developmental trajectory already in evidence between preexisting temperament-adjustment difficulties and associated outcomes of psychopathology 8 years later.

There is some evidence that children whose parents eventually divorce exhibit higher levels of problematic behavior prior to divorce than children whose families remain intact (Block, Block, & Gjerde, 1986). The implication is that, in marriages which eventually dissolve, adverse family processes such as parental conflict or disrupted childrearing (most likely due to preoccupation with marital problems) are in evidence before marital dissolution and precipitate predivorce adjustment problems in children. In order to rule out the possible contaminating effects of divorce-related family processes on children’s adjustment prior to divorce (as demonstrated by Block and colleagues), predivorce differences in child temperament-adjustment between children of families who remained intact versus children of families who divorced were examined.

Only custodial mother families were included in the sample. Because almost one-third of the divorced custodial mothers in this study had remarried, and because the transition into a stepfamily is a major life change with its own complex set of stressors to which children of divorce are often exposed (Bray, 1988; Hetherington, 1987), we examined outcomes as a function of the independent and moderating effects of both single custodial mother family status and stepfamily status. We considered the outcomes to be long-term in that they were measured 8 years after temperament was observed in 100% of the sample, and 4 or more years postdivorce in 68% of the sample (and 2 or more years postdivorce in 82% of the sample). As time since divorce could not be kept constant for the entire sample, it was included as a control variable in all analyses where family status effects were tested.

The authors proposed to answer the following questions:

1. Were there predivorce differences in temperament-adjustment between children who remained in intact families and children whose parents divorced?
2. Is problematic temperament-adjustment in childhood a risk for later psychopathology in adolescence?
3. Is divorce and/or remarriage a risk for later psychopathology after accounting for the risk of earlier temperament-adjustment problems?
4. Does divorce and/or remarriage after the pathway established between preexisting temperament-adjustment characteristics and later outcomes of psychopathology?
5. Are there significant sex differences in the effects of temperament-adjustment or divorce and/or remarriage on later psychopathology, or in the effects of the temperament-adjustment to psychopathology trajectory as modified by divorce and/or remarriage?

Method

Sample

The study sample was drawn from the Children in the Community Project, a longitudinal investigation of risk factors for child psychopathology initiated in 1975 in a randomly selected sample of 976 families with children between 1 and 10 years old living in two New York upstate counties. (See Kogan, Smith, & Jenkins, 1977, for a description of sampling procedures and response rates.) The original sample was broadly representative of families living in the northeastern United States. Both mother and child participated in three follow-up studies in 1983, 1986, and 1992. Data from the initial wave in 1975 and first follow-up study conducted 8 years later (in 1983) are employed here.
Of the original 976 families interviewed in 1975, 805 had biological mothers and fathers and were maritally intact. These families comprised the eligible sample for the current study. In 1983 when the children were ages 9 to 18, 699 or 86.8% of the 1975 eligible families were reinterviewed. Youngest children from poor urban families were slightly more likely to be lost; otherwise, demographic characteristics closely matched those of the 1975 eligible sample as tested by chi-square analyses. Of those reinterviewed, 648 (92.7%) of the 699 families fulfilled study criteria for retention in the current study: 508 families had remained married (intact families), 99 had divorced or separated with single mothers retaining custody of the study children (single custodial mother or SCM families), and 41 had divorced with custodial mothers remarrying (stepfamilies). The remaining 51 families were eliminated because the mothers were no longer living with the children (21), the mothers had been widowed (13), the mothers had obtained more than one divorce (6), the fathers were institutionalized (6), or data were incomplete (5). Of the 140 divorced families, 11 separated or divorced within 1 year after the 1975 assessment and 129 (92%) separated or divorced I or more years after the 1975 assessment. Average interval between the 1975 assessment and separation or divorce was 3 1/2 years. Demographic characteristics of the current study sample grouped by family status are presented in Table I.

Materials

The 1975 data set consisted of information obtained from a 1-hour structured interview of mothers covering demographic characteristics, health and pregnancy problems, childrearing practices, and behavioral and emotional functioning of their children. In the 1983 follow-up study, both mothers and children participated. Data obtained from the mothers included portions of the original interview, a self-administered questionnaire on mother and child relationships and personality characteristics, and a child diagnostic interview. Data obtained from youths included parallel personality and diagnostic measures. Youths and their mothers were interviewed separately in their homes by two trained lay interviewers. See Lewis, Gorsky, Cohen, and Hartmark (1985) for a description of interviewer selection and training, and explanation to study participants.

Predivorce Measures (1975). The four broad-band measures of children’s temperament-adjustment used here were based on factor analyses of 14 scales reflecting a range of childhood behaviors which captured the temperament constructs originally defined by Thomas, Chess, and Birch (1968). The scales were derived from psychometric analyses of a priori items representing narrow-band symptom sets (Cohen & Brook, 1987). Sample scale items and Cronbach alphas for the four broad-band measures [TABULAR DATA FOR TABLE I OMITTED! employed, Immaturity, Anxiety, Behavior Problems, and Affective Problems, are provided in Table II. Study sample mean scale scores and ranges are 19.9, 16.3 to 25.2 for Immaturity; 20.0, 15.8 to 26.6 for Anxiety; 49.9, 42.1 to 60.6 for Behavior Problems; and 10.0, 8.8 to 14.4 for Affective Problems. High scores on Immaturity indicate a clumsy, distractible child who has much difficulty adapting to a variety of social or learning situations. High scores on Anxiety indicate a sensitive, shy, fearful child. High scores on Behavior Problems indicate an aggressive, argumentative, excitable child with excessive motoric activity. High scores on Affective Problems indicate a child who is frequently unhappy and derives little enjoyment in social situations.

Families’ 1975 socioeconomic status (SES) was measured by a sum of standardized measures of father and mother education, occupational status of the fathers, and family income.

Postdivorce Measures (1983). Diagnoses meeting DSM-III-R criteria (American Psychiatric Association, 1987) were derived primarily from mother and child responses to the Diagnostic Interview Schedule for Children (DISC-1; Costello, Edelbrock, Dulcan, Kalas, & Klaric, 1984). However, [TABULAR DATA FOR TABLE II OMITTED! DSM-III-R diagnostic demands were not always met because the DISC-1 was developed to obtain information to satisfy DSM-III criteria (American Psychiatric Association, 1980). Therefore, it was necessary to supplement the information obtained from the DISC-1 with information from items elsewhere in the protocol in order to adequately assess DSM-III-R symptoms. Syndrome scales of Attention Deficit Hyperactivity Disorder, Separation Anxiety Disorder, and Major Depressive Disorder, each based on all relevant items, were constructed. We considered diagnostic criterion to be met with affirmation of the presence of a symptom by either mother or child. For the current study, children who met diagnostic criteria and also had scaled scores of more than 1 standard deviation above the population mean were give a diagnosis. See Cohen et al. (1993) and Cohen, Velez, Kohn, Schwab-Stone, and Johnson (1987) for additional information regarding the construction of scales and diagnoses. Support for the reliability, criterion validity, and construct validity. of these diagnoses was reflected in improved agreement with clinically obtained diagnoses when scaled measures were used (Cohen, O’Connor, Lewis, & Malachowski, 1987; Cohen, Velez, et al., 1987). The diagnoses also have been related in theoretically meaningful ways to risk factors measured prospectively (Velez, Johnson, & Cohen, 1989), to pregnancy and early health problems (Cohen, Velez, Brook, & Smith, 1989), and to treatment (Cohen, Kasen, Brook, & Streuning, 1991).

Mothers provided reports of divorce or remarriage since their previous interviews.

Data Analytic Procedure

The first step was to examine whether any negative influences of divorce-related processes were indeed in operation prior to actual physical separation of spouses (as suggested by Block et al., 1986). Predivorce temperament-adjustment differences between children of families who remained intact and children whose parents later divorced (including mothers who had remarried or remained single) were tested by analyses of covariance (ANCOVAs). Analyses were conducted separately for girls and boys (controlled for age and SES), and for the pooled sample (with sex added to the control variables).

For the longitudinal analyses, we began by using set correlation procedure to examine the overall main and interaction effects of 1975 temperament-adjustment and family status on the set of six 1983 psychiatric diagnoses as a safeguard against Type 1 error (Cohen & Cohen, 1983). In set correlation, multiple independent and dependent variables can be tested simultaneously, and the effects of control variables can be partialed out of both the independent and dependent variable sets. Thus, because a smaller number of significance tests are conducted, and because the effects of control variables are eliminated, the likelihood of obtaining significant results by chance is reduced. When, in set correlation, a significant main or interaction effect on the set of six 1983 psychiatric diagnoses was observed, each diagnostic outcome was regressed against the specific risk(s) and all control variables to obtain the odds ratio (i.e., the degree of increased risk for a given disorder). Logistic regression is the procedure of choice for analyzing the effects of continuous or categorical risk variables on a binary outcome variable in terms of the odds of being in one of two categories versus the other (Fleiss, Williams, & Dubro, 1986).

There is substantial empirical evidence available to support a differential effect of divorce for boys and for girls (see reviews by Zaslow, 1988, 1989). However, although a sex-specific analysis of divorce effects may be considered more informative than an analysis which employs sex as a control variable, it also precludes testing for significant sex differences. Moreover, to do both becomes quite unwieldy with such a large set of variables. Therefore, all longitudinal analyses addressing main and interaction effects of temperament-adjustment and family status on outcomes were tested for sex differences, and where significant, separate results for girls and boys are reported.

Following the cross-sectional analysis of predivorce differences, the longitudinal analyses of main and interaction effects of temperament-adjustment and family status proceeded as follows:

• Main Effects of Temperament-Adjustment. The set of six 1983 psychiatric diagnoses was regressed against each 1975 temperament-adjustment scale after partialing out the main effects of age, sex, and SES. The temperament-adjustment scales were standardized for all regression analyses so that the unit change in the dependent variable set would be comparable across all six diagnoses.
• Main Effects of Family Status. The set of six 1983 psychiatric diagnoses was regressed against family status after partialing out the main effects of age, sex, SES, time since divorce, and all four 1975 temperament-adjustment scales. Family status was entered into all equations as two dummy variables: single custodial mother families versus intact families and stepfamilies, and stepfamilies versus intact families and SCM families. When considered simultaneously, the regression coefficients reflect the differences between SCM families and stepfamilies, respectively, with intact families (Cohen & Cohen, 1983, p. 194).
• Interaction Effects Between Temperament-Adjustment and Family Status. The set of six 1983 psychiatric diagnoses was regressed against the interaction between each 1975 temperament-adjustment scale and family status after partialing out the main effects of age, sex, SES, time since divorce, family status, and all four 1975 temperament-adjustment scales.
• Sex Differences. Potential sex differences in main and interaction effects of temperament-adjustment and family status, and subsequent pairwise comparisons of effects for each sex, were examined.

Results

Predivorce Differences

If predivorce processes were in operation, we would expect the 1975 temperament-adjustment scales to differentiate between children of families who remained intact and children of families who divorced by 1983. However, there were no significant predivorce differences in Immaturity, Anxiety, Behavior Problems, or Affective Problems between girls of intact families and girls of divorced families, or between boys of intact families and boys of divorced families. We repeated these analyses for the pooled sample to increase power, but still found no predivorce differences. Mean family status group differences for predivorce temperament-adjustment are given in Table III.

A Developmental Trajectory for Vulnerability

Significant main effects of predivorce temperament-adjustment on 1983 psychiatric diagnoses were observed with set correlation for Immaturity, F(6, 638) = 2.72, p [less than].05, Behavior Problems, F(6, 638) = 8.85, p [less than].001, and Affective Problems, F(6, 638) = 2.50, p [less than!.05, and followed up with logistic regression. No overall effect was observed for Anxiety, F(6, 638) = 0.35, p [greater than].05. Youths who scored high on Immaturity were at significant increased risk for ADHD, oppositional defiant disorder, and conduct disorder. Youths who scored high on Behavior Problems were at significant increased risk for ADHD, oppositional defiant disorder, conduct disorder, overanxiety disorder, and major depressive disorder. And youths who scored high on Affective Problems were at significant increased risk for oppositional defiant disorder. There was a near significant trend (p [less than].10) for youths who scored high on Immaturity to be at increased risk for major depressive disorder, and for youths who scored high on Behavior Problems to be at increased risk for separation anxiety disorder. See Table IV for odds ratios and confidence intervals. (An odds ratio of 3.0 is interpreted as being three times more at risk for disorder, whereas an odds ratio at a fractional level, e.g., 1.5, is interpreted as being at 50% increased risk for disorder.)

A significant two-way interaction effect between sex and Behavior Problems on 1983 psychiatric diagnoses was observed with set correlation, F(6, 637) = 2.46, p [less than].05, and followed up with logistic regression. Odds ratios and confidence intervals are given in Table V. Compared to boys who scored high on Behavior Problems, girls with comparable scores were at significant increased risk for ADHD and separation anxiety disorder, and showed a near significant trend to be at increased risk for oppositional defiant disorder and conduct disorder. Examination of sex-specific odds ratios revealed a similar predictive pattern of high scores on 1975 Behavior Problems and later disorder across sex; however, the data suggest a somewhat stronger effect for girls than for boys overall.

Main Effects of Family Status on Adolescent Psychopathology

Significant independent main effects of family status on 1983 psychiatric diagnoses were observed with set correlation, F(12, 648) = 3.31, p [less than!.001, and followed up with logistic regression. Odds ratios and confidence intervals are given in Table VI. Youths in single custodial mother families were three times more at risk for conduct disorder, almost two times more at risk for overanxiety disorder, and almost three times more at risk for separation anxiety disorder than youths in intact families. In addition, there was a near significant trend for youths in SCM families to be almost two times more at risk for oppositional defiant disorder than youths in intact families. Youths in stepfamilies were over four times more at risk for ADHD and almost four times more at risk for conduct disorder than youths in intact families.

A significant two-way interaction effect between sex and family status on 1983 psychiatric diagnoses was observed with set correlation, F(6, 637) = 2.46, p [less than].05, and followed up with logistic regression (see Table VI). The risk of major depressive disorder was significantly higher in SCM family boys than in SCM family girls. Inspection of sex-specific odds ratios revealed that boys in SCM families were over five times more at risk for major depressive disorder than were boys in intact families. Girls in SCM families, on the other hand, were no more at risk for major depressive disorder than were girls in intact families. In general, there were more pervasive effects of SCM family status for boys than for girls (albeit only significantly more so for major depressive disorder). Compared to boys in intact families, boys in SCM families were at significant increased risk for four of the six diagnoses (conduct disorder, overanxiety disorder, separation anxiety disorder, major depressive disorder). In contrast, SCM family status effects on girls were more modest, and did not reach significance (compared to intact family girls) when treated independently of the effects for boys, despite the larger ratio of girls to boys in that family context and subsequent increase in power. There was no significant difference in risk of psychiatric disorder between girls and boys in stepfamilies.

The observed variations in pattern of outcome and family status when girls and boys were treated independently are worthy of note (see Table VI). Compared to intact family girls, SCM family girls were not at significant increased risk for any of the six disorders, whereas stepfamily girls were at a significant increased risk for ADHD, and at a near significant increased risk for major depressive disorder and overanxiety disorder. Compared to intact family boys, SCM family boys were at a significant increased risk of conduct disorder and all three internalizing disorders, whereas stepfamily boys were at a significant increased risk of disruptive disorders (ADHD, conduct disorder) only. Thus, in these reduced, independently treated samples, girls were at significantly elevated risk for psychiatric disturbance only in stepfamilies, whereas boys were at significantly elevated risk for psychiatric disturbance within both the SCM family and stepfamily contexts, but the expression of risk varied.

Interaction Effects of 1975 Temperament-Adjustment and Family Status on Adolescent Psychopathology

Significant two-way interaction effects between Immaturity and family status, F(12, 1,260) = 2.54, p [less than].01, between Behavior Problems and family status, F(12, 1,260) = 1.80, p [less than].05, and between Affective Problems and family status, F(12, 1,260) = 3.03, p [less than!.001, on 1983 psychiatric diagnoses were observed with set correlation and followed up with logistic regression. Odds ratios and confidence intervals are given in Table VII. A near significant trend for Anxiety and family status to interact was also observed, F(12, 1,260) = 1.65, p [less than!.10. Among youths who scored high on Immaturity, those in stepfamilies were over two times more at risk for oppositional defiant disorder than those in intact families; however, those in single custodial mother families were at only about half the risk for overanxiety disorder compared to those in intact families. Among youths who scored high on Affective Problems, those in stepfamilies were over five times more at risk for oppositional defiant disorder than those in intact families. There was also a near significant trend for SCM family youths who scored high on Affective Problems to be more at risk for oppositional defiant disorder than were intact family youths with comparable scores. Interaction effects between Behavior Problems and family status were not observed with logistic regression: Increased risk of later disorder given high scores on Behavior Problems was observed for all family status groups.

A significant three-way interaction effect between sex, Immaturity, and family status on 1983 psychiatric diagnoses was observed with set correlation, F(12, 1,250) = 2.44, p [less than].01, and followed up with logistic regression. The risk of conduct disorder was significantly higher in stepfamily boys who scored high on Immaturity than in stepfamily girls who scored high on Immaturity. Stepfamily boys with high Immaturity scores were over six times more at risk for conduct disorder than were intact family boys with comparable Immaturity scores. Stepfamily girls with high Immaturity scores, on the other hand, were no more at risk for conduct disorder than were intact family girls with comparable Immaturity scores. In addition, SCM family boys who scored high on Immaturity were at a near significant increased risk for separation anxiety disorder compared to SCM family girls who scored high on Immaturity. The risk of separation anxiety disorder increased almost three times among SCM family boys with high Immaturity scores compared to intact family boys with comparable Immaturity scores. SCM family girls with high Immaturity scores, however, were no more at risk for separation anxiety disorder than were intact family girls with comparable Immaturity scores. Sex-specific odds ratios and confidence intervals are given in Table VIII.

Discussion

Our intention here was to test risk associated with problematic predivorce temperament-adjustment and nonintact family status for later disturbance as interpreted from a developmental trajectory framework. Data are longitudinal and were obtained prospectively from a randomly selected community sample prior to and independently of divorce. Outcomes were measured 8 years after temperament-adjustment was observed, and 4 or more years postdivorce for 68% of the sample, and 2 or more years postdivorce for 82% of the sample. Mother reports were used for the predivorce temperament-adjustment characteristics and combined mother and child reports were used for the diagnostic outcomes so that no informant was the sole provider of both the independent and dependent variables. To the best of our knowledge, this is the only large-scale study that demonstrates divorce effects in an epidemiological sample after controlling for predivorce temperament-adjustment characteristics. Thus, findings obtained here take on added import for the study of children’s response to marital dissolution.

Before we consider the implications of these results, some methodological issues should be addressed. First, because ethnic and cultural norms involving divorce, child custody, and remarriage vary (Cherlin, 1992), generalizability of findings from a primarily white Catholic sample may be limited. Second, although combined mother-child reports were used for the 1983 psychiatric diagnoses, only mothers provided information about predivorce temperament-adjustment. It is reasonable to assume that reports by single informants are more subject to individual biases than are the reports of multiple informants (Hetherington & Clingempeel, 1992). However, as these biases would affect children of both intact and divorced families similarly (suggested here by the lack of significant predivorce temperament-adjustment differences between these two groups), this weakness is less problematic in the current study. Furthermore, the strong association between temperament-adjustment and later disturbance found here (the latter derived from combined mother and youth reports) offers some validation of mother reports of earlier behaviors. Finally, although not a methodological issue, we are well aware that children’s responses to divorce may be less contingent upon family status than upon certain aspects of the home environment, namely, a close, trusting, and mutually respectful relationship between a child and his or her parents, and parental establishment and enforcement of standards of behavior (Maccoby et al., 1993). Family status, on the other hand, is more structural in nature, and cannot capture the complexity of postdivorce family reorganizations and changes in roles and relationships. However, permanent disruption of intact family status does imply a number of substantial stressors and life changes that require adaptive responses and coping skills on the part of children, no matter how smoothly postdivorce processes may be operating. Thus, in this study, where the focus is on predivorce vulnerability and subsequent outcomes 8 years later, we have used family status as an abbreviated proxy for the demands placed by divorce and consequent family restructuring on children’s adaptive resources and coping strategies. Nonetheless, this does not obviate the need to study postdivorce family processes in children whose adaptive and coping capabilities may be impaired.

Predivorce Temperament-Adjustment

Predivorce temperament-adjustment did not differentiate girls or boys whose parents divorced from girls or boys whose families remained intact, despite the unusually large sample and substantial statistical power. These findings are in contrast to those of Block et al. (1986), who did find preexisting differences, particularly among boys, and attributed them to divorce processes in effect prior to the actual event. Comparable to Block et al.’s study, we examined divorce as occurring in an 8-year period and on average had a 3 1/2-year interval between predivorce assessment and marital dissolution. However, the differing methodologies employed may help to explain why findings of predivorce differences were not replicated here. We used a random community sample of families with 1 to 10-year-old children and interviewed mothers, whereas Block and colleagues used a university preschool sample of children and their parents (most of whom were middle and upper class) and administered a Q-sort task to teachers. Thus, sample selection criteria, instrumentation, and informants were not at all comparable across studies. In addition, in the Block et al. study, the principal findings were based on a relatively small sample of thirty-two 7-year-old boys, five of whom came from families who subsequently divorced and 27 of whom came from families who remained intact over the 8 years from the age 6 to 7 to the age 14 to 15 assessments. Moreover, predivorce differences between the five divorced family boys and 27 intact family boys reaching conventional significance levels (i.e., p [less than].05) at the 7-year-old assessment included nine of the 63 Q-sort items. Although these differences may, to some degree, account for the inconsistent outcomes, this discrepancy between two major studies underscores the need for future investigations to gather prospective longitudinal data in order to reconcile the confounding of predivorce processes with divorce effects.

Temperament-Adjustment Effects on Later Disturbance

A strong developmental trajectory of vulnerability was supported here after adjusting for age, sex, and SES. Children who had problematic temperament-adjustment 8 years earlier, specifically those who were socially immature or who had disruptive or affective problems, were at increased risk for later psychiatric disturbance. Prediction of later disturbance by temperament-adjustment supported the developmental trajectory posited. This finding, coupled with that of no predivorce temperament-adjustment differences between children of families who divorced and children whose families remained intact, allowed us to use temperament-adjustment as a salient but nonconfounded control for examining effects of divorce independent of prior problematic adjustment.

Single Custodial Mother Family Status Effects

In comparison with youths in intact families, youths living in single custodial mother families were at increased risk for both disruptive disorders (conduct and oppositional defiant) and anxiety disorders (overanxiety and separation anxiety). These findings are in accord with those of studies of long-term divorce effects reporting both externalizing and internalizing problem behavior in adolescent offspring residing with divorced nonmarried mothers (Hetherington & Clingempeel, 1992; Wallerstein, 1985).

Externalizing symptoms in boys within this family context have been attributed to loss of the fathers in their role as a socializing agent and model for identification (Emery, 1988). Contact with nonresidential fathers tends to wane following the postdivorce transitional period (Furstenberg, Peterson, Nord, & Zill, 1983), and, when in effect, may be less characteristic of a parenting relationship than of a more peer-like friendship one (Emery, 1988). This latter phenomenon may explain the inconsistent findings regarding the relationship between continued contact with nonresident fathers and adjustment outcomes in children. While some studies of divorce have reported positive effects of regular and frequent contact with nonresident fathers on offspring adjustment (Guidubaldi, Cleminshaw, Perry, & McLaughlin, 1983; Lewis & Wallerstein, 1987), others have failed to find benefits to be derived from continued involvement (Clingempeel & Segal, 1986; Furstenberg, Morgan, & Allison, 1987; Kurdek, Blisk, & Siesky, 1981; Luepnitz, 1986). Discrepant findings regarding the impact of nonresident father contact may also be due in part to the risk of loyalty conflicts in adolescent children who attempt to maintain close ties to two warring parents (Buchanan, Maccoby, & Dornbusch, 1991). Accordingly, there is also evidence that regular and frequent contact between nonresident fathers and mother-resident offspring may be an artifact of low levels of interparental contact (Hetherington, Cox, & Cox, 1978, 1982). Nonetheless, whichever of these factors bears the brunt of responsibility for discordant outcomes, a consequence of the limited and altered exposure to biological fathers formerly in residence may be a decline in paternal monitoring and control. Coupled with the maternal decline in monitoring and control following divorce, this may provoke a disruptive response from sons that could persist over time. Given the escalating coercive negativity reported in divorced nonremarried mother-adolescent son relationships (Hetherington & Clingempeel, 1992), it is also possible that after a hiatus of decreasing problematic behavior, boys revert to an exaggerated masculine response in adolescence by engaging in machismo-like behavior as a means of separating from their mothers and establishing their own identities. Girls, on the other hand, who typically experience close, harmonious relationships with their unremarried mothers after an initial adjustment period (e.g., Hetherington et al., 1979, 1985), become argumentative and openly hostile in adolescence (e.g., Hetherington, 1991a; Kalter, Riemer, Brickman, & Chen, 1985; Newcomer & Udry, 1987). This has been attributed in part to the increasing sexual tensions and heterosexual activities that are typical of the adolescent period. Further, conflicted mother-daughter relationships may be exacerbated by the increased monitoring of daughters’ activities in adolescence by divorced nonremarried mothers (Hetherington & Clingempeel, 1992).

In contrast to a disruptive response in children of divorce, reports of symptoms of anxiety and depression have been much more equivocal. This has been attributed in part to less reliable assessment of internalizing disorders (Achenbach, McConaughy, & Howell, 1987), and to minimal awareness on the part of divorced mothers of their offsprings’ feelings (Wallerstein & Kelly, 1980). The latter may hold true even more so in adolescence. Findings of internalizing disorder here may be due in part to the use of instruments with empirically supported reliability and validity, and to the use of diagnostic criteria equally dependent on self and mother informants. It is also possible that ongoing stressors experienced by some nonremarried mothers, e.g., task overload, household disorganization, child-rearing problems, loneliness, and depression (Hetherington & Clingempeel, 1992), may contribute to the stressors encountered by children in the developmental transitions to adolescence.

Although single custodial mother effects for boys appeared to be more pervasive, sex differences were not significant with the exception of effects on depression. When treated independently of girls, boys living with their nonremarried mothers were over five times more at risk for major depressive disorder than were intact family boys. Others have suggested that boys respond more adversely than girls to psychosocial stressors, and particularly to residing with nonremarried mothers (Rutter, 1970; Zaslow & Hayes, 1986). However, heretofore most reports have indicated that they do so in an undercontrolled hostile manner (Emery, 1982, 1988; Emery, Hetherington, & DiLalla, 1985). From a psychoanalytic perspective and the conceptual work of John Bowlby on attachment, separation, and loss (Bowlby, 1973, 1980), parental loss via divorce or other circumstances for that matter is also a formidable risk factor for depression. For boys, father absence through divorce may not only mean loss of a male model and authority figure, but the more painful psychological loss of a primary attachment figure. It also may be that depression among boys within this family context is more likely with the passage of time, particularly as the child reaches the early and middle adolescent years and has the cognitive capacity to better understand the finality of the divorce (Kurdek et al., 1981). This perception may very well be reinforced by the diminished utilization of visitation rights by nonresidential fathers once the immediate crisis period following divorce has passed (Furstenberg et al., 1983). Empirical support of internalizing symptoms has been reported in long-term studies of divorce effects (Hetherington et al., 1985; Wallerstein, 1985, 1987), and in the 5-year follow-up of the National Survey of Children (e.g., Peterson & Zill, 1986), after a number of years had elapsed since the divorce and most children were in their preadolescent to adolescent years. However, observation of a significantly greater depressive response in adolescent boys than in adolescent girls to long-term residence in a nonremarried mother household has not previously been reported.

Stepfamily Status Effects

In comparison with youths in intact families, youths living in stepfamilies were at increased risk for ADHD and conduct disorder. Persistent reports of school problems among children of divorce due more to disruptions in concentration and behavioral problems than to deficits in learning ability (e.g., Guidubaldi et al., 1984; Hetherington, Camara, & Featherman, 1983), may be a repercussion of one or both of these syndromes for some children living in stepfamilies. The findings here lend support to reports that children in stepfamilies are more at risk for developing behavioral problems, psychological stress, and lower social competency than are comparably aged children in intact families (Anderson & White, 1986; Bray, 1988; Hetherington, 1987; Hetherington & Clingempeel, 1992). Data reported by Bray and Berger (1993) suggest that children who experience a remarriage are at elevated risk for developing psychological problems and adjustment difficulties during the transitional phase of stepfamily formation and after 5 years in a stepfamily, especially as they enter adolescence. Furthermore, Hetherington and Clingempeel (1992) reported that after 2 years of remarriage, thus implying a period of restabilization, adolescent step-children were still demonstrating more behavioral problems and psychological distress than were same age peers in intact families. Other findings suggest that stepfamilies are at greater risk than intact families for developing conflictual parent-child relationships (Hetherington, 1987; Santrock, Warshak, Lindberg, & Meadows, 1982), and that these relationships are related to developmental problems during adolescence (Bray & Berger, 1993; Garbarino, Sebes, & Schellenbach, 1984; Hetherington & Clingempeel, 1992).

No significant sex interaction was observed here within the stepfamily context. Although significant sex differences have been observed with younger children residing in stepfamilies (Clingempeel, Brand, & Ievoli, 1984; Hetherington et al., 1979, 1982, 1985), few differences have been reported in adolescence (Bray & Berger, 1993; Hetherington & Clingempeel, 1992). When treated independently, stepfamily girls in our sample were at elevated risk for ADHD, overanxiety disorder, and major depressive disorder, whereas stepfamily boys were at elevated risk for the disruptive disorders only (ADHD and conduct disorder). Living in a stepfather family may be particularly stressful for adolescent girls, where the potential for sexual tension (Hetherington, 1991a; Newcomer & Udry, 1987) and interrupted closeness with the mother (Kalter, 1977; Kalter et al., 1985) may foster internalizing as well as externalizing symptoms.

Temperament-Adjustment/Family Status Interaction Effects

There was some support for the alteration of a pathway to disorder, i.e., the trajectory toward future disturbance initiated by problematic temperament-adjustment characteristics was altered by the impact of divorce. Living in a stepfamily significantly increased the likelihood of oppositional defiant disorder given earlier social immaturity or affective problems. Being a part of a stepfamily requires accommodating to a second set of postdivorce transitions, sometimes before those of the single-parent family have been resolved (Hetherington & Clingempeel, 1992). The child merging into a stepfamily is asked to adapt to a new family member in a position of parental authority, and oftentimes to a new home, school, and peers. Children who have a history of social immaturity or affective problems may be less able to adapt to new situations and thus more vulnerable to disorder when such stressors are present. The significant three-way interaction between sex, Immaturity, and stepfamily status suggests that socially immature stepfamily boys respond with an increased risk of conduct disorder as well. Custodial mothers and their new spouses, attempting to solidify their relationship, may not be able to meet the increased demands for attention or give the extra support necessary to bolster the vulnerable child, who in turn may respond by acting out in a defiant manner. We also found an increased risk for oppositional defiant disorder to occur, although to a lesser extent, in children with earlier affective problems residing with their single custodial mothers. While not having to negotiate the changes characteristic of stepfamily adaptation, these children must still adapt to a stressful situation where the old rules do not apply, thus overwhelming their already vulnerable coping capacities.
On the other hand, we also found that among youths who had high predivorce social immaturity, those in single custodial mother families were at decreased risk for overanxiety disorder compared to those in intact families; this occurred with boys and, to a lesser extent, with girls, DSM-III-R diagnostic criteria for overanxiety disorder include excessive worry over past and future behaviors, excessive concern over competence, and marked self-consciousness. Children living with single custodial mothers are often asked to play more responsible roles in family matters, e.g., caring for younger siblings, having an equal say in major family decisions, and/or acting as confidantes and peers to their newly divorced mothers (Glenwick & Mowerey, 1986; Hetherington et al., 1983; Reinhard, 1977; Weiss, 1979), so that the mother-child relationship becomes more of a mutually dependent one rather than a hierarchical nurturer-dependent one. It is possible that these more “adult” experiences have enhanced immature children’s social skills and self-esteem, making it less likely that they become overanxious compared to immature children growing up in intact families who may not be afforded similar opportunities to develop earlier maturity. It also may be that mothers would be more likely to turn to their sons than their daughters to fill the role formerly held by their spouses, particularly in decision making, thus explaining the stronger effects found here for boys. An alternative explanation for this effect in boys is that mothers, in general, are less demanding of maturity in their adolescent male offspring than are fathers, and therefore socially immature boys are less anxious when they are not being pushed to act more maturely by fathers (Baumrind, 1991a, 1991b).

Conclusions and Implications

In sum, although previous reports of predivorce behavioral differences in children were not confirmed, we found the expected effects of early temperamental risk and of family disruption on the development of disorders assessed eight years later. Furthermore, effects of divorce on increased risk of psychiatric disorder were independent of prior problematic temperament-adjustment. Effects of single custodial mother family status tended to be larger for boys, not only for disruptive disorders as shown previously but also for affective and anxiety disorders, and negative effects of living in a stepfamily were more likely to be expressed with internalizing disorders for girls and with externalizing disorders for boys. Thus, sons may experience a more formidable psychological loss when fathers leave than do daughters, whereas daughters experience internalizing symptoms with the addition of a stepfather. Perhaps the presence of a stepfather is tantamount to the psychological loss of the mother for daughters in terms of divided time, attention, and loyalty. The potential for sexual tension in the adolescent stepdaughter-stepfather relationship may also be a factor. Increased risk of conduct disorder was observed in both single custodial mother family boys and stepfamily boys. The risks associated with earlier vulnerable temperament were increased by divorce and remarriage, especially with regard to subsequent oppositional defiant disorder, and for boys, conduct disorder as well. On the other hand, living in a single custodial mother family was also associated with a decline in the trajectory toward overanxiety disorders in those with an earlier vulnerable temperament, thus reflecting the complicated process of temperament-environment interactions.

In terms of future research, there are several useful directions to take. First, stress models hypothesize that protective factors may buffer the negative effect of stressful events. Thus, when the family structure is examined within the context of the larger family system, there may be alternative explanations for outcomes. For instance, it is conceivable that having a supportive sibling or a supportive substitute parental figure (such as a grandparent) could buffer the negative effects of living in single custodial mother family or stepfamily situations. Second, we have alluded to some of the potential mediating mechanisms between family structure and adolescent outcomes. Future research would benefit from focusing attention on other possible mechanisms. For example, the impact of family structure on some adolescent outcomes may be mediated by the peer group. A case in point is conduct disorders. Children residing with divorced nonremarried mothers experience less parental control and monitoring, and so may be more likely to associate with deviant peers which in turn increases the risk for conduct disorders. Finally, further research should study the manner in which children and adolescents use coping strategies to alter the likelihood of adverse emotional and behavioral outcomes when living in single custodial mother families and stepfamilies. We are mindful that family status is a structural variable, and certainly not homogeneous with regard to the complex economic, social, and psychological processes occurring within family structures. However, further investigation of the negotiation of formidable psychosocial stressors as a function of individual traits and characteristics in children will clarify the role of divorce processes on children’s well-being.

The findings of long-term divorce effects on children’s well-being have implications for intervention. Given the independent effects of living in a single custodial mother family or a stepfamily on DSM-III-R disorders, preadolescent and adolescent children of divorce may benefit from support and guidance at a professional level so that they may cope more efficiently with the typical stressors characteristic of this developmental stage. Given the increased risk of disorder among already troubled children, this may be especially critical for those with previous problematic adjustment. Adolescents have limited independent access to mental health services (Cohen & Hesselbart, 1993); however, utilization of clinical services within the school system is an accessible resource for adolescents. Alerting school guidance counselors and psychologists, particularly at the secondary school level, would facilitate obtaining such services. In addition, educating parents within these family contexts to recognize when a syndrome of symptoms connotes disturbance at a clinical level may provide the impetus necessary to motivate them to seek out this help for their children in a wider variety of clinical settings in the community.

References

Achenbach, T. M., McConaughy, S. H., & Howell, C. T. (1987). Child/adolescent emotional problems: Implications of cross-informant informants for situational-specificity. Psychological Bulletin, 101, 213-232.

Amato, P. R., & Keith, B. (1991a). Parental divorce and the well-being of children: A meta-analysis. Psychological Bulletin, 110, 26-46.

Amato, P. R., & Keith, B. (1991b). Parental divorce and adult well-being: A meta-analysis. Journal of Marriage and the Family, 53, 43-58.

American Psychiatric Association (1980). Diagnostic and statistical manual of mental disorders (3rd ed.). Washington, DC: Author.

American Psychiatric Association (1987). Diagnostic and statistical manual of mental disorders (3rd ed., rev.). Washington, DC: Author.

Anderson, J. Z., & White, G. D. (1986). Dysfunctional intact families and stepfamilies. Family Process, 25, 407-422.

Atkeson, B. M., Forehand, R., & Rickard, K. M. (1982). The effects of divorce on children. In B. B. Lahey & A. E. Kazdin (Eds.), Advances in clinical child psychology (Vol. 5, pp. 255-281). New York: Plenum Press.

Baumrind, D. (1991a). Effective parenting during the early adolescent transition. In P. A. Cowan & E. M. Hetherington (Eds.), Family transitions (pp. 111-164). Hillsdale, NJ: Erlbaum.

Baumrind, D. (1991b). The influence of parenting style on adolescent competence and substance abuse. Journal of Early Adolescence, 11, 56-94.

Block, J. H., Block, J., & Gjerde, P. F. (1986). The personality of children prior to divorce: A prospective study. Child Development, 57, 827-840.

Bowlby, J. (1973). Attachment and loss (vol. 2): Separation. New York: Basic.

Bowlby, J. (1980). Attachment and loss (vol. 3): Loss, sadness, and depression. New York: Basic.

Bray, J. H. (1988). Children’s development in early remarriage. In E. M. Hetherington & J. D. Arasteh (Eds.), The impact of divorce, single parenting and stepparenting on children (pp. 279-298). Hillsdale, NJ: Erlbaum.

Bray, J. H. (1990). Impact of divorce on the family. In R. E. Rakel (Ed.), Textbook of family practice (4th ed., pp. 111-122). Philadelphia: W. B. Saunders.

Bray, J. H., & Berger, S. H. (1993). Developmental issues in stepfamilies research project: Family relationships and parent-child interactions. Journal of Family Psychology, 7, 76-90.

Brody, G., & Forehand, R. (1988). Multiple determinants of parenting: Research findings and implications for the divorce process. In E. M.

Hetherington & J. Arasteh (Eds.), Impact of divorce, single parenting and stepparenting on children (pp. 117-134). Hillsdale, NJ: Erlbaum.

Buchanan, C. M., Maccoby, E. E., & Dornbusch, S. M. (1991). Caught between parents: Adolescents’ experience in divorced homes. Child Development, 62, 1008-1029.

Camara, K. A., & Resnick, G. (1988). Interparental conflict and cooperation: Factors moderating children’s post-divorce adjustment. In E. M.

Hetherington & J. Arasteh (Eds.), Impact of divorce, single parenting and stepparenting on children (pp. 169-196). Hillsdale, NJ: Erlbaum.

Caspi, A., Elder, Jr., G. H., & Herbener, E. S. (1990). Childhood personality and the prediction of life-course patterns. In L. Robins & M.

Rutter (Eds.), Straight and devious pathways from childhood to adulthood (pp. 13-35). New York: Cambridge University Press.

Cherlin, A. J. (1992). Marriage, divorce, and remarriage. Cambridge, MA: Harvard University Press.

Clingempeel, W. G., Brand, E., & Ievoli, R. (1984). Stepparent-stepchild relationships in stepmother and stepfather families: A multimethod study. Family Relations, 33, 465-473.

Clingempeel, W. G., & Segal, S. (1986). Stepparent-stepchild relationships and the psychological adjustments of children in stepmother and stepfather families. Child Development, 57, 574-584.

Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum.

Cohen, P., & Brook, J. S. (1987). Family factors related to the persistence of psychopathology in childhood and adolescence. Psychiatry, 50, 332-344.

Cohen, P., Cohen, J., Kasen, S., Velez, C. N., Hartmark, C., Johnson, J., Rojas, M., Brook, J., & Struening, E. L. (1993). An epidemiological study of disorders in late childhood and adolescence – I. Age- and gender-specific prevalence. Journal of Child Psychology and Psychiatry, 34, 851-867.

Cohen, P., & Hesselbart, C. S. (1993). Demographic factors in the use of children’s mental health services. American Journal of Public Health, 83, 49-52.

Cohen, P., Kasen, S., Brook, J. S., & Struening, E. L. (1991). Diagnostic predictors of treatment patterns in a cohort of adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 30, 989-993.

Cohen, P., O’Connor, P., Lewis, S. A., & Malachowski, B. (1987). A comparison of the agreement between the DISC and K-SADS-P interviews of an epidemiological sample of children. Journal of the American Academy of Child and Adolescent Psychiatry, 26, 662-667.

Cohen, P., Velez, C. N., Brook, J. S., & Smith, J. (1989). Mechanisms of the relationship between perinatal problems, early childhood illness, and psychopathology in late childhood and adolescence. Child Development, 60, 701-709.

Cohen, P., Velez, C. N., Kohn, M., Schwab-Stone, M., & Johnson, J. (1987). Child psychiatric diagnosis by computer algorithm: Theoretical issues and empirical tests. Journal of the American Academy of Child and Adolescent Psychiatry, 26, 631-638.

Costello, A. J., Edelbrock, C. S., Dulcan, M. K., Kalas, R., & Klaric, S. H. (1984). Testing of the NIMH Diagnostic Interview Schedule for Children (DISC) in a clinical population (Final report to the Center for Epidemiologic Studies, National Institute for Mental Health). Pittsburgh: University of Pittsburgh.

Emery, R. E. (1982). Interparental conflict and the children of discord and divorce. Psychological Bulletin, 92, 310-330.

Emery, R. E. (1988). Marriage, divorce, and children’s adjustment. Developmental clinical psychology and psychiatry: Vol. 14. Newbury Park, CA: Sage.

Emery, R. E., Hetherington, E. M., & DiLalla, L. F. (1985). Divorce, children, and social policy. In H. Stevenson & A. Siegal (Eds.), Child development research and social policy (pp. 189-266). Chicago: University of Chicago Press.

Farrington, D., Loeber, R., & Van Kammen, W. B. (1990). Long-term criminal outcomes of hyperactivity-impulsivity-attention deficit and conduct problems in childhood. In L. Robins & M. Rutter (Eds.), Straight and devious pathways from childhood to adulthood (pp. 62-81). New York: Cambridge University Press.

Fleiss, J. L., Williams, J. B., & Dubro, A. F. (1986). The logistic regression analysis of psychiatric data. Journal of Psychiatric Research, 20, 145-209.

Forehand, R., Long, N., & Brody, G. (1988). Divorce and marital conflict: Relationship to adolescent competence and adjustment in early adolescence. In E. M. Hetherington & J. Arasteh (Eds.), Impact of divorce, single parenting and stepparenting on children (pp. 155-167). Hillsdale, NJ: Erlbaum.

Furstenberg, F. F., Jr., Morgan, S. P., & Allison, P. D. (1987). Paternal participation and children’s well-being after marital dissolution. American Sociological Review, 52, 695-701.

Furstenberg, F. F., Jr., Peterson, J. L., Nord, C. W., & Zill, N. (1983). The life course of children of divorce: Marital disruption and parental contact. American Sociological Review, 48, 656-668.

Garbarino, J., Sebes, J., & Schellenbach, C. (1984). Families at risk for destructive parent-child relations in adolescence. Child Development, 55, 174-183.

Garmezy, N., Masten, A. S., & Tellegen, A. (1984). The study of stress and competence in children: A building block for developmental psychopathology. Child Development, 55, 97-111.

Gersten, J. C., Langner, T. S., Eisenberg, J. G., & Simcha-Fagan, O. (1977). An evaluation of the etiological role of stressful life change events in psychological disorders. Journal of Health and Social Behavior, 18, 228-244.

Glenwick, D. S., & Mowerey, J. D. (1986). When parent becomes peer: Loss of intergenerational boundaries in single parent families. Family Relations, 35, 57-62.

Guidubaldi, J., Cleminshaw, H. K., Perry, J. D., & McLaughlin, C. S. (1983). The impact of parental divorce on children: Report of the nationwide NASP study. School Psychology Review, 12, 300-323.

Guidubaldi, J., & Perry, J. D. (1984). Divorce, socioeconomic status, and children’s cognitive-social competence at school entry. American Journal of Orthopsychiatry, 54, 459-468.

Guidubaldi, J., & Perry, J. D. (1985). Divorce and mental health sequelae for children: A two-year follow-up of a nationwide sample. Journal of the American Academy of Child Psychiatry, 24, 531-537.

Guidubaldi, J., Perry, J. D., & Cleminshaw, H. K. (1984). The legacy of parental divorce: A nationwide study of family status and selected mediating variables on children’s academic and social competencies. In B. B. Lahey & A. E. Kazdin (Eds.), Advances in clinical child psychology (Vol. 7, pp. 109-151). New York: Plenum Press.

Hetherington, E. M. (1972). Effects of parental absence on personality development in adolescent daughters. Developmental Psychology, 78, 313-326.

Hetherington, E. M. (1987). Family relations six years after divorce. In K. Pasley & M. Ihinger-Tallman (Eds.), Remarriage and stepparenting today: Research and theory (pp. 185-205). New York: Guilford Press.

Hetherington, E. M. (1991a). Families, lies, and videotapes. Journal of Research on Adolescence, I, 323-348.

Hetherington, E. M. (1991b). The role of individual differences and family relationships in children’s coping with divorce and remarriage. In P. A. Cowan & M. Hetherington (Eds.), Family transitions (pp. 165-194). Hillsdale, NJ: Erlbaum.

Hetherington, E. M. (1993). An overview of the Virginia Longitudinal Study of Divorce and Remarriage with a focus on early adolescence. Journal of Family Psychology, 7, 39-56.

Hetherington, E. M., & Anderson, E. R. (1987). The effects of divorce and remarriage on early adolescents and their families. In M. D. Levine & M. D. McAnarney (Eds.), Early adolescent transitions (pp. 49-67). Lexington, MA: D. C. Heath.

Hetherington, E. M., & Camara, K. A. (1984). Families in transition: The process of dissolution and reconstitution. In R. D. Parke (Ed.), Review of child development research: The family Vol. 7 (pp. 398-439). Chicago: University of Chicago Press.

Hetherington, E. M., Camara, K. A., & Featherman, D. (1983). Achievement and intellectual functioning of children in one-parent households. In J. Pence (Ed.), Assessing achievement (pp. 206-284). San Francisco: W. H. Freeman.

Hetherington, E. M., & Clingempeel, W. G. (1992). Coping with marital transitions. Monographs of the Society for Research in Child Development (Vol. 57, No. 2-3, Serial No. 227).

Hetherington, E. M., Cox, M., & Cox, R. (1978). The development of children in mother-headed families. In H. Hoffman & D. Reiss (Eds.), The American family: Dying or developing? (pp. 117-145). New York: Plenum Press.

Hetherington, E. M., Cox, M., & Cox, R. (1979). Play and social interaction in children following divorce. Journal of Social Issues, 35, 26-49.

Hetherington, E. M., Cox, M., & Cox, R. (1982). Effects of divorce on parents and children. In M. E. Lamb (Ed.), Nontraditional families: Parenting and child development (pp. 233-288). Hillsdale, NJ: Erlbaum.

Hetherington, E. M., Cox, M., & Cox, R. (1985). Long-term effects of divorce and remarriage on the adjustment of children. Journal of the American Academy of Child Psychiatry, 24, 518-530.

Hetherington, E. M., Stanley-Hagan, M., & Anderson, E. R. (1989). Marital transitions: A child’s perspective. American Psychologist, 44, 303-312.

Johnston, J. R., Campbell, L. E. G., & Mayes, S. S. (1985). Latency children in post-separation and divorce disputes. Journal of the American Academy of Child Psychiatry, 24, 563-574.

Kalter, N. (1977). Children of divorce in an outpatient psychiatric population. American Journal of Orthopsychiatry, 47, 40-51.

Kalter, N., Riemer, B., Brickman, A., & Chen, J. W. (1985). Implications of divorce for female development. Journal of the American Academy of Child Psychiatry, 24, 538-544.

Kelly, J. B. (1988). Longer-term adjustment in children of divorce: Converging findings and implications. Journal of Family Psychology, 2, 119-140

Klein, R. G., & Manhuzza, S. (1991). Long-term outcome of hyperactive children: A review. American Academy of Child and Adolescent Psychiatry, 30, 383-387.

Kogan, L., Smith, J., & Jenkins, S. (1977). Ecological validity of indicator data as predictors of survey findings. Journal of Social Science Research, 1, 117-132.

Kurdek, L. A., Blisk, D., & Siesky, A. E., Jr. (1981). Correlates of children’s long-term adjustment to their parent’s divorce. Developmental Psychology, 17, 565-579.

Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. New York: Springer.

Lewis, J. M., & Wallerstein, J. S. (1987). Family profile variables and long-term outcomes in divorce research: Issues at a ten-year follow-up. In J. P. Vincent (Ed.), Advances in family intervention, assessment and theory (Vol. 4, pp. 121-142). Greenwich, CT: JAI Press.

Lewis, S. A., Gorsky, A., Cohen, P., & Hartmark, C. (1985). The reactions of youth to diagnostic interviews. Journal of the American Academy of Child and Adolescent Psychiatry, 24, 750-755.

Loeber, R., & Dishion, T. (1983). Early predictors of male delinquency: A review. Psychological Bulletin, 94, 68-99.

Luepnitz, D. A. (1979). Which aspects of divorce affect children. Family Coordinator, 28, 79-85.

Luepnitz, D. A. (1986). A comparison of maternal, paternal and joint custody: Understanding the varieties of post-divorce family life. Journal of Divorce, 9, 1-12.

Maccoby, E. E., Buchanan, C. M., Mnookin, R. H., & Dornbusch, S. M. (1993). Postdivorce roles of mothers and fathers in the lives of their children. Journal of Family Psychology, 7, 24-38.

Mannuzza, S., Klein, R. G., Bonagura, N., Malloy, P., Giampino, T. L., & Addalii, K. A. (1991). Hyperactive boys almost grown up. Archives of General Psychiatry, 48, 77-83.

Newcomer, S., & Udry, J. (1987). Parental marital status effects on adolescent sexual behavior. Journal of Marriage and the Family, 49, 235-240.

O’Leary, K. D., & Emery, R. E. (1984). Marital discord and child behavior problems. In M. D. Levine & R. P. Satz (Eds.), Middle childhood: Development and dysfunction (pp. 345-364). Baltimore: University Park Press.

Olweus, D. (1980). Familial and temperamental determinants of aggressive behavior in adolescent boys: A causal analysis. Developmental Psychology, 16, 644-660.

Peterson, J. L., & Zill, N. (1983, April). Marital disruption, parent-child relationships, and behavioral problems in children. Paper presented at the meeting of the Society for Research in Child Development, Detroit.

Peterson, J. L., & Zill, N. (1986). Marital disruption, parent-child relationships, and behavior problems in children. Journal of Marriage and the Family, 48, 295-307.

Plomin, R. E. (1993). Interface of nature and nurture in the family. In W. B. Carey & S. C. McDevitt (Eds.), Prevention and early intervention: Individual differences as risk factors for the mental health of children (pp. 179-189).

Reid, W. J., & Crisafulli, A. (1990). Marital discord and child behavior problems: A meta-analysis. Journal of Abnormal Child Psychology, 18, 105-117.

Reinhard, D. W. (1977). The reaction of adolescent boys and girls to the divorce of their parents. Journal of Clinical Child Psychology, 2, 21-23.

Rutter, M. (1970). Sex differences in children’s responses to family stress. In E. J. Anthony & C. Koupernik (Eds.), The child in his family (pp. 165-196). New York: Wiley.

Rutter, M. (1983). Stress, coping, and development: Some issues and some questions. In N. Garmezy & M. Rutter (Eds.), Stress, coping and development in children, New York: McGraw-Hill.

Rutter, M. (1987). Psychosocial resilience and protective mechanisms. American Journal of Orthopsychiatry, 57, 316-331.

Santrock, J. W., Warshak, R. A., Lindberg, C., & Meadows, L. (1982). Children’s and parents’ observed social behavior in stepfather families. Child Development, 53, 472-480.

Thomas, A., & Chess, S. (1980). The dynamics of psychological development. New York: Brunner/Mazel.

Thomas, A., Chess, S., & Birch, H. G. (1968). Temperament and behavior disorders in children. New York: New York University Press.

Velez, C. N., Johnson, J., & Cohen, P. (1989). A longitudinal analysis of selected risk factors for childhood psychopathology. Journal of the American Academy of Child and Adolescent Psychiatry, 28, 861-864.

Wallerstein, J. S. (1985). Children of divorce: Preliminary report of a ten-year follow-up of older children and adolescents. Journal of the American Academy of Child Psychiatry, 24, 545-553.

Wallerstein, J. S. (1987). Children of divorce: Report of a ten-year follow-up of early latency-age children. American Journal of Orthopsychiatry, 57, 199-211.

Wallerstein, J. S. (1991). The long-term effects of divorce on children: A review. Journal of the American Academy of Child Psychiatry, 30, 349-360.

Wallerstein, J. S., & Kelly, J. (1980). Surviving the break-up: How children actually cope with divorce. New York: Basic.

Weiss, R. (1979). Growing up a little faster: The experience of growing up in a single-parent household. Journal of Social Issues, 35, 97-111.

Zaslow, M. J. (1988). Sex differences in children’s response to parental divorce: 1. Research methodology and postdivorce family forms. American Journal of Orthopsychiatry, 58, 355-378.

Zaslow, M. J. (1989). Sex differences in children’s response to parental divorce: 2. Samples, variables, ages, and sources. American Journal of Orthopsychiatry, 59, 118-140.

Zaslow, M. J., & Hayes, C. D. (1986). Sex differences in children’s response to psychosocial stress: Toward a cross-context analysis. In M. E.

Lamb, A. Brown, & B. Rogoff (Eds.), Advances in developmental psychology (Vol. 4, pp. 285-337). Hillsdale, NJ: Erlbaum.

Zill, N., Morrison, D. R., & Coiro, M. J. (1993). Long-term effects of parental divorce on parent-child relationships, adjustment, and achievement in young adulthood. Journal of Family Psychology, 7, 91-103.