Introduction
Personality is a complex construct in psychology because it has different variables that require diverse tests. Given that different tests can give varied outcomes of same variable, it is imperative that there is validation of both the tests and variables to establish their consistency and reliability in measuring certain constructs of personality.
For decades, sociologists and psychologists have validated or rejected personality tests and measurements based on individual differences and the emergence of new methods of validating personality. Since personality is a complex construct, psychologists have come up with various methods of validating personality and standardizing measurements to enhance reliability and validity of personality.
In 1959, Campbell and Fiske developed a multi-trait multi-method (MTMM) matrix, which assesses a construct of personality using convergent and divergent validation. Raykov (2008) argues that, MTMM matrix is applicable when one is measuring multiple traits using various tests or methods (p.3). Convergent validation assesses whether measurements of same construct using different tests have any correlation.
On the other hand, divergent validation assesses if measurements of different constructs by different tests have any correlation. Thus, for a construct to be valid, measurements of convergent validation must have a high correlation, while those of divergent validation should have low correlation.
To explore the concept of personality validation, this essay defines personality and discusses convergent and divergent validations of MTMM matrix with a view of exploring construction of Minnesota Multiphasic Personality Inventory-2 (MMPI-2).
Personality
Debate regarding the concept of personality is still raging because personality has diverse constructs that are difficult to define using a single construct. Complexity of personality has made sociologists and psychologists to have varying definition of personality.
Colendrino-Bucu, Guerrero, Pascula, and Mateo (1993) argue that, personality is an interaction and integration of diverse traits that a person possesses, thus a product of many traits rather than just a single trait (p.308). Hence, diversity of traits that a person possesses determines personality. The nature of traits and extent of their interaction and integration result into diverse personalities that differentiate one person from another.
Usually, a person observes physical traits of people and describes them as either beautiful or handsome among other attributes that are conspicuous. In media, it is evident that actors and celebrities have certain personalities that are attractive and charming, for they appeal to many fans.
Given that celebrities and actors change their attributes depending on their performance, they complicate the definition of personality as a character trait of a person. Hence, concept of personality is complex because it consists of diverse constructs that define it.
Sociologists and psychologists define personality basing on character traits. Character traits are physical attributes that people can observe and use them to describe the behavior of a person. For instance, a good tone of voice and pleasant smile are physical attributes that can effectively appeal to other people because they stimulate emotions.
Character traits significantly depict personality of a person and determines how one appeals to people. Therefore, it means that family, friends, and society play a critical role in molding characters traits that reflect different behaviors depending on their appeal.
Given that people have different characters, Colendrino-Bucu, Guerrero, Pascula, and Mateo (1993) define personality as the sum of all characteristics and traits that make people unique in society (p.309). Since diverse characteristics and traits form part of the personality, it means that personality is dynamic.
Experiences of life constantly influence how individuals develop and shape their personalities, hence making it an experiential behavior. Thus, it is plausible to define personality as the sum of all traits that integrate and interact effectively and cause development of certain behavior.
Moreover, one can define personality from the perspective of psychoanalytic theory. Psychoanalytic theory perceives personality as an inherent attribute of a human being that lies in the psyche. The psyche consists of three components, ego, superego, and id, which play a significant role in defining personality.
According to psychoanalytic theory, process of the psyche influence emotions, thoughts, and actions, hence development of personality in an individual. Extent of consciousness influences one’s personality as either extrovert or introvert. Thus, psyche has the capacity to influence personality through emotions, actions, and thoughts that differ from one person to another.
According to Coon and Mitterer (2008), delicate balance and interaction of ego, superego, and id influence human thoughts, actions, and emotions, which determine development of personality (p.399). Hence, psyche influences stability of actions, thoughts and emotions that result into varied personalities that people possess.
Convergent Validation
To validate certain personality construct, one needs to identify specific construct that is measurable using many methods or various tests. Convergent validation assesses whether measurements of specific construct correlate when measured using different tests.
Hence, the first step of performing convergent validation is to identify a personality construct. One should identify the construct carefully because there are other variables that may confound measurements, thus invalidate the results.
Campbell and Fiske (1959) argue that, although test measurements may seem to originate from a given construct, the results have composite effects emanating from confounding variables that have similar effects on tests (p. 85). Thus, one need to choose a construct that is measurable and do not have confounding variables that significantly influence measurements.
Moreover, after identifying a construct, one needs to select several tests or methods that can measure the construct. To enhance the validity of the construct, one can choose more than five tests that are reliable when testing the construct. Each test much measure the construct and measurements recorded for correlation analysis.
Reliability of tests is critical for it determines accuracy of measurements that all tests give and subsequently influence correlation and validity of the personality construct. To compare how various tests give respective measurements of a construct, one should draw a matrix or correlation table for analysis of measurements. Correlation of measurements will determine whether the personality construct is valid or not.
In the analysis of measurements, convergent validation determines whether measurements of the construct obtained using different tests have any correlation. For measurements to be valid, they must have a high correlation. Correlation among measurements means that, the construct is stable, and tests are reliable to measure specific construct of personality.
According to Campbell and Fiske (1959), validity value of a construct must be high relative to correlations derived from the construct and other constructs (p.82). Hence, in convergent validation, measurements must show a high correlation for them to be valid as well as validate tests used in measuring the construct.
Low correlation or lack of correlation means that, measurements are unreliable and thus not valid in determining certain personality construct. Given that personality constructs are so diverse and complex, it is difficult to attribute specific construct to observed measurements.
Thus, to ensure that observed constructs give certain measurements, convergent validation is imperative to correlate measurements of specific construct. Use of different tests to measure the same construct should give the same measurements that correlate for a personality construct to be valid.
Divergent Validation
To perform divergent validation of a given personality, one requires different constructs and tests or methods. Campbell and Fiske (1959) contend that, determination of divergent validity requires at least one construct and at least one test (p.81). Divergent validation entails assessing whether measurements of different constructs measured using different tests have any correlation.
Thus, for one to conduct divergent validation effectively, it is imperative that one chooses various constructs that do not have any relationships. Constructs that have a close relationship may invalidate results since they have the same impact on different forms of tests.
Thus, different constructs should only be measurable but also unique so that if any correlation exists, it will be attributable to tests, thus invalidate the personality. Hence, differences in constructs are significant in determining divergent validity of tests and personality constructs.
Besides selecting a variety of constructs to measure, one also needs to choose different tests that are reliable in measuring all constructs. Like the constructs, tests also must have their variability lest they give similar measurements that correlate and invalidate the personality.
Raykov (2008) reasons that, divergent validation needs different traits, as well as different tests, to enhance validity contrast by using heterotrait-heteromethod (p.5). Thus, heterogeneity of tests is particularly critical in determining the validity of measurements obtained from diverse constructs.
Therefore, after identifying different constructs and tests, one needs to measure each construct using different tests before entering the results in a matrix or correlation table for analysis.
Each test much measure all different constructs so that during analysis, it is be possible to attribute each test to specified measurements and conduct comparative analysis. Moreover, the measurements need to be in summary to ease correlation analysis of each construct measured.
Divergent validity should show that there is low or no correlation among measurements for personality constructs to be valid. Insufficiency of correlation among measurements means that different constructs and tests do not have any relationships that confound measurement of personality, and thus validate personality.
According to Raykov (2008), for tests and constructs to be valid, divergent validation must show that there is no significant correlation (5). If measurements of different constructs correlate, it is an indication that, the tests or constructs are invalid. Hence, one needs to perform correlation analysis using heterotrait-monomethod and heterotrait-heteromethod.
Heterotrait-monomethod compares if measurements of the same method on different constructs have any correlation. If there is no correlation, it means that the test is reliable, and the constructs are valid; however, if there is a significant correlation among measurements, it means that test and constructs are invalid.
Regarding heterotrait-heteromethod, if measurements show any significant correlation, they indicate that both tests and constructs are invalid. In contrast, lack of correlation implies that, the measurements are valid and are reliable for analyzing specified personality.
Minnesota Multiphasic Personality Inventory-2 (MMPI-2)
MMPI-2 is the most effective instrument of assessing personality because of its reliability in evaluating different constructs. Psychiatrists and clinicians mainly use MMPI-2 when assessing patients with mental conditions and disorders.
Given that other psychometric instruments were not reliable, development of MMPI-2 enhanced validity of personality assessment particularly in health care where psychiatrists had difficulties in assessing patients. Developers employed convergent validation in construction of MMPI-2 because they used varied tests and scales to differentiate hypochondriacs from normal people.
MMPI-2 has numerous questions that developers formulated empirically using diverse techniques. Since the questions emanated from empirical tests, it is difficult for one to guess answers, thus enhancing convergent validity of the MMPI-2. MMPI-2 has many tests that make it impossible for patients to guess answers.
Since patents with a given disorder cannot guess answers, it means that their assessment using MMPI-2 gives measurements with high correlation, hence validate a given construct. Friedman, Lewak, and Nichols (2001) indicate that, developers of MMP-2 added new items and scales to MMP-1 to enhance clinical validity in its application since it has 567 items (p.11).
Extra items that MMPI-2 has compared to MMP-1 have significantly enhanced its convergent validity. Thus, developers constructed MMP-2 adding more items and scales obtained through convergent validation of empirical studies.
Developers of MMP-2 also employed divergent validation when they constructed MMPI-2 because they used numerous constructs and tests in the formulation of additional scales and items.
Given that divergent validation requires the use of different tests and different constructs in assessing a certain personality trait, developers of MMPI-2 utilized diverse items and constructs to evaluate patients with different mental disorders. Measurements from different patients and different tests provided additional items and scales that divergently validated MMPI-2 as a valid psychometric test.
According Friedman, Lewak, and Nichols (2001), MMPI-2 expanded its items and scales basing on empirical studies conducted on 2600 adults, who provided critical information that increased reliability and validity of measurements (p.12).
Hence, more than 2500 participants with varied constructs and more than 500 tests gave reliable items that formulated MMPI-2. Therefore, it is exceptionally difficult for MMPI-2 to give measurements that correlate because it has a remarkably high divergent validity.
Overall, developers of MMPI-2 employed convergent and divergent validations in construction of MMP-2 to enhance the validity of assessing personality. Convergent validity is applicable because MMPI-2 has 567 items that act as various tests, while measurement of patients with hypochondriasis represents a single construct.
Hence, consistency of measurements among hypochondriacs indicates that MMP-2 is valid because of significant correlation among measurements. Moreover, divergent validation is applicable in MMPI-2 because developers used 2600 participants as constructs while MMPI-2 has 567 items, which are tests that measure varied constructs.
Since divergent validation determines whether measurements of different tests and constructs have any correlation, MMPI-2 developers have employed diverse tests and constructs in the formulation of MMPI-2.
Friedman, Lewak, and Nichols (2001) assert that, modifications of MMPI have enhanced psychometric analysis of patients and many clinical practitioners have adopted MMPI-2 (p.10). Hence, convergent and divergent validations have enabled effective modification of MMPI-2.
Conclusion
Since personality is a complex construct, sociologists and psychologists have formulated diverse theories to elucidate it and constructed numerous tests to assess it. Despite of the fact that diverse theories and tests have tried to define and measure, there is still contention regarding validation of personality as different people come up with different measurements of personality.
In this view, Campbell and Fiske developed MTMM matrix to measure convergent and divergent validation of personality constructs. Convergent validity tests whether measurements of a construct using different tests correlate because high correlation means that they are valid.
On the other hand, divergent validity determines whether measurements obtained from different constructs using different tests have any correlation. Lack of correlation among measurements indicates that personality constructs measured are valid. Therefore, since the definition of personality is complex, convergent, and divergent validations have helped in construction of MMPI-2, which is an effective test of personality.
References
Campbell, D., & Fiske, D. (1959). Convergent and Discriminant Validation by the Multitrait-Multimethod Matrix. Psychological Bulletin, 56(2), 81-105.
Colendrino-Bucu, L., Guerrero, M., Pascula, B., & Mateo, R. (1993). Introduction to Psychology: A Textbook in General Psychology (2nd Ed.). Quezo City: Rex Bookstore.
Coon, D., & Mitterer, J. (2008). Introduction to Psychology: Gateways to Mind and Behavior. New York: Cengage Learning.
Friedman, A., Lewak. R., & Nichols, D. (2001). Psychological Assessment with the MMPI-2. New York: Routledge.
Raykov, T. (2008). Evaluation of Convergent and Discriminant Validity with Multitrait-Multimethod Correlations. British Journal of Mathematical and Statistical Psychology, 1-26.