Introduction
Previous submissions covered the formulation of a two-stage, mixed methods approach to add to the body of knowledge about the traits and antecedent’s characteristic of workplace bullies. The initial quantitative stage will employ the Workplace Aggression Research Questionnaire (WARQ) to define the prevalence of bullying in a diverse range of organizations in some cities. The second stage, employing in-depth interviews, will generate the insights needed to add to the body of knowledge about the personal characteristics and self-reported antecedents of bullies.
Two-Stage Data Analysis
The WARQ is really a series of ordinal variables for employing incidence frequency as the screening criterion for bullying experience: “no,” “sometimes” and “regularly”. Given the random sample and natural order in the ordinal scale, this variable type is ordinarily consistent with articulating a stage 1 hypotheses like “Is the observed distribution of bullying prevalence different across organization type, level, gender, educational attainment, type of upbringing, domestic situation and self-described physique?” Statistical analysis would then be performed employing the non-parametric Kolmogorov-Smirnov test.
On the other hand, habitual bullying is the true criterion variable in the study. Since only “regular” experience of inflicting or reaping bullying matters for the purposes of this study, the data can be transformed to considering just the proportion across sub-samples that meet this criterion. In that case, data analysis will be based on t tests for the significance of differences in the proportion of regular bullying across the analytical sub-groups enumerated above.
Since the original WARQ comprised 33 items and it is possible to derive a composite score that corresponds to having experienced at least one-third of the situations on a regular basis, the primary report format for stage 1 will likely comprise 34 tables multiplied by the 7 classificatory variables listed at the top of page one.
In the qualitative stage, the plan covers both the standard mechanics for synthesizing verbatim responses and a preliminary set of constructs around which to formulate both the interview guide and data analysis. Broadly, data analysis consist of transcribing the tapes into word-processing files and subjecting the verbatim transcripts to content analysis. The transcripts then have to be coded, either manually or with content analysis software (Shank, 2005).
After the interviews themselves, coding is the single most critical activity in all-qualitative research that generates text or verbatim data. Coding is a two-step process of categorizing the data into meaningful groups and assigning each an arbitrary numeric value for ease of encoding and tallying. The work of reviewing and creating categories is tedious and time-consuming, even when spread among several knowledgeable assistants, because the interview questions will deliberately be left open-ended in order to seem more innocuous and generate the widest possible range of character traits or antecedents (Patton, 2002; Schram, 2005).
The four steps in the coding process
- List all responses – Feasible in this study since it is likely that the respondents for the depth interviews will be a manageable subset of the 30 to 60 who might respond to the postal survey.
- Consolidate responses – This is the process of grouping the disparate descriptions of self-described traits and upbringing background into categories that are homogenous (or nearly so) and completely span the range of responses but are distinct from each other. No two coders will make exactly the same judgments but this is a given in qualitative studies.
- Set codes – This means simply assigning numeric codes to the entire consolidated list of responses. For convenience, it is better to sequence, say, all categories about task-oriented supervision and teamwork into one set of numbers and those categories about showing toughness and assertiveness to maintain standing in the workplace “pecking order” into another number sequence. One concedes that subjectivity once again characterizes the ordering of the entire list. However, the order of categories does not matter at this point because tabulation will reveal which ones are more prevalent or correlate well with high incidence of bullying.
- Data entry – This involves the related stages of:
- Reading the individual responses from the transcripts
- Interpreting or matching each individual response with the finished list of categories
- Retrieving the numeric code for the relevant, most similar category
- Recording the appropriate code in the transcript or respondent record form
- Entering the respondent ID number plus all response codes into a spreadsheet or statistical software package.
External validity of the above principally manual coding procedure can of course be enhanced by employing dual judges or the researcher may well facilitate processing what promises to be a wealth of self-reports from respondents by resorting to computer-aided coding programs like General Inquirer or Text Pack.
In any case, the result is a database that can be ranked by frequency of response or cross analyzed against the aforementioned seven independent variables.
The working guide of constructs applicable to the research goal of this study is provided by prior research into bully personalities, some concededly dated and virtually all carried out in European settings that uncovered the link with aggression. In addition, it has been observed at one time or another that bullies tend to be power-driven, abusive, bigoted, abrasive to the point of being authoritarian, anti-social, and even so violent as to be sadistic. Save for those who verge on psychotic syndromes as revealed with the use of personality inventories, these traits will alert the researcher to precisely the Type A personalities who are noted for getting results and hence, being promoted to positions of power over others (Seigne, Coyne, Randall and Parker, 2007).
Conclusion
To this point, therefore, the research plan has defined the data analysis and reporting parameters that will hold for both the quantitative and qualitative stages of the study.
References
Patton, M. Q. (2002), Qualitative research and evaluation methods (3rd Ed.) London: Sage.
Schram, T. H. (2005). Conceptualizing and proposing qualitative research. Upper Saddle River (NJ): Pearson Merrill Prentice Hall.
Seigne, E., Coyne, I., Randall, P. & Parker, J. (2007). Personality traits of bullies as a contributory factor in workplace bullying: An exploratory study. International Journal of Organization Theory and Behavior, 10 (1); 118-132.
Shank, G. D. (2005). Qualitative research: A personal skills approach (2nd Ed).