Balance Theory Revisited: Relationship Issue Relevance Affects Imbalance-Induced Tension in Workplace Relationships

ABSTRACT

The present work applies and extends balance theory by examining the role of relevance of issue to the relationship in balance theory processes within the context of workplace relationships. In Experiment 1, a sample of working adults (N = 81) reported greater job tension when self-supervisor dissimilarity involved a relationship-relevant (vs. non-relationship) ethical dilemma. In Experiment 2, a sample of working students (N = 185) who perceived greater self-supervisor dissimilarity about workplace (vs. family) ethics reported greater job tension, and in turn, less job satisfaction and organizational commitment. Perceiving dissimilarity with a work supervisor in attitudes about relationship-relevant issues may negatively affect outcomes at work. Importantly, these experiments demonstrated that not all dissimilarity is likely to yield negative outcomes; only relationship-relevant (vs. non-relevant) dissimilarity was a catalyst for imbalance-induced tension.

KEYWORDS:

• Attitude similarity
• balance theory
• issue relevance
• job satisfaction
• organizational commitment
• tension

Acknowledgments

The authors thank Elizabeth Nobukuni, Julie Kittel, and Stephen Short for their assistance.

Funding

Online data collection was supported by an Institutional Grant from the Center for Clinical and Translational Research (CCTR) and VCU Technology Services. Views expressed herein are those of the authors and do not necessarily reflect those of the funding agencies.

Notes

¹ We again examined the model with the inclusion of sex to test for any possible interaction between sex and our two independent variables. Sex was not related to job tension on its own (p = .86) or in interaction with our independent variables (ps ranged from .07 to .30).

² To ensure that the relative direction of dissimilarity between individuals and supervisors was not an issue in our research, we examined the participants' reports of their perceived ethics compared to their perception of their supervisors' ethics. On the workplace ethics scale, participants rated themselves as being only marginally less unethical than they rated the supervisor, but this difference was not significant, t(179) = −1.74, p = .08, d = −0.26. Compared to their ratings of the supervisor, 38.90%, 14.40%, and 46.70% of participants rated themselves as being more unethical, equally unethical, and less unethical, respectively. Importantly, no significant differences emerged in level of perceived job tension among participants who reported being more unethical, less unethical, or equally ethical to their supervisors, F(2, 176) = 1.41, p = .25, η_p² = .02. On the family ethics scale, participants rated themselves as being only marginally less unethical than they rated the supervisor, but this difference was not significant, t(179) = −1.85, p = .07, d = −0.28. Compared to their ratings of the supervisor, 42.20% rated themselves as being more unethical, 7.80% rated themselves as being equally ethical, and 50% rated themselves as being less unethical. There were no significant differences in level of perceived job tension among participants who reported being more unethical, less unethical, or equally ethical to their supervisors, F(2, 176) = 0.15, p = .86, η_p² = .00. These findings indicate that our results are not merely a product of participants perceiving their supervisors as unethical.

³ We created parcels using the item-to-construct balancing technique, which pairs items possessing higher standardized factor loadings with items possessing lower standardized factor loadings. Utilizing parcels requires fewer model parameter estimates, increases the reliability of the indicators, and reduces sampling error (Little, Cunningham, Shahar, & Widaman, 2002).

⁴ We used difference scores because our research concerns dissimilarity between individuals and supervisors, and does not concern the relative direction of those differences nor does it concern change within the individual participant. Research has suggested that concern about the use of difference scores is often erroneously expressed irrespective of data analysis approach (Thomas & Zumbo, 2012) and that difference scores may even be preferred under certain circumstances (Kisbu-Sakarya, MacKinnon, & Aiken, 2013). Researchers concluded that the use of difference scores is appropriate when relevant to the subject matter and the analysis has enough power (Thomas & Zumbo, 2012). Our adequately powered analysis tests research questions that are primarily concerned with the relative distance between two individuals on a construct (self vs. supervisor workplace ethics and self vs. supervisor family ethics). The use of difference scores allows an appropriate and most direct method of examining our research questions.

⁵ We examined the same model, but included the variables of sex to control for its relationship with organizational commitment. Model fit was acceptable, χ² (49, n = 177) = 107.49, p <.001, RMSEA = .082_{(.061, .103)}, CFI = .95, TLI = .94, SRMR = .09. Sex significantly predicted organizational commitment, standardized coefficient = .11, p = .03. The patterns of significance and direction for all other direct and indirect paths from the original model remained the same. Thus, we chose to retain our original hypothesized model.

Additional information

Notes on contributors

Chelsea A. Reid

Chelsea A. Reid is an assistant professor in the Department of Psychology at the College of Charleston. She earned her Ph.D. in social psychology at Virginia Commonwealth University.

Jody L. Davis

Jody L. Davis is an associate professor in the Department of Psychology at Virginia Commonwealth University. She earned her Ph.D. in social psychology at the University of North Carolina, Chapel Hill.

Jeffrey M. Pollack

Jeffrey M. Pollack is an assistant professor of management in the Poole College of Management at North Carolina State University. He earned his Ph.D. in management at Virginia Commonwealth University.

Richard S. Coughlan

Richard S. Coughlan is an associate professor of management in the Robins School of Business at the University of Richmond. He earned his MBA and Ph.D. in management at the University of Arizona.