Adulthood Social Class and Union Interest: A First Test of a Theoretical Model

ABSTRACT

A serial mediation model of union interest was tested. Based on theoretical notes provided by Mellor and Golay (in press), adulthood social class was positioned as a predictor of willingness to join a labor union, with success/failure attributions at work and willingness to share work goals positioned as intervening variables. Data from U.S. nonunion employees (N = 560) suggested full mediation after effects were adjusted for childhood social class. In sequence, adulthood social class predicted success/failure attributions at work, success/failure attributions at work predicted willingness to share work goals, and willingness to share work goals predicted willingness to join. Implications for socioeconomic status (SES) research and union expansion are discussed.

keywords:

• Labor unions
• union psychology
• social class
• attribution theory

Acknowledgments

The reference to “we” in this article is real. Above and beyond receiving course credit, the following students should be acknowledged for their skilled survey work: Jocelyn Claudio, Matthew Croteau, Mario Moreno, Nathanael Park, Virginia Otero, Anne Rathey, Alexander Sorvillo, and Kenan Turkic.

Notes

¹ Mellor and Golay (in press) is an online journal article queued for print. It can be viewed or downloaded at http://link.springer.com/journal/12144.

² In regression terms, childhood social class can be labeled as a suppressor variable (Horst, 1941), a variable that is uncorrelated with the outcome variable (the criterion) but can nonetheless improve the prediction when added to the equation. This pattern occurs in cases where the added predictor is correlated with a predictor already in the equation. Labeling the added variable a suppressor reflects the supposition that it removes or suppresses criterion-irrelevant variance from the initial predictor. Regarding the same phenomenon in tests of sequential models, the added variable can be relabeled as an enhancer variable, in that inclusion of the variable improves estimation and clarifies the effect of the focal predictor on the outcome variable (MacKinnon, Krull, & Lockwood, 2000). Of course, suggesting suppositions and positioning focal predictors are matters of theory.

³ A table of χ² values for each subgroup comparing proportions in the sample and the population is available from the author.

⁴ We also ran independent t-tests based on employees grouped by whether they indicated a self-serving bias (attributing their success at work to internal factors and their failure at work to external factors vs. not; Miller & Ross, 1975). The analyses revealed that employees in the self-serving group (n = 224) indicated less willingness to join a union than employees in the non-self-serving group (n = 336), Ms = .46 and .54, respectively, t(558) = 2.132, ps < .05, d = .18. With employees grouped by whether they indicated the self-serving bias known as the fundamental attribution error (attributing their failure at work to external factors and their coworkers' failure to internal factors vs. not; Ross, 1977), the analyses revealed that employees in the attribution error group (n = 117) indicated less willingness to join than employees in the nonattribution error group (n = 443), Ms = .40 and .53, respectively, t(558) = 2.980, p < .01, d = .31.

Additional information

Notes on contributors

Steven Mellor

Steven Mellor is a PhD from Wayne State University who specializes in union-management psychology research. He is an Associate Professor in the Department of Psychological Sciences at the University of Connecticut.