Poor Expression: Concealing Social Class Stigma

Abstract

Three studies found that people from lower-class backgrounds are less expressive toward interaction partners from upper-class backgrounds except in contexts where they share minority status on another dimension. In Study 1, White lower-class dyad members behaved less expressively than their upper-class interaction partners, while in African American and Latino dyads, upper- and lower-class individuals were similarly expressive. In Study 2, lower-class White participants reported feeling generally less comfortable about interacting with an upper-class partner but not when they shared numerical minority status of being residents of the same state traveling away from home. Finally, Study 3 revealed that lower-class individuals intentionally act differently with upper-class individuals but not with lower-class ones. Upper-class individuals act the same with lower- and upper-class partners alike.

ACKNOWLEDGMENTS

We thank John Darley, Susan Fiske, Dale Miller, and Phil Zimbardo for helpful comments, Heng Li for statistical assistance, and Bryan Harrison, Mitch Meyle, Eric Provins, Alex Radetsky, and Irina Yudovich for data collection assistance.

Notes

¹Some participants later reported income between $30,000 and $80,000.

^a Parental education was measured on a 7-point scale on which 1 = elementary school, 2 = some high school, 3 = high school graduate, 4 = some college, 5 = college graduate, 6 = some graduate school, and 7 = graduate or professional degree.

²For data with a round robin structure, we cannot assume that dyads within a block will have homogenous variance and correlation. We applied Li's procedure (1995; Li, Hallahan, & Rosenthal, 1997) to decompose a round robin block's error variance into these two distinct components. For these data, the within-block error variance attributable to dyads that do not share a common member was similar to the between-block error variance, so these two sources of variance were pooled (Green & Tukey, 1960) to serve as the error term for hypothesis tests involving the fixed between-block factors race and income. For more information, contact Mark Hallahan ([email protected]).