Abstract
Prior research on the use of stereotypes in social judgments has shown that whether young adults make stereotype-consistent or -inconsistent judgments depends in part upon the response scale that is used. This shifting standards effect in stereotype use was examined in the present study to determine whether older adults, who tend to rely on stereotypes more than younger adults, would also show a similar effect. Young and older adults evaluated the height of male and female targets using either an objective or subjective scale. No age differences were found, with both age groups producing stereotype-consistent judgments (i.e., men are taller than women) on an objective scale, but stereotype-inconsistent judgments (i.e., men and women are equally tall) on a subjective scale. These results suggest that the shifting standards effect holds across the adult life span.
Acknowledgments
This study was conducted as part of Carolyn Hoessler's honors thesis and was supported by a grant from the Social Sciences and Humanities Research Council of Canada awarded to Alison Chasteen. The authors thank Monica Biernat for supplying the target photographs and Dominic Packer and Sonia Kang for comments on an earlier draft of this article.
Notes
1It should be noted that the stereotype-inconsistent judgments produced by subjective scales can take the form of either no difference between groups or a pattern opposite to the stereotype (Biernat, Citation2003).
2For the education measure the choices were (1) elementary school, (2) grade 10, (3) high school diploma, (4) undergraduate degree, (5) master's degree, (6) doctoral degree.
3Participants in the objective condition were permitted to choose which units (feet and inches or centimeters) they would prefer to use to estimate height. Any estimates made in centimeters were converted to inches.
Note. Values in parentheses are standard deviations. Means with matching subscripts differ significantly, p < .001.
4It was surprising that there was a main effect for response scale, given that the scores were standardized within scale condition. We suspect this difference occurred due to rounding procedures in the statistical package and due to the large number of participants per response scale condition (cell Ns = ∼ 50). Indeed, the means of the z scores were nearly 0 (subjective scale M = .001, objective scale M = − .011).