Implicit and Explicit Age Stereotypes for Specific Life Domains Across the Life Span: Distinct Patterns and Age Group Differences

Abstract

Background/Study Context: Drawing on research that shows the importance of age stereotypes across the life span, the authors investigated domain-specific implicit and explicit age stereotypes in different age groups.

Methods: Implicit (Implicit Association Test [IAT]; Greenwald, McGhee, & Schwartz, 1998, Journal of Personality and Social Psychology, 74, 1464–1480) and explicit age stereotypes were assessed for the domains of family and health in a sample of N = 90 younger, middle-aged, and older adults.

Results: Overall, age stereotypes were negative for the health domain but not for the family domain. Distinct patterns of age group differences emerged depending on domain and assessment method. In the family domain, older participants held the least positive explicit age stereotypes, whereas implicit stereotypes in this domain were most positive for this age group compared with the young and middle-aged groups. For the health domain, implicit and explicit age associations indicated that middle-aged participants showed the most negative age-associations.

Conclusions: These findings suggest that implicit and explicit age stereotypes in different life domains represent largely independent constructs. Differential age group effects are assumed to reflect the result of accommodative and assimilative processes that are used to cope with age-related changes. Implications for future studies of implicit and explicit age stereotypes and their influence on developmental regulation are discussed.

Notes

¹ Previous research has demonstrated that IAT effects reflect relations between category labels rather than between the exemplar stimuli that are used to represent the categories (e.g., De Houwer, 2003; for the old/young IAT, see Gast & Rothermund, 2010); therefore, implicit and explicit measures of our study are entirely matched with regard to their content.

² Some previous studies used pictures instead of names of older persons as target stimuli (e.g., Hummert et al., 2002). We decided to use names, because drawing attribute and target stimuli from the same domain usually yields more robust IAT effects (Chang & Mitchell, 2011). However, since both names and faces typically reveal highly similar patterns of results in IATs (Rudman, 2011), employing faces instead of names would probably have yielded an identical pattern of results in our study. Moreover, since the exemplar stimuli for the categories old and young were identical in all IATs that we used, differences between age groups and domains cannot be due to the choice of exemplar stimuli.

³ Due to the repeated measures design that was employed in our study, it is unlikely that the domain differences in age gradients reflect an artificial inflation of effects due to age-related slowing. Any such influence should have affected both IAT effects in a similar way and thus cannot explain differences in the age gradients for the two measures. In order to completely eliminate influences of age-related slowing in our findings, we conducted an additional analysis that controlled for individual differences in average RT by dividing each single RT by the slowing factor of the respective participant (relation of an individual’s mean RT to the grand mean of all RTs in the complete sample). This procedure eliminates all differences in average RTs between participants and—by implication—between age groups, and thus efficiently controls for all effects that are due to general age-related slowing. Applying this correction procedure yielded a highly similar pattern of results as for the standard analyses; specifically, the Age Group × Explicit/Implicit interaction was significant for the family domain (F(2, 87) = 10.53, p < .001) but not for the health domain (F(2, 87) = 2.57, p = .08). For the health domain, a main effect of age group was found that indicated a quadratic trend across both measures (t(87) = 1.90, p < .05 [one-tailed]). For the family domain, we observed a linear decrease in implicit age stereotypes across the age groups (t(87) = 3.34, p = .001).

Additional information

Funding

This work was supported by a grant of the VolkswagenStiftung (AZ 86 758) to Klaus Rothermund.