Self as a default target in thinking about traits

Abstract

Two experiments provide evidence for the recruitment of self as a default target in thinking about traits. Results show that semantic processing of trait labels, when no specific target is named, facilitates subsequent trait judgments more for self than for other social exemplars, including exemplars of highly familiar others. The effect is not qualified by trait desirability. However, as predicted, such a privileged position of self in thinking about traits, applies to thinking about covert, unobservable, manifestations but not to thinking about overt, observable, manifestations of traits.

Notes

1. In using response latency as an indicator of accessibility and/or cognitive processing researchers typically rely on dichotomous responses rather than on a 5-point Likert-type scales (however, see Cucina & McElreath, 2005; Fazio & Williams, 1986; Huston & Fazio, 1989; Karylowski, 2005; Karylowski et al., 2000; Lischetzke, Cuccodoro, Gauger, Todeschini, & Eid, 2005; Staats & Skowronski, 1992, for some exceptions).The dichotomous scale is often preferred because it reduces the error variance of response latencies by minimizing the contribution of time required to locate and to press the appropriate response key. Our decision to use a 5-point response scale was dictated mainly by the desire to make our data more compatible with results of other studies (Karylowski, 2005; Karylowski et al., 2000). Those studies, which provided evidence against the notion of self as a default point of reference in judgments of others, used a 5-point scale. See Fazio (1990) for an argument on acceptability of using latency measures with a Likert-type scale.

2. This was done to reduce the possibility of participants anticipating the trait-judgment question while answering the semantic priming question. Because, correct answers to subtraction questions were always integers ranging from 1 to 5, the same five numerical keys could be used to enter responses to those questions.

3. This was done to reduce positive skew in latency data, as recommended by Winer (1971, p. 200). To facilitate presentation, throughout the article mean latencies are reported in milliseconds.

4. Because of missing data, this analysis was conducted for 48 participants only.

5. In addition, the main effect of desirability was significant, F(1, 47) = 4.60, p < .05, η² = .09, which was due to longer latencies for undesirable (M = 2942 ms) than for desirable traits (M = 2844). Finally, as in the main analysis, the main effect of target was significant, F(3, 141) = 7.34, p < .002, η² = .14.

6. The actual wording of instructions describing the distinction between unobservable (“feels”) and observable (“looks”) characteristics varied across participants. For half of participants, observable characteristics were always mentioned first, for the other half, unobservable characteristics were always mentioned first. This was done for exploratory purposes and to ensure that the results were not contingent upon such order. Preliminary analysis revealed that this variation in wording had no effects on our data and it was not included in the main analysis.

7. An informal examination of those descriptions did not identify any problems in understanding the distinction.

8. Because adding trait valence as the fourth within-subject variable would result in a large number of missing values, a question of possible modifying effect of trait valence on the interaction between priming, modality, and target could not be addressed. However, we conducted a 2 (Trait Priming) × 6 (Target) × 2 (Trait Valence) × 24 (Counterbalancing) mixed-model ANOVA across the two levels of the modality variable. This analysis confirmed significant Trait Priming × Target interaction, F(5, 360) = 5.58, p < .001, η² = .07, and revealed no significant three-way interaction, F(5, 360) < 1, thus confirming conclusion from Experiment 1 that interaction between priming and target did not depend on trait valence.

9. See Karylowski et al. (2000) for a more in-depth treatment of this issue.