Unformatted text preview: he tones. Results and Discussion
No-load participants showed the standard behavior-induced attitude change; there was a greater difference between the mean rank of
the selected and rejected pairs in Phase 3 than there was in Phase 1,
t(14) 2.23, p .05, r .51. Our prediction that participants under a
cognitive load would also show attitude change was supported, t(14)
2.64, p .02, r .58. As shown in Table 3, nearly identical levels of
attitude change were evidenced by participants in the two conditions
in response to the counterattitudinal behavior. A comparison between
the two groups revealed no differences, t(28) 0.71, p .2, r .10.
As in Experiment 1, there was no attitude change in the noncritical set;
this was true for participants in both the no-load and the cognitiveload conditions, t(14)
.3, and t(14)
.3, respectively (see Table 3). 138 VOL. 12, NO. 2, MARCH 2001
Downloaded from pss.sagepub.com at COLUMBIA UNIV on April 18, 2013 PSYCHOLOGICAL SCIENCE M.D. Lieberman et al.
The cognitive-load manipulation did not attenuate attitude change,
but did it affect other kinds of processing? To answer this question, we
examined participants’ ability to identify which of the 15 prints were
presented in the choice phase. Because cognitive load impairs encoding (e.g., Fletcher, Shallice, & Dolan, 1998), we expected that load
would impair memory for the counterattitudinal behavior. This is exactly what we found: Subjects under cognitive load were able to identify only half as many prints (35%) as their no-load counterparts
.67. Though participants under
load were near chance in their identiﬁcation accuracy, they correctly
categorized as chosen or rejected those prints that they accurately
identiﬁed as often as did the no-load participants (67% vs. 73%, respectively), t(30) 0.03, p .2, r .01. In short, even though cognitive load impaired participants’ ability to think about their behavior,
they showed the same amount of attitude change as participants not
under cognitive load.
As in Experiment 1, we assessed the extent to which degree of attitude change correlated with explicit memory measures. For both the
cognitive-load and the no-load participants, there were no signiﬁcant
correlations between attitude change and either memory measure, all
ps .3. Conclusion
Regardless of cognitive-load condition, participants showed substantial attitude change in response to their counterattitudinal behavior. Attitudes changed despite the fact that participants under a
cognitive load were signiﬁcantly impaired in their recall of which
prints were involved in the choice phase of the experiment.
These data, combined with the results of Experiment 1, suggest
that the process of behavior-induced attitude change is a relatively automatic one (cf. Zanna & Aziza, 1976). It is possible that attitude
change requires some minimal amount of attention, conscious awareness, or mental effort, but it is fair to claim that unaltered performance
under cognitive load indicates that this attitude-change process is to- Table 3. Attitude chang...
View Full Document
This document was uploaded on 01/26/2014.
- Fall '13
- Cognitive Dissonance