Image Induction and Social Influence: Explication and Initial Tests

Abstract

Image induction posits that social influencers can manipulate a target's self-image to gain increased compliance. By preceding a request with a question that creates a salient self-image in a target, targets are more likely to behave consistently with that image by complying with the request. Three studies provide empirical tests of image induction. Results indicate that the addition of an image-inducing pretreatment question is associated with increased rates of compliance when compared to a control group. These findings held across two types of image induction manipulations (helpfulness, willingness to try new things), two types of behaviors (agreeing to complete a survey, agreeing to receive a free soft drink sample), and both face-to-face and written requests.

ACKNOWLEDGMENTS

We give special thanks for the advice and guidance received from the editor and the reviewers. The changes made to the manuscript as a result of the editorial process were invaluable.

Notes

¹Due to limitations of nominal level data, information in the current study was ascertained via the contingency coefficient to determine the strength of association between variables. This coefficient can be somewhat harder to interpret however because its upper limit is less than 1. More specifically, in a 2 × 2 table the upper limit of C is 0.707 (Blalock, 1979). Blalock stated that to correct for this difficulty, “the obtained C should be divided by 0.707” (p. 305).

²In Study 3, participants were not used if they answered no to the initial image inducing question. As was aptly noted by the reviewers, eliminating these participants from the data could have stacked the deck in favor of our hypotheses. Although data from these participants were not retained, we know that the approximate number of participants who answered no to the initial image inducing questions were below the required amount necessary to threaten the significance of our findings. Supplemental analyses suggest that adding 12 participants who said no to being willing to help to the experimental group would still produce a significant chi-square, whereas adding 13 participants to this group would lead to a nonsignificant chi-square. Similarly, adding 9 participants who said no to information about a new brand of soda to the experimental portion of Group B would still produce a significant chi-square. However, adding 10 participants who said no to information about a new brand of soda in the experimental portion of Group B would lead to a nonsignificant chi-square. In each of the image-inducing manipulations no more than 5 participants answered no to the initial question.