Please use this identifier to cite or link to this item:

Strengthening vs. weakening effects of inconsistencies in attribute information : a resolution of opposing perspectives

Authors Sengupta, Jaideep
Johar, Gita Venkataramani
Issue Date 2000-11-29
Summary This research examines the effects of evaluative inconsistencies in product attribute information on an important consequence of attitude strength - viz., the link between attitudes and behavior (as measured by behavioral intention). Drawing on the existing literature, which makes opposing predictions as to the effects of inconsistency, we construct a theoretical framework that allows us to specify conditions under which weakening or strengthening effects should be obtained. Specifically, we propose that a strengthening effect should result when exposure to inconsistencies leads to a process of reconciliation; however, a weakening effect should result when respondents are either not motivated to reconcile inconsistencies, or are hindered from doing so. This framework is collectively tested in two experiments. Experiment 1 varies the accessibility of attribute information under default motivation conditions (i.e., high motivation to reconcile). In support of the thesis that high (low) accessibility should facilitate (hinder) the process of reconciliation, inconsistency leads to an increase in the attitude-intention link under high accessibility, but weakens the link under low accessibility. Experiment 2 then keeps information accessibility high across conditions, and studies the effects of processing goal by making people either accountable or non-accountable for their judgments. Results show that when the default reconciliation motive is replaced by a regret minimization motive that is inimical to inconsistency reconciliation (e.g., under high accountability), a weakening effect is obtained even under high accessibility conditions. Taken together, results from these two experiments provide good support for the proposed framework, and bridge opposing theoretical perspectives on the effects of information inconsistency. Process measures obtained in the two experiments are also supportive of our overall conceptualization.
Language English
Format Working paper
Files in this item:
File Description Size Format
mktg00153.pdf 3873923 B Adobe PDF