How the discrepancy between expectation and outcome influences expectation change in relation to depressive symptoms - Results from four experimental studies.

BACKGROUND: Traditional associative learning models and their clinical applications suggest that the greater the discrepancy between expectations and outcome, the greater the adjustment of expectations. Other frameworks, however, propose a "tipping point", above which the discrepancy is so large that the source of information is considered unreliable, resulting in reduced expectation change.

METHODS: To test these competing predictions, we performed four experiments in which participants were provided with five levels of positivity/negativity in two different contexts: social interactions and performance feedback. We also examined how depressive symptoms were related to changes in expectations in response to different levels of positive/negative information.

RESULTS: The results regarding the processing of unexpectedly positive information (Study 1, N = 212; Study 2, N = 158) did not allow unanimous conclusions, but were more consistent with the tipping point hypothesis, as there was strong expectation change when information was moderately positive and a decline in expectation change when information was even more positive. For negative information (Study 3, N = 252; Study 4, N = 286), the results were more consistent with the linearity hypothesis, indicating that the more negative the information, the greater the change in expectations. Depressive symptoms interacted with the level of positive information in affecting expectation change: Depressive symptoms were associated with lower expectation change in response to mildly positive information, but not in response to highly positive information, particularly in Study 2.

CONCLUSIONS: The results suggest that the relationship between the magnitude of an expectation violation and expectation change may depend on the valence of the expectation violation.