Depressive Symptoms and Category Learning: A Preregistered Conceptual Replication Study.

We present a fully preregistered, high-powered conceptual replication of Experiment 1 by Smith, Tracy, and Murray (1993). They observed a cognitive deficit in people with elevated depressive symptoms in a task requiring flexible analytic processing and deliberate hypothesis testing, but no deficit in a task assumed to require more automatic, holistic processing. Specifically, they found that individuals with depressive symptoms showed impaired performance on a criterial-attribute classification task, requiring flexible analysis of the attributes and deliberate hypothesis testing, but not on a family-resemblance classification task, assumed to rely on holistic processing. While deficits in tasks requiring flexible hypothesis testing are commonly observed in people diagnosed with a major depressive disorder, these deficits are much less commonly observed in people with merely elevated depressive symptoms, and therefore Smith et al.'s (1993) finding deserves further scrutiny. We observed no deficit in performance on the criterial-attribute task in people with above average depressive symptoms. Rather, we found a similar difference in performance on the criterial-attribute versus family-resemblance task between people with high and low depressive symptoms. The absence of a deficit in people with elevated depressive symptoms is consistent with previous findings focusing on different tasks.