Sentiment should exhibit its strongest effects on asset prices at times when valuations are most subjective. Consistent with this hypothesis, we show that a one-standard-deviation increase in aggregate uncertainty amplifies the predictive ability of sentiment for market returns by two to four times relative to when uncertainty is at its mean. We find similar evidence for the cross-section of returns; the predictive ability of sentiment for returns of test assets expected to be most sensitive to sentiment and for anomaly returns is substantially larger in times of higher uncertainty. The results hold for both daily and monthly proxies for sentiment and for various proxies for uncertainty. The evidence sheds light on one component of time-series variation in mispricing and suggests that the effects of sentiment are greatest in times of higher uncertainty.