The authors hypothesized that experienced and imagined intergenerational contact should improve older people's math test performance under stereotype threat. In Experiment 1 (N = 51, mean age = 69 years), positive prior contact with grandchildren eliminated stereotype threat, which was mediated partially by reduced test-related anxiety. In Experiment 2 (N = 84, mean age = 72 years), the effect of threat on performance was significantly improved when participants merely imagined intergenerational contact, a situation again mediated by reduced anxiety. Previous research established that intergroup contact improves intergroup attitudes. The findings show that intergroup (intergenerational) contact also provides a defense against stereotype threat.