We monitored chimpanzee welfare during the introduction of on-exhibit cognitive research training and testing, as measured by behavior and interest in such training, and related individual variation to personality assessments. We observed 11 chimpanzees (six males; five females) over a 16-month period and compared their behavior across three conditions: (1) Baseline (nontraining/research situations) and (2) an on-going, off-exhibit program of Husbandry Training and (3) Research Pod Activities, on-exhibit, group training for cognitive testing. There was considerable individual variation in their interest levels during research sessions; females and those scoring higher for Openness were present more frequently (including those who actively participated and those who observed others participating), but interest did not vary in relation to rates of self-directed behaviors (SDBs), rank, or the level of social disruptions within the group (i.e. large-scale displays or fights). The frequency of SDBs was predicted by the Neuroticism personality factor, but did not differ across baseline and training contexts, indicating that these activities do not negatively impact welfare. We also explored vigilance as an indicator of social uncertainty, but social monitoring did not differ in relation to either social context or rank. Finally, we explored how the specific characteristics of the research context impacted on SDBs; namely, social context, reward contingency, and visual access to keepers. SDBs increased only when visual access to keepers was restricted, suggesting that visual contact reduced uncertainty in novel training contexts. Overall, the introduction of a cognitive research program did not compromise welfare, and the chimpanzees’ repeated interest and willingness to participate suggests that the research was enriching. Am. J. Primatol. 74:828-840, 2012. © 2012 Wiley Periodicals, Inc.