We investigated the potential impact of a cohort traumatic exposure, the Troubles in Northern Ireland, on memory functioning in later life, and the potential moderating effect of social activity engagement. Using data from 6571 participants aged 60 + in the Northern Ireland Cohort for the Longitudinal Study of Ageing (NICOLA) cohort, we used a structural equation modelling framework to explore associations between traumatic exposure during the Troubles and memory functioning. As expected, social activity engagement was positively associated with memory functioning, β = .102. Traumatic exposure was also positively associated with memory functioning, β = .053. This association was stronger at low levels of social activity engagement; among those with higher levels, there was little association, interaction β = − 0.054. The positive association between traumatic exposure during the Troubles and memory functioning was not moderated by the age at which the exposures occurred (based on analysis of a subsample with available data), interaction β = − 0.015. We conclude that superior memory functioning was associated with higher levels of traumatic exposure during the Troubles, particularly among those with lower levels of social activity engagement, and regardless of the age at which the exposures occurred. Future longitudinal analyses are required to build on these results, which potentially have implications for life-course epidemiology, in relation to critical periods for traumatising experiences.