In the field of transportation, several studies have researched commute mode choice and its dynamics in the short and in the long term. Relatively less is known about mode choice for discretionary and more flexible activities, such as social visits. These choices are dynamic and depend on personal habits and preferences, reflected to some extent in the history of similar choices. This study adopts the theory of path dependence to take life cycle dynamics and habitual preferences into account. Using a dataset collected in the Netherlands in 2011, a multinomial logit model of mode choice was developed. Results suggest that mode choices for social activities are path dependent, yet not entirely. There is also evidence of switching towards faster and more flexible modes after a life cycle event.