This paper examines the role that public transport last mile problems play in mode choice decisions of commuters, while controlling for trip, built environment, and decision maker related variables. Last-mile problems arise due to lack of adequate connectivity between transit stops and trip origin or termination points. The paper is motivated by previous literature which has pointed out that high-quality public transit needs to consider end-to-end connectivity from trip origins to destinations. In contrast to previous work on transit last mile problems which has focused on physical distance and sidewalks to transit stops, we consider a wider range of area factors including transit availability, job accessibility, parking costs, the quality of the pedestrian environment and risks to pedestrians from vehicular traffic, and social characteristics such as street-level crime. Using a discrete choice model, our goal is to unpack ways in which such factors contribute to the last mile problem in home-based work trips, while controlling for these wider range of factors as well as the usual variables such as cost and trip time that inform mode choice. We find that the prevalence of non-domestic violent crimes reduces the odds of using all types of non-motorized alternatives as well as transit that is accessed either by walking or driving. Using compensating variation to measure welfare changes, we show that there are significant benefits that could be brought to transit service users Preprint submitted to Journal of Transport Geography June 9, 2016 through increasing safety in the transit access trip. By separately controlling for origin and destination transit accessibility, we show that improved destination accessibility significantly boosts transit use to a greater degree than increases in origin level accessibility. These findings argue for improving accessibility and related job densities at employment centers.