Demographic Changes in the US Elderly Population between 2009 and 2017: Implications for Transportation Planning and Transferability of Elderly Person Trip Generation Models

Abstract

US Decennial Census data show that the elderly segment of the US population has continued to grow in magnitude and share in the last few decades. Consequently, there has been increased attention on this population segment in transportation planning to ensure their future travel needs are adequately accommodated. Using data collected in two national household travel surveys conducted in 2009 and 2017, this research determines what changes occurred in selected demographic characteristics and selected travel characteristics of the elderly population between the two survey years and their respective implications for planning and to investigates the temporal transferability of linear regression models of trips generated by the elderly. The study findings include the following: (1) the elderly’s share of the US population grew by 22.41% in contrast to the 6.55% growth in the overall population, with the implication that trips generated by the elderly will be an important component of future travel on transport networks and thus must be explicitly considered in demand forecasting; (2) about 87% of trips they made in both survey years were by automobiles. Given the convenience of this mode, its share will continue to remain high; hence, the safety of the elderly as operators or passengers in automobiles will have to be a particularly important element of transportation planning; and (3) the purpose of their trips was found to be primarily shopping and social/recreational purposes with the majority of them made during daytime off-peak hours. Public transit levels of service are high during peak periods and low in the off-peak hours. Hence, transportation planning will have to consider increased specialized transit services that better accommodate the travel needs of the elderly and offer a good alternative to automobiles. The transferability analysis led to the following conclusions: (1) the trip-generation model parameters were not temporally stable; (2) the transferred model had 64.5% of the explanatory power of the local application context model, indicating that its forecasts can be used for planning purposes; and (3) working with a birth cohort enhanced transfer effectiveness—it led to the transferred model having 73.4% of the explanatory power of the local application context model. Thus, working with cohorts can be exploited to improve the accuracy of forecasts required in short- and medium-term planning analysis.