Identifying transport preferences in a South African township: A case study of Soweto

Rapid urbanisation and apartheid-era land-use legacies leave many South Africans captive to long, costly, and unsafe commutes. Yet rigorous evidence on how township residents value transport attributes is scarce. This study addresses that gap with the face-to-face stated-preference survey dataset for travelling to work and higher education institutions in South Africa. Using an efficient best-worst design, 201 face-to-face interviews were conducted in Soweto, eliciting dual responses on cost, in-vehicle time, first/last-mile effort, comfort, and safety for both public and private transport options. Mixed-logit best-worst models capture unobserved taste heterogeneity and deliver the township's first mode-specific elasticities and willingness-to-pay measures. Results show no systematic behavioural differences in mode choice between workers and adult students. Instead, attribute valuations are dominated by safety concerns. The willingness to pay for a “very safe” service is over four times that for a one-hour reduction in travel time. Elasticities confirm that safety improvements have the greatest potential to shift demand, whereas marginal fare increases sharply deter low-income users. The findings stress that enhancing perceived and actual safety is a prerequisite for sustainable mode shift in South African townships. They also provide policy-ready parameters for appraisal of bus, rail and minibus-taxi upgrades in similarly segregated African cities.