Investigating Drivers’ Variance in Route Choice between the Home and Work Commute Trip in the AM and PM

The study takes place in the Metropolitan Transport Area of East London, South Africa, a coastal city with a population of 530 000 spread over an area of 210 km 2. Like Auckland, development of its transport network is severely constrained by physiographic conditions such as steep terrain and 4 large river valleys and environmental /social constraints. A questionnaire survey was undertaken to determine the variance amongst drivers route choice by comparing the AM to the PM Home and Work commute trip. The study set out to determine this variance by measurement of the spatial variance and the variance in route choice factors, which drivers apply when making these two types of trip. Parametric and non-parametric statistical analyses were used to measure the difference in the spatial values between the trips. The variance in route choice factors was compared with other explanatory variables to determine if the responses were measuring the same or underlying values. The data was further examined to determine if factors such as socio-demographic and personal characteristics influenced the route choices.The study found that 25 % of drivers significantly varied (<50% of the route is common) their route between the AM and PM. Time is still considered an important factor, but the unexplained 24 minutes of residual time found between arrival at work and the official work start time, is inconclusive as to how time really affects the route choices. Factors such as congestion and predictability of journey time which could influence departure and journey times appear to be a non-issue. Other factors such as socioeconomic and personal characteristics influence the spatial variance of route choice. Route choice reasons appear to be duplicate for the AM and PM trips, but 45 % of the whole sample of drivers use different routes for the AM and PM, with the amount of overlap of the routes ranging from 0% to 95%. This casts doubt on the current method of traffic modelling in South Africa and indeed most part of the world which predicts the AM conditions and then reverses the flow for the PM conditions. This method leads to an over supply of expensive infrastructure. Despite the use of different routes, journey times and distances are similar for the two spatially different trips. The spatial variance between the two routes is best expressed as a percentage of route commonality and the physical lateral separation.