A Note of the Use of the Shift Technique in the Measurement of Population Change

This paper examines the issue of modeling dynamic aspects of shopping trip-making behavior using time series data and presents the results of an empirical analysis based on a two-week travel diary survey of households in Hamilton, Ontario. It is concluded that logit models incorporating time dependent variables perform significantly better in terms of both calibrated goodness-of-fit and predictive capabilities than do models which assume no time dependency between choices.