Abstract
THIS ARTICLE IS CONCERNED WITH THE CONSTRUCTION OF A STATISTICAL MODEL OF TRIP FREQUENCY WHICH WOULD DESCRIBE THE DATA GENERATING PROCESS MORE COHERENTLY THAN THE CLASSICAL LINEAR REGRESSION MODEL (CLRM). A MORE DATA-COHERENT MODEL CAN BE OBTAINED BY A GENERALIZATION OF THE POISSON DISTRIBUTION, NAMELY THE NEGATIVE BINOMIAL DISTRIBUTION. AN APPLICATION OF THIS MODEL, TO THE ANALYSIS OF SHOPPING DATA FROM THE SUSSEX HOUSEHOLD SHOPPING SURVEY, IS REPORTED AND COMPARISON IS DRAWN WITH THE CORRESPONDING REGRESSION RESULTS.