Table 4.3 shows the independent variables, estimated coefficients, and summary
statistics for McCarthy’s study. The independent variables included measures of
automobile costs (price and operating cost per mile), physical characteristics and
vehicle style (horsepower, length, government crash test results, and vehicle type),
quality (results of a consumer satisfaction index), manufacturer, consumer search
activities (the number of first and second visits to different dealers and whether
the consumer repurchased the same brand as previously), and household socioeconomic
data. The study used multinomial logit analysis, a special form of regression
analysis in which the dependent variable is a discrete variable (to purchase or
not purchase a specific vehicle) rather than a continuous variable (the number of
vehicles purchased).