Multiple regression and market research
N J S Clemens and P Duncan-Jones, International Journal of Market Research
This paper discussed multiple regression and its possible use in market research and marketing. It won the MRS Gold Medal in 1965.
This paper discussed multiple regression and its possible use in market research and marketing. It won the MRS Gold Medal in 1965. The first part of the paper describes the linear multiple regression equation, which relates a dependent variable to a number of independent (or predictor) variables. It has two uses: to forecast the effect of a change in one or more predictors, or to interpret what is going on as a guide to further action. It is pointed out that this involves an assumption about causality, which is never implicit in the regression model itself. Two books are recommended: 'Quantitative techniques in marketing analysis' (eds. Frank, Kuehn and Massey), and 'Diffusion of Innovations' (Everett Rogers). The second part describes a case history, a political opinion survey, in which Efroymson's computerised technique (described in an appendix) for selecting the most effective variables for multiple regression was used for the analysis.