Chapter 3: Multiple Regression and Curve Estimation

The most important advantage of multiple regression is that it shows the effect of each variable individually by keeping the other variables under control, as well as showing the collective effect of the independent variables.

Multiple regression is best applied when the dependent and independent variables are normally or nearly normally distributed. Another assumption of multiple regression is that there should be no multiple correlations between independent variables. Categorical variables can be used in multiple regression with the dummy coding technique. It is also assumed that the effects of the independent variables are linear.

Let’s assume that the motivation of 20 candidates who were interviewed to get a job as marketing personnel in the marketing department of an enterprise, their excitement levels in front of the jury, and their success in the exam (interview) were measured, and these scores are given as in Table 3.1. For motivation, scores of 0–15 were obtained, with “0” indicating “not motivated at all” and “15” indicating “highest level of motivation.” Exam success scores are also given between 0 and 20, and “0” indicates “not successful at all.” In the variable “excitement in front of the jury,” which scores between 0 and 10, “0” indicates “not at all excited.”