Analysis with dummy variables (dummy variable regression) is carried out if the independent variable to be analyzed has a categorical data type. While the dependent variable is not in the form of a category, if the dependent variable is in the form of a category, then the analysis technique used is logistic regression analysis.

Suppose we want to know what demographic variables of entrepreneurs are able to predict business turnover. Our research design is as follows.

• Dependent variable: turnover
• Independent variables: gender and education

We have two independent variables, namely gender (X1) and education (X2). Gender and ethnicity are categorical variables. Gender consists of two categories, namely male and female, while education consists of four categories, namely SMA/SMK and below, Diploma I/II/III, Diploma IV/S1, and S2/S3. Because gender consists of 2 categories, the gender variable only has one dummy variable (2-1), while education consists of 4 categories, so the education variable has 3 dummy variables (4-1).

Code dummy variables
As mentioned above, the number of categories is determined by the formula k-1, while the codes given are 1 and 0. All respondents who are members of the category are coded 1 while respondents who are not in that category are coded from 0. In this way, each respondent will have a code of 1 in the category that corresponds to it and a code of 0 in the category that does not fit it.
Further data processing is the same as simple or multiple linear regression.