Question

With multiple regression, the main focus is on variables that are significant within the model and contribute to the variation occurring on the dependent variable. When multiple variances within the model are insignificant, then the reliability of the model is reduced. Therefore, we can not depend on the model for future reference. In this analysis, the dependent variable is ethical behavior that can be determined by the course taken, age, gender, and personality character of an individual. This model can be improved by considering the religious background, behavior at work, racial categories and marital status. Age, gender, and other variables are measurable (are quantitative or have codes associated with the qualitative variables). How would you measure/quantify/qualify religious background and behavior at work?

Answer 1

As we can see, religious background and behavior at work are non-numerical variables i.e., they are categorical variables. That means they are in the format of categories i.e., for example, take religion, Christians is one category, Hindu is the second category and Muslim is another category so on...While doing regression, we can't keep these categorical variables as it is, we need to create dummy variables for different categories in the variables. And each dummy variable takes either 0 or 1 as value.

For example, in the example of religion, we need to create a separate dummy variable for Christians and the values it can take are either 0 or 1, similarly Hindu and Muslims.

Therefore, in this way we need to deal with categorical variables such as religion or behavior etc..