Whenever we construct a statistical model (such as a linear regression model) we come across two kinds of variables – endogenous variable and exogenous variables, depending on whether they are affected by the variables in our model or not.
The exogenous variables are not affected by the variables in the model, that is, they lie outside (“exo”) the model. Exogenous variables should not be confused with independent variables.
Examples of exogenous variables:
- When studying a regression model for predicting income, the eye colour of the individual is an exogenous variable. This is because the eye colour is not affected by the income or any factors which affect income.
- When studying the effect of IQ on marks obtained in a test, gender is an exogenous variable as intelligence and gender are unrelated to each other.
Exogenous variables can be identified by simply verifying that their values remain the same no matter how much the variables inside the model change.
The endogenous variables are affected by the variables in the model, that is, they lie inside (“endo”) the model. The dependent variable in a linear regression model is an example of an endogenous variable because it is affected by the independent variables in the model.
Examples of endogenous variables:
- When studying a regression model for predicting income based on education levels, the income is an endogenous variable. This is because the income is affected by the education level.
- When studying the relation between IQ and marks obtained in a test, the marks obtained are an example of endogenous variable because the marks obtained depend on the IQ of the student.
We know a variable is endogenous when we observe that the value of the variable changes according to the changes in the variables in our model.