Skip to main content
Cross Validated

Return to Answer

added 96 characters in body
Source Link
DWin
DWin
  • 7.8k
  • 23
  • 35

The collinearity is not between those two variables. They both have coefficients. Rather it is a joint collinearity between the two variables and the interaction variable. The interaction variable is the one that has the NA coefficient.

Collinearity does not need to be between only two variable but can exist between three or more variables. If any number of variables can exactly predict another variable in the model, then multi-collinearity exists and the predicted variable is given an NA by the regression function. ChangingThe process of dropping coefficients from consideration is called “aliasing” by the R authors. Changing to logistic regression will not solve this problem.

The collinearity is not between those two variables. They both have coefficients. Rather it is a joint collinearity between the two variables and the interaction variable. The interaction variable is the one that has the NA coefficient.

Collinearity does not need to be between only two variable but can exist between three or more variables. If any number of variables can exactly predict another variable in the model, then multi-collinearity exists and the predicted variable is given an NA by the regression function. Changing to logistic regression will not solve this problem.

The collinearity is not between those two variables. They both have coefficients. Rather it is a joint collinearity between the two variables and the interaction variable. The interaction variable is the one that has the NA coefficient.

Collinearity does not need to be between only two variable but can exist between three or more variables. If any number of variables can exactly predict another variable in the model, then multi-collinearity exists and the predicted variable is given an NA by the regression function. The process of dropping coefficients from consideration is called “aliasing” by the R authors. Changing to logistic regression will not solve this problem.

added 350 characters in body
Source Link
DWin
DWin
  • 7.8k
  • 23
  • 35

The collinearity is not between those two variables. They both have coefficients. Rather it is a joint collinearity between the two variables and the interaction variable. The interaction variable is the one that has the NA coefficient.

Collinearity does not need to be between only two variable but can exist between three or more variables. If any number of variables can exactly predict another variable in the model, then multi-collinearity exists and the predicted variable is given an NA by the regression function. Changing to logistic regression will not solve this problem.

The collinearity is not between those two variables. They both have coefficients. Rather it is a joint collinearity between the two variables and the interaction variable. The interaction variable is the one that has the NA coefficient.

The collinearity is not between those two variables. They both have coefficients. Rather it is a joint collinearity between the two variables and the interaction variable. The interaction variable is the one that has the NA coefficient.

Collinearity does not need to be between only two variable but can exist between three or more variables. If any number of variables can exactly predict another variable in the model, then multi-collinearity exists and the predicted variable is given an NA by the regression function. Changing to logistic regression will not solve this problem.

Source Link
DWin
DWin
  • 7.8k
  • 23
  • 35

The collinearity is not between those two variables. They both have coefficients. Rather it is a joint collinearity between the two variables and the interaction variable. The interaction variable is the one that has the NA coefficient.