Multi-collinearity

Multiple regression analysis is based on the assumption that the independent variables are not correlated with each other, whenever the independent variables are highly correlated with each other then it is very difficult to isolate the effect of each one of these on the dependent variables. This occurs when there is a simultaneous movement of two or more independent variables in the same direction and almost at the same time. This condition is called multi-collinearity.

We can use the correlation matrix to determine whether 2 independent variables are highly correlated. If a correlation value of more than 0.8 exists between two independent variables, then the problem of multi-colinearity is bound to occur. Alternatively if the correlation coefficient between the two variables is greater than the multiple correlation coefficients, then multi-colinearity problem will occur. To remove the problem of multi-colinearlity, we drop one of the correlated variables. You can drop any of the variables.