Question

2. Some will claim that adding a variable to a model is justified as long as the adjusted R-square increases and the t-test on the added variable indicates significance, despite the theoretical validity of the added variable. The claim is that adding the variable under these conditions is justified and there are no adverse consequences on the other variables in the model. Do you agree with the claim? Explain why or why not.

Answer 1

Solution:

Adjusted R square is the measure of variance explained by the regression and also a statistically insignificant variables can add information's.

Adding the number of variables in the given model gives higher R square adjusted value. but the model is not always adequate.

For example- suppose for model with 3 variables we have a higher R square adjusted value but on variable become insignificant. And also for the model with 2 variable has a lower R adjusted but the two variables are significant. Hence the model is adequate.

Hence add only significant variables.