1
$\begingroup$

I have created a multiple regression model that checks if final result is explained by following predictor variables:

Time spent studying Class attendance (%) Cash spend on extra materials and classes etc.

Now I have a group of 50 boys and a group of 50 girls. I run my model for each gender independently. Can I compare coefficient of (for example) Class attendance for boys with coefficient of same variable for girls and say:

"Boys final results are less dependent on their class attendance than girls?"

If not how can I compare them?

$\endgroup$

1 Answer 1

2
$\begingroup$

Put your 100 children's data together. Add one variable: Gender = 1 for male, = 0 for female. Call this variable $Z$.

Fit the following model: $$Y = \beta_0 + \beta_1 X_1 ...+\beta_k X_k + \alpha_0 Z + \alpha_1 ZX_1 ...+\alpha_k ZX_k +\epsilon$$ where $X_1,...,X_k$ are Time spent studying Class attendance (%) Cash spend on extra materials and classes etc..

Then you can check the $\alpha$s to see is there difference between male and female on their results on $X$. For example if $X_1$ is Class attendance, then $\alpha_1$ will tell you the difference of the effect of Class attendance on the results between male and female.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.