logistic regression question

Home > Statistics Every Writer Should Know > The Stats Board > Discusssion

logistic regression question
Message posted by Tina Hong on January 3, 2000 at 12:00 AM (ET)

Hello, My question is: Can you determine how much each independent variable uniquely contributes to the model using logistic regression? Any response or information including web links would be greatly appreciated.

Thanks in advance

READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: logistic regression question
Message posted by Jack Tomsky on January 4, 2000 at 12:00 AM (ET)

There are different parametrizations of a logistic regression. Let Y be either 0 or 1 and suppose the model is of the form, Y = z(X1, X2, ..., Xn)/[1+z(X1, X2, ..., Xn)], where z(X1, X2, ..., Xn) = exp(sum(ai*Xi)).

Then the contribution by the i-th factor is the sum of squares of the n Xi's about Xibar, times the square of the estimated coefficient ai.

Your $5 contribution helps cover part the $500 annual cost of keeping this site online.