confusing coefficient signs

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

confusing coefficient signs
Message posted by clark (via 142.106.180.133) on February 1, 2002 at 10:34 AM (ET)

could someone explain why one might get negative coefficients when you expected a positive coef. from a logistic regression analysis, i'm using spss...this occurs with only one predictor variable so it's not multicollinearity...

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

Re: confusing coefficient signs
Message posted by JG (via 128.8.22.11) on February 4, 2002 at 3:39 AM (ET)

Negative coefficients due to error - regression just an unconstrained optimization procedure.

Re: confusing coefficient signs
Message posted by clark (via 142.106.180.133) on February 5, 2002 at 1:16 PM (ET)

i'm sure you're right, i guess i was hoping for a more eloquent alternative explanation

thanks for your reply

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