2
$\begingroup$

I have performed a binary logistic regression with re-contracted as the DV (not re-contracted =0, re-contracted =1).

Team ladder position at the end of the year is a significant predictor. This varible is a rank (1-18) with 1 being "top" of the ladder and "18" being the bottom. The regression coefficient for this variable is -0.044 and the exponentiated coefficient is 0.957.

I am confused by the interpretation given that that a one unit increase in the predictor variable would actually be a worse result e.g. a higher number is toward the bottom of the ladder. Does this then mean that "a one unit increase in ladder position (towards the bottom) reduces the probability of being re-contracted by -0.044? and thus, the odds of being re-contracted are reduced by 0.957 for each one unit increase in ladder position (towards the bottom of the ladder)?

I'm going round in circles! have I interpreted correctly?

$\endgroup$
2
$\begingroup$

No. The interpretation would be that a one-unit increase in ladder rank is associated with an expected (1 -.957)% = 4.3% decrease in the odds of being contracted. To obtain the probability at rank level $X$, you'd need to calculate, $\hat{p}$=$exp(X\hat{\beta})\over{1+exp(X\hat{\beta})}$.

$\endgroup$
  • $\begingroup$ Thank you very much for your help with this interpretation. $\endgroup$ – Courtney Feb 25 '16 at 6:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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