Please enable JavaScript to view this site.

The effect of changes in the β0 and β1 coefficients calculated by simple logistic regression can't easily be translated to changes in probability. Why not?

The reason β0 and β1 can’t easily be applied directly to an increase or decrease in probability is because the change in probability depends on the specific value of X. When the probability is near 0.5, a one unit increase in X will have a much larger effect than, for example, when the probability is near 0 or 1. If you’re interested in this, you can read more about the relationships of log odds, odds, and probability.


© 1995-2019 GraphPad Software, LLC. All rights reserved.