When logit analysis first became popular, a major complaint by those who resisted its advance was that the coefficients had no intuitive meaning. Admittedly, they’re not as easy to interpret as coefficients in the linear probability model. For the linear probability model, a coefficient of .25 tells you that the predicted probability of the event increases by .25 for every 1-unit increase in the explanatory variable. By contrast, a logit coefficient of .25 tells you that the log-odds increases by .25 for every 1-unit increase in the explanatory variable. But who knows what a .25 increase in the log-odds means?

The basic problem is that the logit model assumes a nonlinear relationship between the probability and ...

Start Free Trial

No credit card required