Logit vs Probit

In a previous article, we illustrated the usage of logit model to predict the potential PTPTN defaulters (Read more here).  One could use the probit model to run the analysis as well.  In the logit model the log odds of the outcome is modelled as a linear combination of the predictor variables.  Meanwhile, in the probit model, the inverse standard normal distribution of the probability is modelled as a linear combination of the predictors.

The following chart shows the comparison between logit and probit results.  Both models gave similar results, with logit model giving slightly “fatter” tails.  The capability of their predicting power is the same for both models.

The choice between logit and probit model is purely personal.  But it is always good to have an alternative model to cross check the results.  They are used in various fields, including machine learning.  Logit is more popular in medical fields and social sciences while probit is often used in advanced econometric, finance and political sciences.