In marketing, a nonlinear probabilistic model based on a normal distribution is used to analyze binary variables. Several variants of the probit model exist, adapted to different types of analyses. According to Yves Evrard and his colleagues, a comparison between Logit and Probit models reveals that the Probit approach has two drawbacks compared to Logit: it relies on complex theoretical foundations and makes a strong assumption of normality, which can be problematic in many econometric contexts. In cases where this assumption is not met, it is preferable to opt for a Logit model.
Return to the glossary indexProbit (model)
Return to the glossary index