This is a question we receive a lot. The question is asked in the spirit of wanting to know whether a predictive model can be trusted. It makes sense when approaching a predictive model as being a causal explanation of some behavior.
In “What’s in the Model?” Capture Higher Ed’s Pete Barwis breaks down why we ask this question, and whether it is the right question to be asking. He argues that, rather than focusing on what goes into the model, the more illuminating and useful question is, “What amount of confidence should I put in the model outputs?”