Although Lott found that the carry law was associated with reductions
in crime in his multivariate model, whether these reductions are real
depends on whether the model is correct.
The best test of such a model is its predictive power, unfortunately,
as we shall see below, it fails such a test.
Other evidence that
the model is incorrect is that it throws up many spurious correlations.
Other researchers have proposed alternative models that do not have the
crime reductions found in Lott's model.