Current tests rely on comparisons of expected values coming out of a model fit vs. actual values which act as a snapshot.

Ideally, we would refactor the test to look for general properties being extracted rather than the exact value.