The log-likelihood of a model is a measure of model fit that can be used to compare different kinds of models (or variations on the same model). Higher values (that is, less negative values) correspond to better fit. The log-likelihood is available for all models.
To view a model’s log-likelihood:
For categorical outcome variables, the log-likelihood is broken down by outcome. By looking for changes in each individual category’s log-likelihood, you can see whether changing the model achieves a better fit for all of the outcome categories, or only some of them.