Logistic regression models can be compared with respect to their predictive power by taking the difference of their respective -2LLs.

Question

Quizplus · Accepted Answer

The -2LL (log-likelihood) is commonly used as a measure of model fit in logistic regression. Smaller -2LL values indicate better fit and, therefore, stronger predictive power. Comparing the difference in -2LLs between two models can indicate which one is a better fit and has stronger predictive power.

Logistic Regression Models Can Be Compared with Respect to Their