Likelihood Ratio Test

This calculates a P-value for a term in a regression model based on how much the likelihood worsens when the term is removed from the model. Under the null hypothesis of no actual association of the term with the outcome being modeled, the change in minus 2 times the log likelihood follows the chi-squared distribution with 1 degree of freedom. This is used to calculate the p-value. A set of two or more terms can also be tested, using the chi-squared distribution with degrees of freedom equal to the number of terms removed from the model. Likelihood ratio tests must be performed with the models fitted to exactly the same set of observations, so care is needed when using software that drops observations with missing values.