calculate_lr_test: Function to calculate p-values for likelihood-ratio test

In mfp2: Multivariable Fractional Polynomial Models with Extensions

Function to calculate p-values for likelihood-ratio test

Description

Function to calculate p-values for likelihood-ratio test

Usage

calculate_lr_test(logl, dfs)

Arguments

logl	a numeric vector of length 2 with log-likelihoods. Typically ordered in increasing order (i.e. null model first, then full model) and used to test the ratio logl[1] / logl[2].
dfs	a numeric vector with degrees of freedom.

Details

Uses Wilk's theorem that -2log(LR) (LR = likelihood ratio) asymptotically approaches a Chi-square distribution under the null hypothesis that both likelihoods are equal.

Model likelihoods can then be compared by computing D = -2 log(likelihood reduced model / likelihood full model), and then use a Chi-square distribution with df_full - df_reduced degrees of freedom to derive a p-value.

This is basically the same way as stats::anova() implements the likelihood ratio test.

Value

A list with two entries for the likelihood ratio test for the ratio logl[1] / logl[2].

• statistic: test statistic.

• pvalue: p-value

mfp2 documentation built on Nov. 15, 2023, 1:06 a.m.