Comparison of regression models.

Note: R2 is the proportion of variance explained by the model; ΔR2 is the difference in R2 between the current model and first model; −2LL is the log likelihood of the regression model multiplied by −2; p was calculated to estimate change in −2LL by Chi-square distribution with df that equals the difference of dfs between models. AIC (Akaike's information criterion) and the BIC (Bayesian information criterion) are information-theory measures of goodness of model fit.