pchisqC: Comparing Logistic Regression Models with Chi-Squared Test

In poliscidata: Datasets and Functions Featured in Pollock and Edwards, an R Companion to Essentials of Political Analysis, Second Edition

Description

Comparing Logistic Regression Models with Chi-Squared Test

Usage

pchisqC(reduced, full, digits = 3)

Arguments

reduced	The reduced logisitic regression model as an object. This is model with fewer independent variables.
full	The full logisitic regression model as an object. This is model with more independent variables.
digits	The number of digits to display after decimal point, default is 3.

Value

The chi-squared statistic, df, and p-value as a vector of numbers to test null hypothesis that full model no better than reduced model.

Examples

   library(poliscidata)
   
   obama_vote_model_reduced <- svyglm(obama_vote ~ ft_dem + white, design=nesD, 
                                      family="quasibinomial")
   obama_vote_model_full <- svyglm(obama_vote ~ ft_dem + white + (owngun_owngun=="1. Yes"), 
                                   design=nesD, family="quasibinomial")
   pchisqC(obama_vote_model_reduced, obama_vote_model_full)

Example output

Warning: These models have different numbers of observations.
[1] "<.001"

poliscidata documentation built on July 8, 2020, 6:23 p.m.