R Script: Chi-Square Tests for Parametric and Non-parametric Models

# Chi square tests for parametric and non-parametric models

# Conduct chi-square tests on nested parametric models
for (i in 1:length(taxa.names)) {

     print(taxa.names[i])
     resp <- dfmerge[,taxa.names[i]] > 0

     # Fit a model that is only a constant
     modcmp <- glm(resp ~ 1, family = binomial, data = dfmerge)

     # Compare original model with constant model using
     # a chi-square statistic
     modout <- anova(modlist.glm[[i]], modcmp, test = "Chi")
     print(modout)

     # Select p < 0.05 as statistically significant
     if (modout[2,"P(>|Chi|)"] < 0.05) {
          print("Model significant compared to constant")
     }

     # Fit a model with only a linear explanatory variable
     modcmp <- glm(resp ~ temp, family = binomial, data = dfmerge)

     # Compare original model with constant model using a
     # chi-square statistic
     modout <- anova(modlist.glm[[i]], modcmp, test = "Chi")
     print(modout)
     if (modout[2,"P(>|Chi|)"] < 0.05) {
          print("Model with b2 significant (p < 0.05) improvement over linear model")
     }

}

# Conduct chi-square tests on nested non-parametric models
library(gam)
for (i in 1:length(taxa.names)) {

     print(taxa.names[i])
     resp <- dfmerge[,taxa.names[i]] > 0
     modcmp <- gam(resp ~ 1, family = binomial, data = dfmerge)
     modout <- anova(modlist.gam[[i]], modcmp, test = "Chi")
     print(modout)
     if (modout[2,"P(>|Chi|)"] < 0.05) {
          print("Model significant compared to constant")
     }

     # Fit a model with only a linear explanatory variable
     modcmp <- gam(resp ~ temp, family = binomial, data = dfmerge)
     modout <- anova(modlist.glm[[i]], modcmp, test = "Chi")
     print(modout)
     if (modout[2,"P(>|Chi|)"] < 0.05) {
          print("Nonparametric model with two degrees of freedom significant over linear model.")
     }

}