Consider all ways of adding a single additional predictor to model 2. Which one results in the smallest p-value of the Wald test for the significance of its regression coefficient? What is the p-value and how would you interpret the effect of this predictor has on the chance of developing diabetes?

library(faraway) library(dplyr) data(pima) summary(pima) pimafiltered = filter(pima, bmi > 0 & triceps >0)

model2 = lm(test~bmi+age+triceps+pregnant, data=pimafiltered)

pima data is used in model 2 and from faraway package