Constructing the Confidence Interval with MythBusters Yawn

Let’s first construct the bootstrap distribution for $\hat{p}_{seed}$ − $\hat{p}_{control}$ and then use this to construct 95% confidence intervals for $\hat{p}_{seed}$ − $\hat{p}_{control}$. We’ll do this using the infer workflow again. However, because the difference in proportions is a new scenario for inference, we’ll need to use some new arguments in the infer functions along the way.

Specify variables

Let’s take our mythbusters_yawn data frame and specify() which variables are of interest using the y ~ x formula interface where:

• Our response variable is yawn: whether or not a participant yawned. It has levels "yes" and "no".

• The explanatory variable is group: whether or not a participant was exposed to yawning. It has levels "seed" (exposed to yawning) and "control" (not exposed to yawning).