Visualizing Confidence Intervals

Click here to go to the applet.

Click here to go to the applet

This applet is meant as an aid for students in an undergraduate introductory statistics course. Students often find it difficult to grasp the (rather unintuitive) interpretation of a confidence interval: in an infinite number of independent experiments, a 95% confidence interval would contain the true population parameter 95% of the time. The applet is an attempt to make this more apparent by drawing a large(ish) number of samples, creating confidence intervals around each sample mean, and showing what proportion of these intervals contained the population mean.

Despite a few minor bugs that are still present, this applet placed first in the Penn State statistics department's Shiny Contest in 2015. It was built using R Shiny. The code is available on GitHub.