Tutorial illustration

Tutorial description

Categorial data is all around us. It’s in the latest opinion polling numbers, in the data that lead to new breakthroughs in genomics, and in the troves of data that internet companies collect to sell products to you. In this tutorial you’ll learn techniques for parsing the signal from the noise; tools for identifying when structure in this data represents interesting phenomena and when it is just random noise.

This tutorial also builds on statistical inference ideas we’ve learned about so far and applies them to situations where the response variable for our analysis is a numerical variable. In this tutorial we will learn how to conduct statistical inference in such scenarios using both simulation based methods as well as methods based on the central limit theorem.
We will encounter several new point estimates and a couple new distributions. In each case, the inference ideas remain the same: determine which point estimate or test statistic is useful, identify an appropriate distribution for the point estimate or test statistic, and apply the ideas of inference.

Learning objectives

Lessons

1 - Inference for a single proportion

In this lesson you will learn how to form a confidence interval on a single proportion using bootstrapping and an approximation-based method. You’ll explore how the sample size and the true proportion effect the width of the interval.

2 - Hypothesis tests to compare proportions

This lesson dives deeper into inference for a single parameter by performing hypothesis tests. Then, you’ll extend this technique to the difference between two proportions. Finally, this lesson wraps up with an exploration of what happens when you know the null hypothesis is true and the types of statistical errors that are baked in to this approach.

3 - Chi-squared test of independence

This part of the tutorial will teach you how to use both bootstrapping and approximation methods to test for the independence of two categorical variables. This commonly-used method is called the chi-squared Test of Independence.

4 - Chi-squared goodness of fit test

The tutorial wraps up with two case studies using election data. Here, you’ll learn how to a second Chi-squared test to check the goodness of fit between your data and a hypothetical distribution. You’ll study election results from Iran and Iowa and evaluate whether we can detect fraud using something called Benford’s Law.

5 - Bootstrapping for estimating a parameter

Construct confidence intervals for a median, mean, and any other statistic for the distribution of a numerical variable using bootstrapping.

6 - Introducing the t-distribution

Introduce the t-distribution for theoretical inference on means and constrast it to the normal distribution.

7 - Inference for difference in two means

Use simulation based and theoretical methods for inference in the difference of two means.

8 - Comparing many means

Compare many means using ANOVA.

Instructors

Andrew Bray

Andrew Bray

Reed College

Andrew Bray is an Assistant Professor of Statistics at Reed College and lover of all things statistics and R.


Mine Çetinkaya-Rundel

Mine Çetinkaya-Rundel

University of Edinburgh, Duke University, RStudio

Mine Çetinkaya-Rundel is Associate Professor of the Practice position at the Department of Statistical Science at Duke University and Data Scientist and Professional Educator at RStudio. Mine’s work focuses on innovation in statistics and data science pedagogy, with an emphasis on computing, reproducible research, student-centered learning, and open-source education as well as pedagogical approaches for enhancing retention of women and under-represented minorities in STEM. Mine works on integrating computation into the undergraduate statistics curriculum, using reproducible research methodologies and analysis of real and complex datasets. She also organizes ASA DataFest, an annual two-day competition in which teams of undergraduate students work to reveal insights into a rich and complex dataset. Mine has been working on the OpenIntro project since its founding and as part of this project she co-authored four open-source introductory statistics textbooks (including this one!). She is also the creator and maintainer of datasciencebox.org and she teaches the popular Statistics with R MOOC on Coursera.