Chapter 5: Inference for Numerical Data

This set of labs discusses methods of inference for numerical data (two-sample tests and analysis of variance), in addition to related topics (statistical power and multiple testing). The last lab uses simulation and the principles of conditional probability to highlight some specific misconceptions about p-values; it can also be viewed as an informal introduction to the paradigm of Bayesian inference.
Lab Notes

1. Two-Sample Tests

Introduces hypothesis testing in the two-sample context, discussing the two-sample t-test for paired data and independent group data.

Handout Template Solutions

2. Statistical Power

Discusses the control of Type I and Type II error and explores the factors influencing the power of a statistical test via simulation.

Handout Template Solutions

3. Analysis of Variance (ANOVA)

Introduces the analysis of variance procedure for comparing the means of several groups.

Handout Template Solutions

4. Multiple Testing

Examines the multiple testing problem and concept of experiment-wise error in the context of the Golub leukemia data.

Handout Template Solutions

5. A Closer Look at the P-Value

Integrates the ideas of conditional probability and hypothesis testing to present a broader understanding of p-values, Type I error, and statistical power in a research context.

Handout Template Solutions

R (Base), Useful Links

The statistical software R is a widely used and stable software that is free. RStudio is a user-friendly interface for R.