Workshop Statistics: Discovery with Data, Fourth Edition

TOPIC 5

Designing Experiments

Do strength shoes (modified athletic shoes with a 4-cm platform attached to the front half of the sole) really help a person jump farther? How would you design a study to investigate this claim? What factors in a memory experiment affect how many letters a person can memorize correctly? Can a nicotine lozenge help smokers who want to quit? One common aspect of these questions is an interest in finding out whether or not one variable has an effect on another variable. The question is not simply whether those who take a nicotine lozenge tend to quit smoking, but whether you can say that their quitting is because of the lozenge and not some other factor. This topic teaches you how to design studies so they produce data that can answer such questions.

Overview

In the previous topic, you studied the idea of random sampling as a fundamental principle by which to gather information about a population. Sometimes, however, your goal is not to describe a population but to investigate whether one variable has an effect on another variable. This topic introduces you to the design of controlled experiments for this purpose, contrasting them with the observational studies that you explored in earlier topics. You will discover principles for designing controlled experiments and learn how to avoid the problem of confounding variables. You will also explore properties of random assignment and see how this process differs from random sampling, which you studied in ...