Workshop Statistics: Discovery with Data, Fourth Edition

TOPIC 4

Random Sampling

In the mid-1990s, public information campaigns urged parents to put babies on their backs to sleep, in an effort to reduce the occurrence of sudden infant death syndrome (SIDS). How can statistics help to determine whether or not these campaigns were successful? Can statistics help to resolve authorship of literary works whose origins are hotly debated? And what do these two research questions have in common? In both cases, it is not feasible to interview every member of the population; instead, you turn to sampling, which you will investigate more thoroughly in this topic.

Overview

Much of statistics concerns generalizing about observed findings from a sample to a larger population. For example, in the previous topic, you saw that a radio station wanted to estimate the proportion of all listeners who believe Elvis is alive by asking them to phone in with their opinions. However, you saw that generalizing information from the sample to the larger population of interest can be very misleading when the sampling method is biased, giving you a sample that is not representative of the population. This topic explores methods for selecting a sample so that trends and patterns observed in the sample can be reasonably generalized to the larger population of interest.

Preliminaries

1. Guess how many words are in the Gettysburg Address. (Activity 4-1)
2. Guess the proportion of American infants who sleep on their backs. (Activity 4-5)

Collect Data

1. Think about and then ...