Example 1

Example 1: In 1936, Literary Digest magazine attempted to predict the winner of the contest for President of the United States between Alf Landon (R) and Franklin D. Roosevelt (D). They attempted to poll ten million people by sending out ballots to magazine subscribers, automobile owners, and telephone users. Over two million ballots were returned which is a large number of respondents. The polls predicted that Alf Landon would capture 57% of the vote, which would have made him the winner by a landslide.

This poll result carried a lot of weight since Literary Digest magazine had used similar methods to accurately predict the winner in the 1916, 1920, 1924, 1928, and 1932 elections, so it was a surprise when Alf Landon only received 37% of the 44,431,180 votes cast in the 1936 election. In this example what was the target population and what was the sample? (Answer: Population = the collection of voting preferences of the entire eligible adults in the USA in 1936. Sample = the over two million poll respondents. The poll had a major selection bias—i.e. it was not a random selection of the population, not everyone had a car or a phone or a magazine membership in 1936, so not everyone member of the population had an equal chance of being selected.)