|
|
 |
|
1. In 1928 and 1932, the magazine Literary Digest polled its mailing list to predict the presidential elections, and was within 4 and 2 percentage points of the actual vote share. In 1936, it added names from auto registrations, voter registrations, and telephone directories to create a mailing of 10 million ballots, from which they received 2.3 million responses. It predicted that Alf Landon (R) would defeat Franklin Roosevelt (D) with a 54% vote share. FDR won that election with over 60% of the vote. What most likely went wrong? |
|
| The sample was biased towards individuals with higher incomes.
|
| The sample was dependent upon voluntary responses.
|
| The sample was just not large enough.
|
| Both A and B
|
|
|
2. In the various polls you see in the media, a margin of error is often reported along with the results. What is that margin of error accounting for? |
|
| The poll’s questions may have been badly worded and confusing.
|
| People who were supposed to be in the sample didn’t participate.
|
| The sample was biased in some way and not representative of the actual population
|
| By simple chance, the majority opinion of this sample is different from that of the overall population.
|
|
|
3. What approximate sample size is a 3% sampling error (commonly seen in opinion polls) generally associated with? |
|
| 100
|
| 1,000
|
| 10,000
|
| 100,000
|
|
|
4. Another method of predicting elections is to use a model of voter behavior. Professor Ray Fair of Yale University has developed a model which predicts the vote share in a presidential election, given party status, incumbency status, and three economic variables: inflation, output growth, and number of “good news” quarters (quarters in which the growth in GDP per capita was greater than 3.2%). The model has been quite accurate in the past, and this year’s prediction is a Bush victory by a sizable margin (58.7 share, with a standard error of 2.4). Does this mean a Bush victory is more or less a sure thing? |
|
| Yes, the model’s previous success indicates its accuracy.
|
| No, voters may weigh factors other than economic growth, inflation, and “good news” quarters more heavily now than in the past.
|
| No, the prediction itself has a large standard error.
|
| Yes, since the economy would need to perform unrealistically poorly to change the outcome.
|
|
|
|
| Created with QuizScript |