Quiz: A Brief History of Polling and Political Measurement

Questions

1. The earliest recorded straw poll in American politics was conducted in which year? - (a) 1789 - (b) 1824 - (c) 1860 - (d) 1916

2. The Literary Digest poll of 1936 surveyed approximately how many people? - (a) 50,000 - (b) 500,000 - (c) 2.4 million - (d) 10 million

3. What was the primary source of error in the 1936 Literary Digest poll? - (a) The questions were poorly worded - (b) The sampling frame overrepresented affluent Americans - (c) The poll was conducted too early - (d) The poll used probability sampling incorrectly

4. George Gallup's key insight in 1936 was that: - (a) Larger samples always produce better results - (b) Mail surveys are more accurate than phone surveys - (c) Representativeness matters more than sample size - (d) Polling should only be done in the final week of a campaign

5. The method used by Gallup and other early scientific pollsters, in which interviewers filled demographic quotas using their own judgment, is called: - (a) Probability sampling - (b) Stratified random sampling - (c) Quota sampling - (d) Snowball sampling

6. In the 1948 election, all major pollsters predicted that __ would win, but ____ actually won. - (a) Roosevelt; Landon - (b) Dewey; Truman - (c) Truman; Dewey - (d) Eisenhower; Stevenson

7. Which of the following was NOT identified as a cause of the 1948 polling failure? - (a) Interviewers selected unrepresentative respondents within quotas - (b) Pollsters stopped surveying too early - (c) The sample size was too small - (d) Undecided voters were not allocated to candidates

8. Random digit dialing (RDD) is associated with which type of sampling? - (a) Quota sampling - (b) Convenience sampling - (c) Probability sampling - (d) Snowball sampling

9. According to the chapter, telephone polling response rates fell from approximately 36% in 1997 to less than what percentage by 2024? - (a) 20% - (b) 12% - (c) 8% - (d) 4%

10. Which of the following is cited as a major cause of the decline in telephone polling response rates? - (a) People no longer have opinions about politics - (b) Caller ID, robocalls, and declining institutional trust - (c) The elimination of landline telephones by law - (d) Pollsters began asking too many questions

11. The primary methodological concern with non-probability online panels is: - (a) They are too expensive to operate - (b) They cannot include visual elements in surveys - (c) Panelists are self-selected and may not be representative - (d) They can only be used for market research, not political polling

12. Meridian Research Group's hybrid polling approach includes all of the following EXCEPT: - (a) Live telephone interviews with landline and cell phone numbers - (b) Text-to-web surveys sent to randomly selected registered voters - (c) In-person door-to-door interviews - (d) Online panel respondents from a probability-based panel

13. After the 2016 polling errors, what methodological adjustment did many pollsters adopt? - (a) Switching entirely to online panels - (b) Weighting by education level - (c) Eliminating cell phone samples - (d) Reducing sample sizes

14. Vivian Park's "Worry List" is best described as: - (a) A list of competitors who might take Meridian's clients - (b) A running catalog of potential sources of error in the firm's methodology - (c) A record of polls that Meridian got wrong - (d) A list of survey questions that are too sensitive to ask

15. The chapter's central theme, "Measurement Shapes Reality," means: - (a) Polls always change the outcome of elections - (b) The methods used to measure opinion determine whose opinions are counted and how they are represented - (c) Better measurement technology guarantees better predictions - (d) Political reality can be understood only through quantitative measurement

Answer Key

  1. (b) 1824. The Harrisburg Pennsylvanian reported candidate preferences at a public gathering in Wilmington, Delaware.

  2. (c) 2.4 million. The Digest mailed out approximately 10 million postcards, and 2.4 million were returned.

  3. (b) The sampling frame overrepresented affluent Americans. The Digest drew names from telephone directories and automobile registrations, which in 1936 were markers of prosperity that correlated with Republican preference.

  4. (c) Representativeness matters more than sample size. Gallup correctly predicted Roosevelt's win with a sample of only a few thousand, compared to the Digest's 2.4 million.

  5. (c) Quota sampling. In this method, interviewers fill specified demographic quotas but have discretion in selecting individual respondents within each quota.

  6. (b) Dewey; Truman. All major pollsters predicted Dewey would win. The Chicago Tribune famously printed the headline "DEWEY DEFEATS TRUMAN."

  7. (c) The sample size was too small. The other three factors---interviewer selection bias within quotas, early cessation of polling, and failure to allocate undecided voters---were all identified as causes. Sample size was adequate.

  8. (c) Probability sampling. RDD generates random telephone numbers, giving every household with a telephone a known probability of selection.

  9. (d) 4%. The chapter states response rates fell to "less than 4 percent" by 2024.

  10. (b) Caller ID, robocalls, and declining institutional trust. The chapter identifies multiple reinforcing causes, including technology that enables call screening, the explosion of robocalls and scams, and declining trust in institutions.

  11. (c) Panelists are self-selected and may not be representative. Non-probability panels recruit through convenience methods, meaning participants volunteer rather than being randomly selected.

  12. (c) In-person door-to-door interviews. Meridian's hybrid approach combines live telephone interviews, text-to-web surveys, and probability-based online panel respondents. In-person interviews are not part of their standard methodology.

  13. (b) Weighting by education level. The AAPOR post-mortem identified education-based nonresponse bias as a key factor in 2016 errors, and many pollsters subsequently began weighting by education.

  14. (b) A running catalog of potential sources of error in the firm's methodology. Vivian updates it after every election cycle to track new and persistent sources of potential bias.

  15. (b) The methods used to measure opinion determine whose opinions are counted and how they are represented. The chapter traces how each era's technology---straw polls, telephone surveys, online panels---shaped which voices were included in "public opinion."