Quiz: The Age of Political Data

Questions

1. Which of the following best describes the relationship between data volume and analytical quality in political analytics? - (a) More data always leads to better analysis - (b) More data can lead to more confident but potentially wrong conclusions - (c) The amount of data is irrelevant; only the method matters - (d) Less data is generally preferable because it is easier to manage

2. The chapter distinguishes between three "worlds" of political data. Which world's primary goal is winning elections? - (a) The media/research world - (b) The civic world - (c) The campaign world - (d) The academic world

3. What is the key difference between analytics and punditry, as described in this chapter? - (a) Analytics uses television; punditry uses print media - (b) Analytics quantifies uncertainty; punditry asserts conclusions with confidence - (c) Analytics is always correct; punditry is always wrong - (d) Analytics focuses on local races; punditry focuses on national ones

4. The 1936 Literary Digest poll is mentioned as an example of: - (a) A poll that correctly predicted a landslide victory - (b) How polling response rates have improved over time - (c) A massive survey that produced a spectacularly wrong prediction - (d) The first scientifically conducted public opinion poll

5. What does the theme "Measurement Shapes Reality" mean in the context of political analytics? - (a) Political polls always change the outcome of elections - (b) The choices made about what and how to measure have real political consequences - (c) Data scientists have more political power than elected officials - (d) Accurate measurement guarantees accurate predictions

6. In the Garza-Whitfield Senate race, what is the approximate demographic breakdown of the state? - (a) 55% white non-Hispanic, 25% Hispanic/Latino, 12% Black, 8% other - (b) 38% white non-Hispanic, 32% Hispanic/Latino, 18% Black, 12% other - (c) 45% Hispanic/Latino, 30% white non-Hispanic, 15% Black, 10% other - (d) 40% white non-Hispanic, 20% Hispanic/Latino, 20% Black, 20% other

7. Which character is described as the analytics director for the Garza campaign? - (a) Carlos Mendez - (b) Jake Rourke - (c) Nadia Osei - (d) Sam Harding

8. Nadia Osei's quote about models being "bets on who shows up" primarily illustrates which analytical challenge? - (a) Campaign fundraising - (b) The difficulty of measuring a changing electorate - (c) Social media analysis - (d) Opposition research techniques

9. Response rates to telephone surveys have declined from approximately 35% in the late 1990s to less than what percentage by the mid-2020s? - (a) 25% - (b) 15% - (c) 10% - (d) 5%

10. OpenDemocracy Analytics (ODA) best represents which "world" of political data? - (a) The campaign world - (b) The media/research world - (c) The civic world - (d) The commercial world

11. The chapter describes a tension between prediction and explanation. Which statement best captures this tension? - (a) Good predictions always lead to good explanations - (b) Explanation is more important than prediction in every context - (c) A model that predicts well may not explain well, and vice versa - (d) Prediction and explanation are different words for the same thing

12. "Calibrated confidence" refers to: - (a) Being as confident as the evidence warrants, and no more - (b) Using calibration tools to adjust polling data - (c) Expressing confidence only when your model has been tested - (d) Reporting results without any uncertainty measures

13. Which of the following is NOT listed as part of the political analyst's core toolkit? - (a) Data acquisition and management - (b) Statistical modeling - (c) Campaign fundraising - (d) Visualization and communication

14. Jake Rourke's approach to campaigns is best described as: - (a) Purely data-driven with no room for intuition - (b) Experience-driven but not anti-data - (c) Opposed to all forms of quantitative analysis - (d) Focused exclusively on digital campaigning

15. The chapter's discussion of the Cambridge Analytica scandal serves as an example of: - (a) How open data benefits democracy - (b) The potential for political data to be misused - (c) Successful application of microtargeting - (d) Academic research in action

Answer Key

  1. (b) More data can lead to more confident but potentially wrong conclusions. The chapter uses the 1936 Literary Digest poll (2.4 million respondents, wrong prediction) and 2016 forecasting models as examples.

  2. (c) The campaign world. The chapter explicitly states that for campaigns, "data is a weapon" and "the goal is simple: win."

  3. (b) Analytics quantifies uncertainty; punditry asserts conclusions with confidence. The chapter contrasts a pundit saying "Garza is going to win" with an analyst giving a probability estimate.

  4. (c) A massive survey that produced a spectacularly wrong prediction. The Literary Digest surveyed 2.4 million people but predicted Landon would beat Roosevelt; Roosevelt won 46 of 48 states.

  5. (b) The choices made about what and how to measure have real political consequences. The chapter gives examples including definitions of "likely voter," district lines, and Census categories.

  6. (b) 38% white non-Hispanic, 32% Hispanic/Latino, 18% Black, 12% other. This is explicitly stated in the state description for the Garza-Whitfield race.

  7. (c) Nadia Osei. She is described as the 31-year-old Ghanaian-American analytics director who left a Stanford PhD program.

  8. (b) The difficulty of measuring a changing electorate. Nadia's quote is presented under "Challenge 1: Measuring a Changing Electorate."

  9. (d) 5%. The chapter states response rates have dropped "to less than 5 percent by the mid-2020s."

  10. (c) The civic world. ODA is described as a civic technology nonprofit dedicated to making political data accessible.

  11. (c) A model that predicts well may not explain well, and vice versa. The chapter discusses how prediction asks "What will happen?" while explanation asks "Why does it happen?" and these goals are "not always aligned."

  12. (a) Being as confident as the evidence warrants, and no more. This is presented as a "Best Practice" with the example that a 55% win probability means "slight favorite, not a sure thing."

  13. (c) Campaign fundraising. The toolkit includes data acquisition, descriptive analysis, statistical modeling, visualization/communication, and critical evaluation---not fundraising.

  14. (b) Experience-driven but not anti-data. The chapter states Rourke "trusts his gut" and values relationships but "is not anti-data" and has "seen enough campaigns to know that good data can sharpen strategy."

  15. (b) The potential for political data to be misused. Cambridge Analytica is discussed in the context of "the darker possibilities" of political data, alongside push polls and misleading microtargeted ads.