Chapter 29 Exercises: Voter Targeting and Microtargeting

Chapter 29 Exercises: Voter Targeting and Microtargeting

Conceptual Exercises

Exercise 29.1 — Universe Building Logic A campaign is preparing for a competitive congressional election in a district with 280,000 registered voters. The current best estimate is that the candidate has approximately 130,000 base supporters (who will vote for her if they vote) and that there are approximately 40,000 genuinely persuadable voters. The candidate needs 145,000 votes to win given expected turnout.

(a) How many net votes does the candidate need to gain through persuasion, mobilization, or both? (b) If personal canvassing has a mobilization effect of adding about 8 percentage points to turnout probability for contacted voters, and the campaign has resources to canvass 45,000 GOTV targets, how many additional votes can it expect to generate from the GOTV program? (c) If persuasion contact (also canvassing) moves about 4% of contacted voters, and the campaign can canvass 15,000 persuasion targets, how many net votes does that generate? (d) How would you balance the GOTV and persuasion canvassing programs given the math above?

Exercise 29.2 — Score Design A campaign wants to build a persuadability score. Describe: (a) what you are trying to measure, (b) why it cannot be observed directly, (c) three data sources you would use to build the model, (d) one validity check you could run to assess whether the model is measuring what you intend it to measure.

Exercise 29.3 — The Third-Party Data Problem A campaign's consumer data vendor provides an "issue affinity" score for immigration, ranking each voter 1–10 on their estimated concern about immigration policy. The vendor says the score is built from "proprietary consumer data signals." Describe: (a) three specific data signals you would expect the vendor to be using, (b) two ways this score could be systematically inaccurate for specific population groups, and (c) what validation test you would run before incorporating the score into your targeting model.

Exercise 29.4 — Targeting Philosophy Compare Nadia Osei's suburban persuasion focus with Jake Rourke's exurban mobilization focus. For each strategy: (a) describe the underlying theory of the electorate that justifies it, (b) identify one condition under which that theory would be correct, (c) identify one condition under which it would fail, and (d) propose a data signal that would, in real time, tell you whether the theory was holding up.

Exercise 29.5 — The Efficiency-Equity Tradeoff Campaign targeting is designed to be efficient — to direct resources where they will have the greatest electoral impact. But efficiency in this sense has distributional consequences: some voters receive extensive contact, others receive none. Write a 400–500 word essay arguing either that (a) campaigns should modify their targeting practices to increase democratic inclusion even at some cost to efficiency, or (b) campaigns should maximize electoral efficiency and leave the problem of democratic inclusion to other institutions.

Applied Exercises

Exercise 29.6 — Segment Analysis You have a voter file with 500,000 registered voters. Each voter has a support score (0–100), a turnout score (0–100), and a persuadability score (0–100). Using the following scoring rules, calculate the size of each universe:

GOTV universe: support score ≥ 65, turnout score 30–69
Persuasion universe: support score 40–64, turnout score ≥ 50, persuadability score ≥ 60
Fundraising universe: support score ≥ 70, consumer wealth percentile ≥ 70
No-contact: support score ≤ 30

Assume the following distribution: - 30% of voters have support ≥ 65 - Of those, 40% have turnout scores 30–69 - 25% of voters have support 40–64, of whom 60% have turnout ≥ 50, of whom 45% have persuadability ≥ 60 - 20% of voters have support ≤ 30

Calculate the approximate size of each universe and the percentage of the voter file in each.

Exercise 29.7 — Message Track Design You are the targeting director for a Senate candidate. Your issue affinity model identifies five issue clusters: healthcare, education, environment, economy, and public safety. Your campaign wants to design three message tracks for direct mail to the persuasion universe.

Design the three tracks: for each, specify (a) the primary and secondary issue emphasis, (b) the target voter profile (what combination of affinity scores defines this track's audience), (c) the candidate attribute or record you would emphasize, and (d) one claim you would need to fact-check before the mail goes out.

Exercise 29.8 — Digital Targeting Transparency Audit Access the Facebook Ad Library (available at facebook.com/ads/library) and search for political ads from any current or recent election campaign in your state or region. For three ads you find: (a) identify the apparent target audience based on the ad content, (b) compare that apparent target to any disclosed targeting information the platform provides, (c) assess whether the claims in the ad are verifiable through public sources, and (d) describe what information about the ad's reach and targeting you cannot determine from the library.

Exercise 29.9 — Ethics Case Analysis A campaign's analytics team builds an issue affinity model and discovers that voters in two specific zip codes — which have majority immigrant-origin populations — score very high on immigration issue affinity. The political director proposes running Spanish-language social media ads in those zip codes emphasizing the candidate's support for a pathway to citizenship, targeted specifically at these high-immigration-affinity voters. The analytics director notes that the same model, in the hands of an opposing campaign, could be used to identify these same voters and deliver fear-based immigration messaging to them.

Write a 300–400 word analysis that addresses: (a) the ethical status of the first campaign's proposed targeted ads, (b) the concern raised by the analytics director, and (c) what, if any, ethical obligations campaigns have regarding how their targeting data could be misused.

Discussion Exercises

Exercise 29.10 — The Manipulation Question The chapter distinguishes between persuasion (offering arguments a rational agent can evaluate) and manipulation (bypassing rational evaluation by exploiting vulnerabilities). In a group discussion or written reflection: (a) identify a real or hypothetical political ad or message that you believe crosses from persuasion into manipulation, (b) explain why it crosses the line, and (c) propose a rule or norm that campaigns could follow to stay on the persuasion side of the line.

Exercise 29.11 — The Disclosure Debate Several proposals have been made to increase transparency in political microtargeting: ad libraries that show all political ads, mandatory disclosure of targeting criteria, prohibition on certain targeting methods, and user rights to see and opt out of political ad targeting. Evaluate each proposal for its likely effectiveness in addressing the transparency problem and its likely operational impact on campaigns.

Research Exercise

Exercise 29.12 — Effectiveness Evidence Review Find and read two academic studies on the effectiveness of voter targeting or microtargeting. For each study: (a) what type of targeting was studied (GOTV, persuasion, digital, issue-specific)? (b) what was the research design (randomized experiment, quasi-experiment, observational study)? (c) what did the study find, expressed in plain terms? (d) what are the main limitations of the study? Then write a paragraph synthesizing what the two studies together tell you about when targeting is most and least likely to be effective.