Chapter 15 Exercises: Political Misinformation and Election Integrity

These exercises range from conceptual analysis to applied research and coding tasks. Complete all exercises assigned by your instructor, noting word count and citation requirements where specified.

Part A: Conceptual and Analytical Exercises

Exercise 15.1 — Definitional Analysis (300 words)

Wardle and Derakhshan distinguish misinformation, disinformation, and malinformation. Select one documented case of political false information from the past five years and classify it using this framework. Explain your reasoning, and identify any ambiguities or limitations of the framework that your case reveals.

Exercise 15.2 — Partisan Asymmetry Evaluation (500 words)

The claim that political misinformation is asymmetrically concentrated among right-wing actors is both empirically studied and politically contested. Write an essay that: - Summarizes the empirical evidence for asymmetry - Identifies methodological critiques of this research - Explains what the policy implications would be if asymmetry is real, versus if it is a methodological artifact - Arrives at your own evidence-based assessment

Exercise 15.3 — Typology Application

For each of the following claims (real or realistic), classify it using the election misinformation typology from Section 15.2. Explain your classification and identify what additional information would affect your assessment:

a. "Election Day has been extended due to high turnout — all registered voters have until 9 PM Wednesday to cast their ballots." b. "Senator X voted to defund the police three times." c. "Dominion Voting Systems is owned by a company founded in Venezuela to rig elections for Hugo Chávez." d. "Your vote won't count if you have any unpaid traffic tickets." e. "The midnight 'ballot dump' in Philadelphia is statistically impossible and proves fraud." f. "Candidate Y admitted privately that she doesn't support the Second Amendment." (Based on a selectively edited video.)

Exercise 15.4 — IRA Operations Analysis (750 words)

Using information from Section 15.3 and additional research into the Senate Intelligence Committee reports, write an analysis that addresses: - The IRA's targeting strategy: why Black American communities? What does this targeting tell us about the IRA's actual goals? - The distinction between "reach" and "impact" in assessing IRA effectiveness - What the IRA operations reveal about the vulnerabilities in US social media platforms that existed in 2016 - What reforms were made after 2016 and whether they would have prevented the operations

Exercise 15.5 — Judicial Record Examination

The chapter states that over 60 lawsuits challenging the 2020 election were dismissed or rejected. Research and answer: a. Find and summarize the reasoning of at least two specific judicial decisions rejecting 2020 election fraud claims, noting the appointing president of each judge. b. What standards do courts apply when evaluating election fraud claims? How does the legal evidentiary standard differ from the standard implied by "thousands of affidavits"? c. Giuliani and Powell made specific claims in press conferences that they did not include in court filings. What ethical and legal rules govern attorney conduct in court versus public statements?

Exercise 15.6 — Comparative Case Study (1,000 words)

Compare two election interference cases from Section 15.7 (choose any two: Brazil 2018, Brexit, France 2017, Philippines). Your comparison should address: - What types of misinformation were deployed in each case - What media platforms were central to each operation - What evidence we have for the operations' effects - What institutional factors may explain differences in outcomes - What lessons each case holds for the other country's future elections

Exercise 15.7 — The Liar's Dividend

Researchers have argued that the most important current effect of deepfake technology is the "liar's dividend" — the ability of public figures to deny authentic recordings. Write a 400-word analysis: - Explain the mechanism: how does the existence of deepfakes allow denial of authentic content? - Provide at least one documented example of a public figure invoking potential deepfake status to deny authentic evidence. - Assess: is this a bigger threat than actual deepfakes? Under what conditions?

Part B: Research and Source Evaluation Exercises

Exercise 15.8 — Source Identification and Evaluation

For each source listed below, identify: (a) the organization's mission and funding, (b) whether it meets IFCN fact-checker accreditation standards, (c) its history of accuracy based on available meta-assessments: - PolitiFact - FactCheck.org - NewsGuard - Snopes - The Daily Wire's fact-checking content - USA Facts

Exercise 15.9 — Primary Source Research

Locate and read at least one of the following primary source documents, then write a 500-word summary of its key findings and methodological approach: a. Senate Intelligence Committee Report on Russian Interference, Volume 2 (Social Media Influence) b. Mueller Report, Volume 1, Section on Internet Research Agency c. New Knowledge / DiResta et al., "The Tactics & Tropes of the Internet Research Agency" (2019) d. Benkler, Faris & Roberts, "Network Propaganda" (any chapter)

Exercise 15.10 — Platform Policy Analysis

Visit the current published policies on election misinformation for Facebook/Meta, YouTube/Google, and Twitter/X. For each platform: a. Describe what types of election content are prohibited, labeled, or reduced in distribution. b. Note any significant changes since 2020. c. Identify gaps or ambiguities in the policies. d. Evaluate: do these policies address the types of misinformation documented in this chapter?

Exercise 15.11 — Voter Suppression Documentation

Research and document one case of voter suppression disinformation not mentioned in this chapter. Your documentation should include: - The false claim(s) and the target community - Available evidence about the source of the operation - Platform(s) or channels through which it spread - Any legal or regulatory response - Estimated reach and likely impact

Exercise 15.12 — International Comparison Research

The chapter briefly describes election interference in France 2017, Brazil 2018/2022, Brexit, and the Philippines. Research one additional country case (suggestions: Hungary, India, Germany, Canada, Mexico) and write a 600-word analysis that addresses the same dimensions covered for the chapter's cases.

Part C: Applied Analysis Exercises

Exercise 15.13 — Misinformation Timeline

Create a detailed timeline of the "Big Lie" narrative from August 2020 (when Trump began making pre-emptive fraud claims) through January 6, 2021. Your timeline should include: - Key Trump statements and tweets with dates - Key allied media claims and their platforming - Platform responses (removals, labels) - Legal filings and outcomes - Key news events Format as a visual timeline or structured table.

Exercise 15.14 — Inoculation Design Exercise

Design a "prebunking" intervention targeting one type of election misinformation. Based on inoculation theory principles (see Chapter 10), your design should: - Identify the manipulation technique being prebunked (not the specific false claim) - Specify the target audience - Describe the format (video, social media post, interactive) - Specify what psychological principles it activates - Explain how you would evaluate its effectiveness

Exercise 15.15 — Platform Recommendation Memo

You have been hired as a misinformation policy consultant to advise a social media platform on its election integrity policies for the next major election cycle. Write a professional memo (750 words) that: - Recommends specific policies for at least three categories of election misinformation - Addresses the tension between free expression and misinformation reduction - Discusses enforcement challenges and proposes solutions - Addresses the asymmetry debate: should policies be applied symmetrically even if misinformation is not?

Exercise 15.16 — Network Analysis Interpretation

The chapter describes coordinated inauthentic behavior networks. Given the following (hypothetical) data about a network of 500 social media accounts: - 80% of accounts were created within a 48-hour window - Accounts use profile photos that reverse-image search as AI-generated - Accounts post nearly identical content within minutes of each other - Accounts have very high engagement rates with each other but minimal engagement with organic accounts - No accounts post original content — only reshared content from 12 specific sources

Write an analysis (400 words) assessing: (a) the likelihood this represents coordinated inauthentic behavior; (b) what additional evidence would strengthen your assessment; (c) what the network's apparent purpose seems to be.

Exercise 15.17 — Deepfake Policy Analysis

The chapter notes significant differences between the US and European approaches to synthetic media regulation. Research current law and proposed legislation in: - The United States (federal and at least two state proposals) - The European Union (AI Act provisions) Write a 600-word comparative analysis of the regulatory approaches, assessing their likely effectiveness and trade-offs.

Exercise 15.18 — The CISA Case

The chapter describes CISA Director Chris Krebs being fired after declaring the 2020 election secure. Research this case further and write a 500-word analysis: - What specifically did Krebs and colleagues say? - Why did the Trump administration find it objectionable? - What does this case reveal about the relationship between technical security expertise and political authority? - What institutional changes, if any, might protect technical election security officials from political retaliation?

Part D: Writing and Communication Exercises

Exercise 15.19 — Op-Ed Writing

Write a 700-word op-ed for a general audience newspaper addressing one of the following topics: a. Why courts' rejection of the stolen election claims should (or should not) resolve the debate b. The ethical responsibility of television networks in deciding whether to broadcast unsubstantiated fraud claims c. What foreign election interference tells us about domestic media vulnerabilities

Your op-ed should be factually accurate, acknowledge complexity, and make a clear argument supported by evidence.

Exercise 15.20 — Fact-Check Exercise

Select a political claim currently circulating on social media. Perform a systematic fact-check following IFCN standards: - State the claim precisely as made - Rate its accuracy (True / Mostly True / Half True / Mostly False / False / Pants on Fire or similar) - Document your sources and reasoning - Address the most persuasive version of the claim, not a straw man - Write a 400-word fact-check suitable for publication

Exercise 15.21 — Explainer Article

Write a 600-word explainer article for a general audience explaining what the Internet Research Agency was, what it did, and what its actual documented impact was. Your article should: - Avoid both over-dramatization and dismissiveness - Accurately represent the state of research on impact - Cite sources appropriately for a general audience (links rather than academic citations) - Address the question: "Should I be worried about this?"

Exercise 15.22 — Counter-Narrative Development

You work for a nonpartisan election administration organization. Develop a counter-narrative strategy for addressing false claims about mail-in voting security in your state. Your strategy should: - Identify the specific false claims you are targeting - Specify your target audience (concerned but persuadable voters) - Develop accurate counter-messages that are emotionally resonant - Specify platforms and formats - Explain how you will avoid creating "implied truth effects" on unlabeled false content

Part E: Coding and Data Analysis Exercises

Exercise 15.23 — Classifier Evaluation

Using the code in example-01-election-misinfo-classifier.py, modify the classifier to include at least 20 additional training examples per category. Evaluate the updated classifier's performance using cross-validation and report: - Precision, recall, and F1 score for each category - Confusion matrix analysis: which categories are most often confused? - Two examples of claims that the classifier misclassifies, with explanation

Exercise 15.24 — Coordination Detection Extension

Using example-02-coordinated-behavior-detection.py as a starting point, add at least one additional feature for detecting coordinated behavior beyond time-series analysis. Options include: - Content similarity clustering - Account creation time clustering - Network graph analysis of reply/retweet patterns Document your added feature and evaluate whether it improves detection rate on the synthetic dataset.

Exercise 15.25 — Partisan Source Analysis

Using example-03-partisan-source-analysis.py, extend the analysis to compute: a. The diversity index of a sample of 10 simulated user feeds (using Shannon entropy or a similar metric) b. The correlation between feed partisanship and low-credibility source exposure c. A visualization showing how feed diversity changes as a function of how many sources a user follows Write a 300-word interpretation of your results.

Exercise 15.26 — Novel Dataset Exploration

Download the publicly available dataset of IRA Twitter accounts released by Twitter (available via the Stanford Internet Observatory or academic repositories). Select a random sample of 1,000 tweets and: a. Compute basic descriptive statistics (posting times, languages, hashtags) b. Identify the most common hashtags and classify them by political topic c. Analyze whether posting times suggest automated behavior (unusual time-zone patterns, regular intervals) d. Write a 500-word summary of your findings

Exercise 15.27 — Visualization Project

Create a data visualization that communicates a key finding from this chapter to a general audience. Your visualization may use any tool (Python matplotlib/seaborn, Tableau, Flourish, R, etc.) and should: - Display data accurately with appropriate uncertainty communication - Be interpretable without a statistics background - Include a clear title, labels, and source citation - Be accompanied by a 200-word caption explaining what the visualization shows