Omar Wasow
LightNews LightNews
banner
owasow.bsky.social
Omar Wasow
@owasow.bsky.social
37K followers 3.8K following 3.1K posts
Asst Prof, UC Berkeley, Political Science. Study protests, stats & race: 1/ Agenda Seeding http://j.mp/agenda-seeding 2/ Race as a Bundle of Sticks http://j.mp/bundle-of
Posts Replies Media Videos
owasow.bsky.social
<40

fiftyplusone.news/polls/approv...
A line chart titled “Donald Trump Approval” shows presidential approval and disapproval trends over 2025, weighted by poll quality and recency, with shaded 95% confidence intervals. The pink line represents disapproval; the green line represents approval. In January, approval briefly rises from the mid-30s to around 47% before trending downward, ending near 40% in November. Disapproval starts near the mid-40s, climbs above 50% by March, and remains between 52–56% through the year, ending around 56%. A label for November 13, 2025 highlights: “Approve 39.9%, Disapprove 55.9% (Disapprove +16).” The chart shows widening divergence over time, with disapproval consistently higher than approval.
November 14, 2025 at 2:31 AM
owasow.bsky.social
New study “explores century-long persistence of buraku discrimination in Japan. We find that in 1912—40 years after liberation edict—land prices in buraku areas were 53% lower. Even in 2018, 150 years post-liberation, buraku areas had approximately 14% lower land prices.” voxdev.org/topic/instit...
Plot showing price penalty for buraku land. The orange dots represent the point estimates of the buraku price penalty in the border design with the linear specification of the buraku dummy. Estimates for 1961–2015 were obtained in the same way. The vertical bars represent the corresponding 95% confidence intervals. All numbers are rounded to the nearest integers.
November 9, 2025 at 2:27 PM
owasow.bsky.social
In our data—108 million homeowner-voters linked to campaign donations—we find Trump underperformed with wealthy in 2016 but, by 2020, mobilized a mass of non-wealthy donors. Bottom 90% nearly matched top 10% of dollars. He didn’t need establishment elites; he built something new. 2/ cup.org/4cfm0Az
Figure 3 in paper. Bar charts showing Percent Change in Fundraising between Romney in 2012 and Trump in 2016 and 2020, by Wealth Rank. In 2016, Trump lost 73% of Romney’s dollars from the top 1%, but lost only 55% of Romney’s dollars from the bottom 90%. In 2020, Trump recovered among the wealthy: the top 1% contributed similar totals to Trump and Romney, and the top 0.1% gave Trump 16% more than Romney. However, the most striking result from 2020 is Trump’s increased fundraising from the non-wealthy. The bottom 90% gave him 172% more than Romney. These findings reveal that Trump’s financing depended much less on the wealthy than did Romney’s. To be sure, Trump fully recovered wealthy dollars in 2020. But he more than recovered non-wealthy dollars. Trump’s non-wealthy dollars and shares are dramatically larger than Romney’s (and the largest in the figure). Open access link to paper: http://cup.org/4cfm0Az
November 8, 2025 at 4:39 PM
owasow.bsky.social
Earlier this year, we showed Trump built a remarkable “plutopopulist” coalition in his campaigns, drawing money from both wealthy elites and an unprecedented surge of small, first-time donors. With Tuesday’s elections, though, that coalition seems to be fraying. What happened? 1/ cup.org/4cfm0Az
Plutopopulism: Wealth and Trump’s Financial Base SEAN KATES, ERIC MANNING, TALI MENDELBERG and OMAR WASOW Comparative scholarship suggests authoritarian candidates often rely on backing from the wealthy. The wealthy are also said to play an important role in American campaign finance. Studies of Donald Trump, however, found that he drew significant support from white Americans with less education and privilege. We evaluate wealthy and non-wealthy Americans’ financial support for Trump, compared to other candidates, by constructing a comprehensive dataset of property values matched to contributions and voter files. We find Trump underperformed among wealthy Republican donors while mobilizing new non-wealthy donors. Trump also diversified the donorate, especially by education. That is, Trump built an unusual coalition of wealthy and non-wealthy donors. Our results support an alternative, “plutopopulist” model of Trump’s financial base.
November 8, 2025 at 4:39 PM
owasow.bsky.social
“Can political activism foster electoral participation?” Study examines “1913 Pilgrimage, a large-scale nationwide march in support of women’s suffrage.” Finds “women’s activism can drive political participation even in the virtual absence of women politicians.” www.cambridge.org/core/journal...
Activism and the Electoral Participation of Women Mona Morgan-Collins and Valeria Rueda Abstract Can political activism foster electoral participation? We investigate this question by examining the role of the British suffragists in fostering women’s electoral participation. While scholars have shown that women politicians increase women’s participation in politics, less attention has been paid to the role of activists. We fill this gap by studying the 1913 Pilgrimage, a large-scale nationwide march in support of women’s parliamentary suffrage. Using a novel database of geocoded electoral registers, we employ a differences-in-differences strategy to show that proximity to the Pilgrimage increased women’s registration in local elections. To explain this effect, we show that exposure to women activists spurred other women’s political mobilization, and that our results are not driven by an endogenous selection of the march path and cannot be attributed to mobilization by other movements or politicians. Our findings highlight that women’s activism can drive political participation even in the virtual absence of women politicians. Figure 2. Map of the march and data availability at the division level. Note: March path in sampled counties is along main roads connecting stopping points and along straight lines outside the sample (see Appendix Figure C.1 for stopping points). Figure 3. The effects of the march using different treatment definitions. Note: Plots the coefficient of interest β, defined using different distance buffers; ninety per cent and ninety-five per cent CIs; standard errors clustered at the parliamentary division level, using Model 2 in Table 3
November 2, 2025 at 9:00 PM
owasow.bsky.social
Hard to see the batter from these seats but can’t beat the view
Photo of Dodgers Stadium from an airplane window.
October 29, 2025 at 3:33 AM
owasow.bsky.social
Lots of reasonable skepticism so some personal data. Plot is V02max estimated by Apple Watch. Notice:
— Early summer, working on a paper, running dropped from 4x per week to 1x, a significant decline
— Late summer, running again, recovery + plateau
— Recently, with 2x20second sprints, a big increase
A screenshot of the Apple Health app’s “Cardio Fitness” graph showing VO₂ max over the past six months. The pink line trends downward from about 48 mL/kg/min in April to a low around 42 mL/kg/min in July, then gradually climbs back up to 49.5 mL/kg/min by mid-October 2025. The selected time period is October 12–18, 2025, with the current value labeled “49.5 VO₂ max.” The “High” fitness level category and “26 weeks” indicator appear below the chart, along with options labeled “Show All Cardio Fitness Levels” and “Update Your Data.” The time-range selector at the top highlights “6M” for six months.
October 22, 2025 at 7:59 PM
owasow.bsky.social
Small study found thirty bouts of four second sprints at all-out intensity, with short recovery time over eight weeks, “elicited a 13.2% increase in VO2peak…a 7.6% increase in total blood volume. Concurrently, maximal anaerobic power increased by 17.2%.” media-site.carolbike.com/documents/Fo...
ABSTRACT: High-intensity interval training (HIIT) is an effective tool to improve cardiovascular fitness and maximal anaerobic power. Different methods of HIIT have been studied but the effects of repeated maximal effort cycling with very short exercise time (i.e., 4-s) and short recovery time (15-30 s) might suit individuals with limited time to exercise. Purpose: We examined the effects of training at near maximal anaerobic power during cycling (PC) on maximal anaerobic power, peak oxygen consumption (VO2peak), and total blood volume in 11 young healthy individuals (age: 21.3±0.5 y) (6 men, 5 women). Methods: Participants trained three times a week for eight weeks performing a PC program consisting of 30 bouts of 4-s at an all-out intensity (i.e., 2 minutes of exercise per session). The cardiovascular stress progressively increased over the weeks by decreasing the recovery time between sprints (30 to 24 to 15-s) and thus total session time decreased from 17 to < 10 min. Results: PC elicited a 13.2% increase in VO2peak (Pre: 2.86±0.18 L/min, Post: 3.24±0.21 L/min, (p=0.003) and a 7.6% increase in total blood volume (Pre: 5,139±199 ml, Post: 5,529±342 ml, p<0.05). Concurrently, maximal anaerobic power increased by 17.2% (Pre: 860±53 watts, Post: 1,009±71 watts; p<0.001). Conclusion: A Power Cycling training program employing 30 bouts of 4-s duration for a total of 2 min of exercise, resulting in a total session time of less than 10 min in the last weeks, is effective for improving total blood volume, VO2peak and maximal anaerobic power in young healthy individuals over 8 weeks. These observations require reconsideration of the minimal amount of exercise needed to significantly increase both maximal aerobic and anaerobic power.
October 22, 2025 at 6:54 PM
owasow.bsky.social
Echoes of the 1980s: “Trump himself ensured workers were told to remove the bronze latticework with blowtorches, separate the friezes from the walls with jackhammers, and throw them down into the interior of the building where they shattered into a million pieces.” news.artnet.com/art-world/do...
Headline: When Donald Trump Razed the Bonwit Teller Building, He Promised the Met Its Art Deco Friezes. A New Book Details How He Pulverized Them Instead
October 20, 2025 at 8:27 PM
owasow.bsky.social
“Nice university you got there…”

Gift link: www.nytimes.com/2025/10/03/u...
“U.S. universities were and are the envy of the world,” Mr. Rowan told Bloomberg Television. “We can destroy that. We can lose it.”
October 4, 2025 at 2:12 PM
owasow.bsky.social
“Contrary to existing theories of protest mobilization, we find that teacher strikes decrease voter turnout, with a substantially larger effect among Republicans than Democrats. Strikes decrease Republican turnout by 3 percentage points, Democratic turnout by 1pp” edworkingpapers.com/sites/defaul...
Title: Teacher Strikes and the Demobilization of Republican Voters Melissa Arnold Lyon University at Albany, SUNY Leslie K. Finger University of North Texas Hyesang Noh University at Albany, SUNY September 2025 AbstractProtests can mobilize both supporters and opponents. Extant research suggests that disruptiveprotests are particularly likely to mobilize opponents, yet strikes—one of the most disruptiveforms of protest—have been largely absent from this literature. We use an original dataset of 716teacher strikes from 2007-2020 to examine the causal effects of teacher strikes on politicalparticipation. Contrary to existing theories of protest mobilization, we find that teacher strikesdecrease voter turnout, with a substantially larger effect among Republicans than Democrats.Strikes decrease Republican turnout by 3 percentage points, on average, while they reduceDemocratic turnout by 1 percentage point. We further show how the effect varies depending onthe parental status and across different types of elections. Additionally, we find that theproximity and duration of strikes also matter, with closer or longer strikes having a moresignificant demobilizing effect. These effects are driven by decreased support for teacherdemands and reduced political efficacy. Our findings reveal that different forms of protest actionmay produce distinct patterns of political behavior, suggesting the need to expand theories ofprotest beyond their present focus on inspiring third party mobilization. Chloropleth map of The Geographic Distribution of Teacher Strikes, 2007-2020. Figure 2. Effects of Strikes by Subgroup. Two side-by-side dot-and-whisker plots titled “Effect of Strikes by Subgroup.” Panel A shows voter turnout effects by partisanship. Teacher strikes are associated with lower turnout across groups: about −3 percentage points for Republicans, −1 point for Democrats, and −2 points for Independents, with 90% confidence intervals that do not cross zero. Panel B further splits by parental status. Republican parents show the largest negative turnout effect, Democratic parents a smaller negative effect, Republican non-parents a modest decline, and Democratic non-parents nearly no effect. Horizontal axis shows the effect of strikes from −0.06 to +0.02. Confidence intervals calculated at the 90% level.
September 22, 2025 at 4:06 PM
owasow.bsky.social
“Since June, the share of adults who say the country is on the wrong track increased 13 percentage points from 62% to 75%. The shift occurred primarily among Republicans.”
apnorc.org/projects/pes...
Line chart titled “Pessimism among Republicans spikes for the first time since Trump’s second inauguration.” It shows the percent of adults saying the country is headed in the wrong direction, by party, from January 2024 to September 2025. Overall pessimism stays mostly steady around 70–75%. Republicans are highly pessimistic in 2024 (near 90%), then drop sharply to 26% in early 2025 before rebounding to 51% by September 2025. Democrats show the opposite pattern: around 50% in 2024, jumping to about 90% by late 2024 and remaining high through 2025. The latest poll (Sept 2025) shows 75% overall, 92% of Democrats, and 51% of Republicans saying “wrong direction.”
September 20, 2025 at 1:17 PM
owasow.bsky.social
“73% of all mass shootings in developed countries occur in the U.S.” www.theviolenceproject.org/key-findings/
Horizontal bar chart comparing the number of mass shootings in different countries from 2000–2022. The United States dominates with 109 incidents, far exceeding all others. France follows with 6, Germany 5, Canada 4, Finland 3, Austria 2, Switzerland 2, Italy 2, the United Kingdom 2, the Netherlands 2, the Czech Republic 2, Norway 1, and Belgium 1. The title emphasizes that the U.S. leads the world in mass shootings during this period.
September 16, 2025 at 3:57 AM
owasow.bsky.social
For more, follow co-authors @jkertzer.bsky.social, Dara Kay Cohen, Dorothy Manevich, and @zeitzoff.bsky.social

jkertzer.sites.fas.harvard.edu/Research_fil...
Title: “If Abortions Aren’t Safe, Neither Are You”: Public Support for Violent Tactics in U.S. Abortion Politics Joshua D. Kertzer,1 Dara Kay Cohen,2 Dorothy Manevich,3 and Thomas Zeitzoff4 September 12, 2025 Abstract: What drives support for political violence: opportunity structures, moralized attitudes and strong identities, or general predispositions? We examine these questions in the context of abortion politics, which has been understudied by political violence scholars despite abortion constituting one of the most frequent targets of politically violent attacks in the United States. Using nationally representative survey data and an embedded survey experiment, we find that Americans view violent tactics as fundamentally distinct from conventional protest. Support for such tactics has little to do with the side respondents take on abortion, or the extremity or moralization of their abortion attitudes and identities. Rather, support for violent tactics is better predicted by general predispositions toward aggression and hostile sexism. Taken together, our findings suggest that an appetite for violence is concentrated not among the most ardent believers, but among a small subset who hold more aggressive and prejudiced attitudes across a range of issues.
September 15, 2025 at 9:02 PM
owasow.bsky.social
“The most significant predictors of support for violent tactics regarding abortion are psychological dispositions not directly related to abortion itself. Respondents’ belief that it is justified for people to use violence to pursue their political goals…displays the strongest association…”
Coefficient plot of predictors of support for violent tactics. Four panels: Macropolitical context: Pro-choice, abortion scope, and abortion legality attitudes show small positive associations. Strength of identities/attitudes: Extremity of abortion scope and legality, and abortion identification strength, show small associations. Dispositions: Political aggression and hostile sexism have strong positive associations (~0.3–0.35). Benevolent sexism is near zero. Demographics: Variables like race, party ID, parent status, sex, income, education, church attendance, and age show little effect. Two series are shown: black points for factor score models, gray points for additive models, both with 95% confidence intervals.
September 15, 2025 at 8:50 PM
owasow.bsky.social
“Public approval of conventional protest tactics is very high, with an overwhelming majority of respondents expressing approval of protests and rallies. Second, public approval of
disruptive or violent tactics is extremely low…”
Source: jkertzer.sites.fas.harvard.edu/Research_fil...
Bar chart of reported acceptability of contentious tactics. The y-axis shows average acceptability ratings (0–0.8). Conventional tactics (protest, rally, clinics, boycott, etc.) in red are rated most acceptable (around 0.55–0.7). Targeted harassment tactics (e.g., protest at home, doxxing, blocking traffic, hacking) in green are less acceptable (around 0.15–0.25). Disruptive/violent tactics (e.g., threats, vandalism, firebombing, physical attacks) in blue are rated lowest (under 0.1). Bars include 95% confidence intervals.
September 15, 2025 at 8:44 PM
owasow.bsky.social
“Support for violent tactics is better predicted by predispositions toward aggression and hostile sexism…findings suggest an appetite for violence is concentrated not among the most ardent believers, but among a small subset who hold more aggressive and prejudiced attitudes across range of issues.”
Title: If Abortions Aren’t Safe, Neither Are You”: Public Support for Violent Tactics in U.S. Abortion PoliticsJoshua D. Kertzer,1 Dara Kay Cohen,2 Dorothy Manevich,3 and Thomas Zeitzoff4 September 12, 2025 Abstract: What drives support for political violence: opportunity structures, moralized attitudes and strong identities, or general predispositions? We examine these questions in the context of abortion politics, which has been understudied by political violence scholars despite abortion constituting one of the most frequent targets of politically violent attacks in the United States. Using nationally representative survey data and an embedded survey experiment, we find that Americans view violent tactics as fundamentally distinct from conventional protest. Support for such tactics has little to do with the side respondents take on abortion, or the extremity or moralization of their abortion attitudes and identities. Rather, support for violent tactics is better predicted by general predispositions toward aggression and hostile sexism. Taken together, our findings suggest that an appetite for violence is concentrated not among the most ardent believers, but among a small subset who hold more aggressive and prejudiced attitudes across a range of issues.
September 15, 2025 at 8:39 PM
owasow.bsky.social
According to the “Global Terrorism Database (GTD)—arguably the most widely-used dataset in the study of political violence—anti-abortion activists constitute the single most common perpetrators of terrorist attacks in the United States.”

Source: jkertzer.sites.fas.harvard.edu/Research_fil...
Bar chart showing terrorist attacks in the United States by perpetrator, 1970–2020. The top panel highlights the ten most active groups, with Anti-Abortion extremists leading at over 200 attacks, followed by Left-Wing Militants, White supremacists/nationalists, FALN, NWLF, Black Nationalists, ALF, JDL, Student Radicals, and ELF. The middle panel shows abortion-related groups, where Anti-Abortion extremists dominate compared to smaller groups like Army of God, Phineas Priesthood, Pro-choice extremists, and Christian Liberation Army. The bottom panel shows Islamist/Jihadi groups, led by Jihadi-inspired extremists (about 40 attacks), with smaller contributions from Muslim extremists, Jamaat-al-Fuqra, Al-Qaida, Hanafi Muslims, AQAP, TTP, and Islamist extremists. Bars are color-coded: blue for abortion-related groups, red for Islamist/Jihadi groups, and gray for others. X-axis is number of attacks, Y-axis lists groups.
September 15, 2025 at 8:33 PM
owasow.bsky.social
Free buses seriously, but not literally www.nytimes.com/2025/09/09/n...
Excerpt from NYTimes article on polling for Mamdani in New York: Some 57 percent of likely voters also said New York City should not follow through on Mr. Mamdani’s plan to make buses free for everyone, even as a majority of New Yorkers support the proposal in theory.
September 9, 2025 at 3:04 PM
owasow.bsky.social
Figure 4 from Zhao et al. speaks most closely to mechanisms when the body tries to break down acetaminophen. Most routes are safe, but some kids—especially with certain impairments—struggle with key pathways, leading more of the drug to turn into a harmful toxin that’s harder for them to clear.
Flowchart showing the metabolism of acetaminophen (APAP) in humans. At the center, APAP branches into three metabolic pathways. Glucuronidation (left, green arrow): Produces beta-D-glucusiduronic acid, labeled as non-toxic and eliminated from the body. This is not the major pathway in babies. Sulfation (right, green arrow): Produces a sulfate conjugate, also non-toxic and eliminated. This is the dominant pathway in babies but is impaired in many children with autism spectrum disorder (ASD). Oxidative pathway (downward red arrow, cytochrome P450-dependent): Produces NAPQI, labeled “highly toxic” with a skull-and-crossbones icon. From NAPQI, a green arrow shows detoxification through glutathione, producing a harmless conjugate, but this glutathione-dependent detoxification is impaired in many children with ASD. Overall, the diagram highlights that when sulfation and glutathione pathways are impaired, more APAP is shunted into the oxidative pathway, increasing NAPQI accumulation and toxicity.
September 8, 2025 at 4:26 AM
owasow.bsky.social
Two quick points: first, meta-analysis cited above does not appear to rely only on retrospective studies. ehjournal.biomedcentral.com/articles/10....

Second, agree ”how we parse as a society“ is a key question and my point is we’re not well suited to debates that cross-cut science and partisanship.
Following recommendations on causal inference in environmental epidemiology[18], we set our findings within a triangulation framework: when distinct study designs that suffer different, ideally opposing, biases reach congruent results, causal inference is strengthened. Our analysis integrates (i) prospective cohorts susceptible to residual confounding but strong temporality, (ii) biomarker-based studies with low recall bias but possible misclassification of dose, and (iii) experimental models largely free of confounding yet limited in external validity. The convergence of these independent sources of evidence increases confidence that the observed associations are not artefactual.
September 7, 2025 at 7:51 PM
owasow.bsky.social
This paper offers a survey of possible mechanisms (See Table 1). “Acetaminophen causes neurodevelopmental injury in susceptible babies and children: no valid rationale for controversy” pmc.ncbi.nlm.nih.gov/articles/PMC...
Early life exposure of susceptible individuals to APAP (acetaminophen/paracetamol) causes neurodevelopmental injury. ⸻ Evidence Categories and Supporting Points Animal Studies 1. Behavioral changes in rat pups 2. Male susceptibility in rat pups 3. Killing of cortical neurons in adult lab animals Mechanism of Action 18. Role of immunity: Exception of CF 19. Role of immunity: Oxidative stressors 20. Established molecular mechanism of APAP toxicity Postnatal Observations 14. Schultz study: APAP + vaccine 15. Observations of parents 16. Schultz study: Regression with APAP use Associations with Human Activities 5. Circumcision 6. Korean manufacturing problems 7. Modern Jewish religious practice 8. Postnatal connection (DNBC) Associations in Time 9. Shift toward regression in 1980 with switch from aspirin 10. Rise starting in 1980 with switch from aspirin 11. Increase with direct-to-consumer advertising 12. Prenatal connection (13 studies) 13. Association: Cord blood APAP Other Considerations 4. Never proven safe 5. Anti-social effect in adult humans ⸻ Summary The diagram argues that multiple lines of evidence converge on the hypothesis that early-life APAP exposure may cause neurodevelopmental injury. Evidence comes from animal studies (behavioral and neuronal changes), mechanistic work (toxicity pathways, oxidative stress, immune effects), postnatal human observations (parental reports, epidemiological studies), associations with human practices (circumcision, religious customs, manufacturing), and historical associations over time (switch from aspirin around 1980, advertising, prenatal exposure). Additionally, APAP is noted as never proven safe, with some studies pointing to anti-social effects in adults.
September 7, 2025 at 5:52 PM
owasow.bsky.social
From the meta-analysis in the original post, some concerns raised about the Swedish sibling study:
A third, large prospective cohort study conducted in Sweden by Ahlqvist et al. found that modest associations between prenatal acetaminophen exposure and neurodevelopmental outcomes in the full cohort analysis were attenuated to the null in the sibling control analyses [33]. However, exposure assessment in this study relied on midwives who conducted structured interviews recording the use of all medications, with no specific inquiry about acetaminophen use. Possibly as a resunt of this approach, the study reports only a 7.5% usage of acetaminophen among pregnant individuals, in stark contrast to the ≈50% reported globally [54]. Indeed, three other Swedish studies using biomarkers and maternal report from the same time period, reported much higher usage rates (63.2%, 59.2%, 56.4%) [47].
September 7, 2025 at 6:18 AM
owasow.bsky.social
Critique of the Swedish study
September 7, 2025 at 6:16 AM
owasow.bsky.social
“A racial reckoning?” New study by @amengel.bsky.social & Cindy Kam:

“Challenging the conventional wisdom, our analyses demonstrate that racial attitudes changed following George Floyd’s murder, but in ways dependent upon attitude measure and population subgroup.” www.cambridge.org/core/journal...
A racial reckoning? racial attitudes in the wake of the murder of George Floyd Andrew M. Engelhardt and Cindy D. Kam Abstract Did George Floyd’s murder and its ensuing protests produce a racial reckoning? Conventional social-science accounts, emphasizing the stability of racial attitudes, dismiss this possibility. In contrast, we theorize how these events may have altered Americans’ racial attitudes, in broadly progressive or in potentially countervailing ways across partisan and racial subgroups. An original content analysis of partisan media demonstrates how the information environment framed Black Americans before and after the summer of 2020. Then we examine temporal trends using three different attitude measures: most important problem judgments, explicit favorability towards Whites versus Blacks, and implicit associations. Challenging the conventional wisdom, our analyses demonstrate that racial attitudes changed following George Floyd’s murder, but in ways dependent upon attitude measure and population subgroup. Figure 1 shows four scatterplots with Lowess smoothing lines comparing Fox (gray squares, gray line) and MSNBC (black circles, black line) coverage of Black Americans in 2020, with a vertical red line marking George Floyd’s murder on May 25. Top left (Daily Counts of Activism Frames): Both networks show a sharp spike in activism-related mentions immediately after Floyd’s murder, then declining through the year, with similar levels across Fox and MSNBC. Top right (Daily Proportion of Mentions containing Activism Frames): The share of mentions with activism frames rises briefly after Floyd’s murder for both networks but quickly declines, with no clear partisan difference. Bottom left (Daily Counts of Backlash Frames): Fox shows a much larger spike in backlash-related mentions (e.g., “violence,” “mob”) after May, while MSNBC increases more modestly. Bottom right (Daily Proportion of Mentions containing Backlash Frames): From June to October, nearly half of Fox’s mentions of Black Americans include backlash frames, roughly twice the rate of MSNBC, which remains lower throughout. Overall, the figure shows that while both networks used more activism frames immediately after Floyd’s murder, Fox emphasized backlash frames much more heavily than MSNBC in the following months. Figure 2 shows six scatterplots with Lowess smoothing lines tracking mentions of racism or race relations as the most important problem in Gallup polls (2017–2021). Each dot is a monthly estimate, with a vertical red line marking George Floyd’s murder in May 2020. Top row: Full Sample: Mentions are low (under 5%) before 2020, then spike sharply in June 2020 (~16%) before falling but remaining above pre-2020 levels. Among Whites: Similar pattern as the full sample, with a ~10-point jump in June 2020, followed by a decline but remaining elevated relative to earlier years. Among Blacks: Higher baseline concern compared to Whites; mentions spike by ~21 points after Floyd’s murder and remain elevated through 2021. Bottom row: Among White Republicans: Very low pre-2020 mentions, a modest rise (~6 points) in June 2020, then rapid decline toward baseline. Among White Independents: Clear but moderate spike in June 2020, with some persistence above baseline. Among White Democrats: Low pre-2020 mentions, sharp June 2020 spike (~14 points), then decline but sustained higher levels through 2021. Overall, the figure shows a sharp discontinuity after Floyd’s murder across all groups, with the most sustained increases among Black respondents and White Democrats, and weaker persistence among White Republicans. Figure 4 presents six scatterplots of predicted weekly average IAT D-scores for 2019 (gray squares, gray line) and 2020 (black circles, black line), with a vertical red line marking George Floyd’s murder on May 25. Higher D-scores indicate stronger implicit anti-Black bias. Top row: Full Sample: Bias was declining before Floyd’s murder and drops further afterward in 2020, diverging from 2019. Among Whites: A clear decline in D-scores appears after Floyd’s murder, with lower bias sustained through 2020. Among Blacks: No sharp discontinuity; scores remain stable across 2019 and 2020. Bottom row: Among White Conservatives: Noticeable decline in anti-Black bias after May 2020, sustained through the year. Among White Neutrals: A clear drop in bias after Floyd’s murder, somewhat larger than among liberals. Among White Liberals: Small decline in bias after May 2020, though less pronounced than among conservatives or neutrals. Overall, the figure shows that implicit anti-Black bias decreased among White respondents across ideological groups after Floyd’s murder, with larger reductions among conservatives and neutrals than liberals. Black respondents show little change, consistent with already low baseline bias.
August 31, 2025 at 3:33 PM