<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Tufts Public Opinion Lab on Medium]]></title>
        <description><![CDATA[Stories by Tufts Public Opinion Lab on Medium]]></description>
        <link>https://medium.com/@tufts-pol?source=rss-e4754e5edb87------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*b0wiVeQ_iiOy8Z-iZUsm8w.jpeg</url>
            <title>Stories by Tufts Public Opinion Lab on Medium</title>
            <link>https://medium.com/@tufts-pol?source=rss-e4754e5edb87------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 20:30:09 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@tufts-pol/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Tufts Public Opinion Lab national pre-election poll shows a big lead for Democrats in the race for…]]></title>
            <link>https://tufts-pol.medium.com/tufts-public-opinion-lab-national-pre-election-poll-shows-a-big-lead-for-democrats-in-the-race-for-57656e85c738?source=rss-e4754e5edb87------2</link>
            <guid isPermaLink="false">https://medium.com/p/57656e85c738</guid>
            <category><![CDATA[polls-and-surveys]]></category>
            <category><![CDATA[public-opinion]]></category>
            <category><![CDATA[2026-elections]]></category>
            <dc:creator><![CDATA[Tufts Public Opinion Lab]]></dc:creator>
            <pubDate>Thu, 29 Oct 2026 13:01:02 GMT</pubDate>
            <atom:updated>2026-10-29T13:01:02.829Z</atom:updated>
            <content:encoded><![CDATA[<h3>Tufts Public Opinion Lab national pre-election poll shows a big lead for Democrats in the race for the U.S. House</h3><p>With less than a week to go before the 2026 midterm elections, a new national poll from the Tufts University Public Opinion Lab shows Democrats poised to win a large victory in U.S. House races across the country. Among likely voters in the poll, Democratic candidates held a 12 point lead over Republicans, 52% to 40%. Five percent of likely voters remain undecided. The margin of error for likely voters is 4.6 percentage points.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LxYX0oWp1dXc9fEIG0R8-g.png" /></figure><p>Unlike most polls which use a generic house ballot question that only provides the party labels, our House vote question piped in candidate names (with their party labels) to more accurately reflect how people actually vote in these races.</p><p>The Tufts Public Opinion Lab poll was fielded online with <a href="https://yougov.com/">YouGov</a>. The poll was fielded on October 23rd — 27th, 2024. The total sample size for the poll is 1,000 American adult respondents. Vote estimates are given for likely voters using a probabilistic likely voter model. Toplines, crosstabs, and methodological details for the poll can be found <a href="https://tufts.box.com/s/mz81q5uqscd3m82y02dg4c50x4mm4n8i">here</a>.</p><p>You can analyze the poll results with <a href="https://bfschaffner.shinyapps.io/tpol2026/">this interactive tool</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=57656e85c738" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Is there a partisan divide when it comes to AI?]]></title>
            <link>https://tufts-pol.medium.com/is-there-a-partisan-divide-when-it-comes-to-ai-cafd0b49a686?source=rss-e4754e5edb87------2</link>
            <guid isPermaLink="false">https://medium.com/p/cafd0b49a686</guid>
            <category><![CDATA[public-opinion]]></category>
            <category><![CDATA[political-parties]]></category>
            <category><![CDATA[polarization]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Tufts Public Opinion Lab]]></dc:creator>
            <pubDate>Thu, 21 May 2026 13:59:17 GMT</pubDate>
            <atom:updated>2026-05-21T13:59:17.116Z</atom:updated>
            <content:encoded><![CDATA[<p><em>by Gunnar Ivarsson (‘28), Evan Russell (‘29), Jackson Deans (‘26), and Matthew Sage (‘26)</em></p><p><em>Are large language models (LLMs) becoming “left language models”? </em>A 2025 report by the Manhattan Institute found that many major-market artificial intelligence (AI) models <a href="https://manhattan.institute/article/measuring-political-preferences-in-ai-systems-an-integrative-approach">exhibit left-leaning bias in their outputs</a>. Artificial intelligence is becoming increasingly integrated across all sectors of society, including politics, and trust in the political system and technology is rarely stable. As AI becomes more integrated into major political strategy, understanding bias is critical, but the next dimension of that question is understanding who is consuming AI products and what trust Americans have in LLMs.</p><p>If one side of the aisle views AI as a tool for progress while the other views it as a biased gatekeeper, the resulting technological rift could further expand already polarized partisan silos. So, what is the effect of party identification on usage and trust of artificial intelligence tools? We expected to find that identifying as a Democrat is associated with both more frequent use of AI and greater trust in its development than Republicans and Independents.</p><p><strong>How we conducted our analysis</strong></p><p>To test the relationship between political identity and technology adoption, we use the Pew Research Center’s American Trends Panel (ATP). Specifically, we analyze <a href="https://www.pewresearch.org/dataset/american-trends-panel-wave-152/">Wave 152</a>, conducted from August 12th to August 18th, 2024. This survey gathered responses from a representative sample of 5,410 U.S. adults about respondents’ use and attitudes toward artificial intelligence. We define the independent variable as party identification. Respondents were categorized as “Democrat,” Republican,” or “Independent.” For our dependent variable, we observe AI Usage, measured on a three-point scale, being “Daily,” “Sometimes,” and “Almost Never.” AI Trust is also measured on a three-point scale in our study, ranging from “High,” “Medium,” and “Low” levels of trust in AI development.</p><p>We suspect two primary mechanisms may be driving a relationship beyond partisanship, which must be addressed to understand the partisan divide in AI use and trust.</p><p>(1) College-educated individuals who are disproportionately Democratic are likely to work in environments using AI.</p><p>(2) Ideological bias of AI systems may make AI less trustworthy to Republicans. Together, these dynamics might suggest that partisan perceptions may shape AI use.</p><p>We controlled for various factors that may affect the causal relationship between AI use and respondents’ party identification. These control variables include education, age, and family income. This allows us to estimate the isolated effect of party identification on an individual’s usage of and trust in AI tools.</p><p><strong>What we found</strong></p><p>The study compares descriptive frequencies with multivariate regression analysis to distinguish between general trends and isolated partisan effects. A primary observation from the descriptive data reveals a clear partisan divide in the frequency of AI tool use. Republicans are significantly more likely to report “Almost Never” when using AI (exceeding 45% of the party subgroup) compared to Democrats and Independents, hovering around 40%. On the other hand, Democrats and Independents show higher rates of “Daily” and “Sometimes” usage than their Republican counterparts. These points suggest that AI integration into American life is indeed being settled more rapidly among those on the left and center of the political spectrum.</p><p>While usage rates vary, trust levels show a different pattern. We find relatively uniform frequencies for the amount of trust each party has in AI tools. Despite theoretical concerns about “Left Language Models,” baseline trust in AI remains similar across all three partisan groups. It could suggest that skepticism may be more tied to practical utility or cultural signaling than a rejection of AI.</p><p>To confirm that these trends are driven by political identity rather than other factors, we turn to the results of our regression models for AI Usage and AI Trust. Partisanship remains a significant predictor even when controlling for socioeconomic factors. The regression shows that Republican identity is associated with an average lower response of 0.066 on a 1–3 scale of AI use compared to Democrats. Our models indicate that Independents’ use of AI tools is at rates very similar to Democrats, further isolating Republicans as the primary group exhibiting less AI interaction.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/938/1*PEsJnQyFUizlvo3i5rJIrA.png" /></figure><p>Our regression model supports our hypothesis that Democrats use AI tools slightly more than Republicans. The Republican 0.066 decrease in AI use on a 1–3 scale is statistically significant (p-value = 0.02), but the raw difference remains relatively small within the broader context of AI adoption across political groups. Our model shows that Independents use AI tools at similar rates to Democrats and do not exhibit a statistically significant partisan effect on trust in AI.</p><p>The biggest caveat for our results is their time-bound nature. AI has advanced rapidly in capability and applicability across the workforce over the past two years, meaning the dataset — gathered in 2024 — may no longer accurately reflect Americans’ opinions and habits in 2026. Future research should rely on updated survey data to more accurately address the research question.</p><p>Regardless, our study contributes to an understanding of the partisan divide over AI policy. Further research should investigate why, despite their less frequent use of AI tools, <a href="https://www.nytimes.com/2026/03/20/us/politics/white-house-unveils-ai-policy-aimed-at-blocking-state-laws.html">Republicans are less likely to support AI regulation policies</a>. Though there is only a slight difference in AI use between Republicans and Democrats, all those using these models — typically left-leaning — might become more exposed to more left-leaning political viewpoints as a result. We might see in the future that individuals identifying as Democrats or Republicans both increasingly consume more left-leaning information. As AI continues to permeate the American political system, the stakes for understanding these adoption patterns will only grow.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cafd0b49a686" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Who is driving polarization? The role of ideological extremity in communication and funding]]></title>
            <link>https://tufts-pol.medium.com/who-is-driving-polarization-the-role-of-ideological-extremity-in-communication-and-funding-52926c4fd2c8?source=rss-e4754e5edb87------2</link>
            <guid isPermaLink="false">https://medium.com/p/52926c4fd2c8</guid>
            <category><![CDATA[campaign-finance]]></category>
            <category><![CDATA[ideology]]></category>
            <category><![CDATA[political-polarization]]></category>
            <category><![CDATA[partisanship]]></category>
            <category><![CDATA[negative-partisanship]]></category>
            <dc:creator><![CDATA[Tufts Public Opinion Lab]]></dc:creator>
            <pubDate>Tue, 19 May 2026 13:40:50 GMT</pubDate>
            <atom:updated>2026-05-19T13:40:50.721Z</atom:updated>
            <content:encoded><![CDATA[<p><em>by Fatima Eisa (‘26), Shayna Levy (‘28), Chloe Walker (‘28), and Eliot Wertheimer (‘29)</em></p><p>Long defined by the two-party system, American politics is no stranger to polarization and partisanship. Fundamental disagreements and bitter fights across party lines have long been a staple of policymaking in Congress, but data shows that something has shifted recently. A recent article from the <a href="https://www.pewresearch.org/politics/2022/08/09/republicans-and-democrats-increasingly-critical-of-people-in-the-opposing-party/">Pew Research Center</a> shows that in the past six years, there has been a “particularly pronounced increase” in how many partisans describe members of the other party as immoral. More research from <a href="https://www.pewresearch.org/short-reads/2024/06/25/americans-see-little-bipartisan-common-ground-but-more-on-foreign-policy-than-on-abortion-guns/">Pew </a>conducted in May 2024 shows that “just 18% of Americans say there is at least some common ground between Republicans and Democrats on abortion,” and very few respondents see bipartisanship on other issues, including gun policy and immigration. Simultaneously, as a recent article from the <a href="https://www.washingtonpost.com/ripple/2026/04/18/can-primary-reform-keep-out-extremist/?itid=hp_brand_promo_3">Washington Post</a> reflects, ideological extremists are being blamed for this polarization.</p><p>In this post, we use empirical analysis to identify whether certain ideological groups in Congress are particularly driving this increase in polarization. Specifically, we analyzed the association between ideological extremity and personal-attack-focused communications, bipartisanship-focused communications, and out-of-state funding. These behaviors, or lack thereof, are often classified as polarizing political behaviors.</p><p>We tested three hypotheses about the association between ideological extremity and Congressional behaviors contributing to political polarization. First, we expected that more ideologically extreme members of Congress would have a greater proportion of personal-attack-focused communications. Second, we expected that more ideologically extreme members of Congress would have a lesser proportion of bipartisanship-focused communications. Finally, we expected that more ideologically extreme members of Congress would have a greater proportion of out-of-state funding.</p><p><strong>Data and Methodology</strong></p><p>Our analsysi draws on the <a href="https://americaspoliticalpulse.com/data/elite/profiles.zip">Legislator Profiles dataset</a> from the <a href="https://americaspoliticalpulse.com/data#section=elites">American Political Pulse Polarization Research Lab</a>, which includes data on all 535 current members of Congress. The dataset provides ideology scores, detailed campaign finance information, and classifications of legislators’ public communications since 2020. Communications are classified as either personal attack-focused, policy-focused, bipartisanship-focused, credit claiming, or other. Our independent variable is ideological extremity, measured as the absolute value of each member of Congress’s ideology score derived from the <a href="https://jbduckmayr.com/publication/ggum/ggum.pdf">Duck-Mayr &amp; Montgomery (2022) scaling method</a>. Taking the absolute value allows us to capture extremity regardless of direction, meaning a very conservative and a very liberal legislator with equivalent ideological distances from the center are treated the same. This variable is measured on a 0 to 1 scale, where values closer to 1 indicate greater ideological extremity.</p><p>We examine how ideological extremity predicts three dependent variables, all measured as proportions from 0 to 1. The first is the proportion of personal-attack-focused communications, which is calculated by dividing the total number of communications classified as personal attacks by the total number of classified communications. The second is the proportion of bipartisanship-focused communications, calculated the same way using attack-focused classifications. The third is the proportion of out-of-state funding, which is found by dividing the total amount of money raised from donors residing outside the legislator’s state by their total campaign fundraising.</p><p>To isolate the effect of ideological extremity on each dependent variable, we control for several factors in our regression models. We control for the political party of the legislator, the total number of communications, and how competitive their 2026 election is expected to be (using the <a href="https://www.cookpolitical.com/">Cook Political Report</a> ratings). The competitiveness is coded on a 1 to 4 scale, where 1 represents a Toss Up (most competitive), and 4 represents a Solid Democrat or Solid Republican seat (least competitive). Including this variable allows us to account for the fact that legislators in more competitive races may behave differently than those in safe seats.</p><p><strong>What we found</strong></p><p>The results of our regression analysis are shown in the figure below. Overall, we find a clear relationship between ideological extremity and both communication strategy and fundraising. As members of Congress become more ideologically extreme, they rely more heavily on attack-focused messaging and less on bipartisan communication. Moving from the least to most extreme members is associated with a 5.2 percentage point increase in the share of attack-focused communications and a 10.6 percentage point decrease in bipartisan messaging. We also find that ideological extremity is associated with a 23.7 percentage point decrease in the share of out-of-state funding. All of these relationships are statistically significant.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/942/1*N-lXCNYdzwTTCqki-r9GNw.png" /></figure><p>These effects are also substantively meaningful. In a political environment with thousands of campaign communications and billions of dollars raised each cycle, shifts of this magnitude can meaningfully alter how candidates present themselves to voters and how they engage with broader political networks.</p><p>Finally, our results present no statistically significant relationship between the competitiveness of a race and communication strategy. While this lack of statistical significance does not mean ideology entirely shapes communication strategy, it suggests that individual characteristics, such as ideological positioning, may be particularly important in shaping how candidates communicate, independent of electoral contexts.</p><p><strong>What this means</strong></p><p>Our findings present notable implications for the causes of American political polarization, which can later be drawn on to find solutions. While we cannot infer causation from our study, our data reveal two distinct patterns: ideological extremists employ more polarizing communication strategies, but receive less on out-of-state funding than ideological moderates. The first pattern suggests that ideological extremists contribute to polarizing discourse, while the second challenges a straightforward view, as their funding profiles do not follow the trend of nationalization commonly associated with polarization.</p><p>Future studies could explore other strategies contributing to political polarization such as the extent to which candidates concentrate on “culture war” issues and limit their media presence to partisan outlets. Exploring these strategies could further support our hypotheses about ideological extremists’ roles in political polarization or complicate them. For now, our findings present a nuanced picture: ideological extremists’ communication strategies are more aligned with those propagating ideological extremism, but their funding sources suggest that their role could be more limited than commonly assumed.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=52926c4fd2c8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The blind leading the blind: Transatlantic differences in how polls treat race]]></title>
            <link>https://tufts-pol.medium.com/the-blind-leading-the-blind-transatlantic-differences-in-how-polls-treat-race-e2df7112951c?source=rss-e4754e5edb87------2</link>
            <guid isPermaLink="false">https://medium.com/p/e2df7112951c</guid>
            <category><![CDATA[polling]]></category>
            <category><![CDATA[race]]></category>
            <category><![CDATA[europe]]></category>
            <category><![CDATA[public-opinion]]></category>
            <dc:creator><![CDATA[Tufts Public Opinion Lab]]></dc:creator>
            <pubDate>Thu, 19 Feb 2026 14:26:00 GMT</pubDate>
            <atom:updated>2026-02-19T14:26:00.147Z</atom:updated>
            <content:encoded><![CDATA[<p><em>by Lucy Morisse-Corsetti (class of 2027)</em></p><p>Claiming that one is “colorblind,” that they do “not see race,” is commonly used to hide patterns of racial exclusion, which <a href="https://journals.sagepub.com/doi/10.1177/0020715206065784">leads to structural oppression of ethnic minorities</a>. However, it is par for the course in European data science, with the EU’s General Data Protection Regulation (GDPR) effectively <a href="https://commission.europa.eu/system/files/2021-09/data_collection_in_the_field_of_ethnicity.pdf">preventing the collection of race-based data</a>. Public opinion polling is a tangible example of “colorblindness” leading to harm. When a variable such as race, which has proven incredibly informative for societal formations and political motivations, is excluded from considerations, political science researchers fail at their job. We exist in a world that is still broadly racialized, and therefore, when no one is ensuring that voices of color are included, they are often shut out entirely.</p><p>While this piece makes the assumption that readers are familiar with the basics of public opinion polling, <a href="https://www.pewresearch.org/course/public-opinion-polling-basics/#how-does-polling-work">Pew Research Center</a> and <a href="https://media.gallup.com/muslimwestfacts/pdf/pollingandhowtouseitr1dreveng.pdf?">Gallup</a> have overviews for those who want a quick refresher. In the United States, race is <a href="https://www.pewresearch.org/politics/2024/04/09/partisanship-by-race-ethnicity-and-education/?">almost always one of the variables</a> that is measured in public opinion polls and is an incredibly predictive form of <a href="https://isps.yale.edu/news/blog/2023/07/racial-identity-explains-presidential-vote-choices-more-than-geography?">stratification in US political polling</a>, both in local and national level races, across the entire country. Furthermore, the <a href="https://www.census.gov/programs-surveys/decennial-census/decade/2020/2020-census-results.html">US census</a> requires the collection of race and ethnicity data in a standardized format, which provides pollsters with known population proportions to compare their sample data to. Groups can be adjusted using weighting, in order to make sure that the data being produced is a more accurate representation of the population’s beliefs.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/434/1*R7T65XJRumV5eN6KzKSi7A.png" /><figcaption>An example of how American pollsters analyze opinion across racial and ethnic groups.</figcaption></figure><p>As <a href="https://belonging.berkeley.edu/europeans-not-talking-about-race">compared to Europe</a>, racial politics are more central to <a href="https://www.pewresearch.org/politics/2024/04/09/partisanship-by-race-ethnicity-and-education/?">US political debates and party coalitions</a>. Race-conscious polling is therefore critical to capturing key facets of political decision making. Furthermore, as the United States has a long and contentious history of shutting out voices of color from political conversations, inclusion of minority voices is critical.</p><p>On the other hand, European public opinion polling does not take race into consideration, despite being the <a href="https://ourworldindata.org/grapher/european-overseas-colonies-and-their-colonizers">world’s colonizers for centuries</a>. Furthermore, while ethnicity historically has been a <a href="https://journals.sagepub.com/doi/10.1177/1354068819864089">more relevant variable</a> (Romani, Jewish, etc.), it is also not considered. The <a href="https://www.gesis.org/en/eurobarometer-data-service/data-and-documentation/standard-special-eb/weighting-overview?">Eurostat weights</a>, which are post-stratification weights commonly used in the EU, do not include race, ethnicity, or religious affiliation. This lack of standardized data raises some key questions for policymakers. How can they verify minority groups are (or aren’t) receiving equal opportunity without data on the ethnic and religious background of their own citizens? Proving that there is systematic discrimination, not just an interpersonal issue, is far easier when there is broader data to back up claims. Furthermore, racial minorities in the US are <a href="https://www.pewresearch.org/decoded/2023/03/07/nonresponse-rates-on-open-ended-survey-questions-vary-by-demographic-group-other-factors/">less likely to respond to polls</a>, and if that holds true in other countries, then under-representation of minority groups would be unnoticed by researchers.</p><p>Many countries are<a href="https://commission.europa.eu/system/files/2021-09/data_collection_in_the_field_of_ethnicity.pdf"> legally restricted </a>from collecting data on race and ethnicity, a legacy of how such data was used during the Holocaust. Some argue that what began as an inclusive measure is in fact closer to what we understand today as “colorblindness” — choosing to ignore racial formations that exist in society, often leading to <a href="https://journals.sagepub.com/doi/10.1177/0020715206065784">marginalization and further oppression</a>. Racialization is still a major problem in the EU, with over half of the population believing there is <a href="https://commission.europa.eu/news-and-media/news/tackling-racial-discrimination-eu-2025-03-21_en">widespread racial discrimination</a> in their countries, most commonly occurring when looking for housing, a job, while at work or educational settings. That being said, in the context of the<a href="https://www.dw.com/en/far-right-populists-eu-germany-afd/a-72888389"> rise of far right parties</a> (such as AfD in Germany and PVV in the Netherlands), it is important to consider that data privacy is a serious concern, particularly for those who are part of minority groups which tend to have a high number of immigrants. This is a complicated tension that has come into view particularly in the last ten years, and will continue to evolve and take shape.</p><p>This debate is an <a href="https://commission.europa.eu/system/files/2021-09/data_collection_in_the_field_of_ethnicity.pdf">ongoing point of contention</a> in the EU, with “Some argue[ing] that this data collection essentialises ethnic groups or contributes to race discrimination” and others pointing out that “migration, language, education level and poverty data are not effective proxies for measuring discrimination based on racial and ethnic origin.” Privacy concerns are a commonly cited reason to prevent the collection of race based data. US institutions have <a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC4631034/?">Institutional Review Boards (IRBs)</a>, which are research ethics committees that work to protect participants in any type of human research, from public opinion work to drug trials. This includes the implementation of privacy protections, including the scrubbing of distinctly identifying information (such as name or street address) from public opinion data sets. Similar systems, called <a href="https://eurecnet.eu/recs-in-europe/?">Research Ethics Committees (RECs)</a>, are used in the EU, but they focus primarily on medical research. By expanding this kind of framework and making sure that it is required for public opinion work, a significant number of the privacy concerns that the GDPR is trying to address could be eliminated. While ignoring the existence of a very real and impactful variable is not a solution, we can recognize that the EU may have some legitimate concerns that can be addressed in other, more efficient ways.</p><p>In order to examine this issue more closely, it is helpful to narrow our analysis down to a specific case study — the Netherlands. While on a broader EU scale, the GDPR prevents the collection of race and ethnicity data, the Dutch statistical service, known as CBS, uses <a href="https://www.cbs.nl/en-gb/news/2022/07/cbs-introducing-new-population-classification-by-origin?">two different “non-racial” components</a> to classify minority individuals: birth location (if you/your parents were born in the Netherlands) and country of origin, both of which can be seen in the figures below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/1*YtQ3TyjP_Nk6ecXESH3gnA.png" /><figcaption>How the Dutch statistical service CBS classifies individuals.</figcaption></figure><p>Country of origin is a more complex method of classification, with four categories or “levels”. Level one distinguishes three categories: The Netherlands, Europe (excluding the Netherlands) and outside Europe. Level two breaks down “outside Europe” into two different categories: Traditional countries of origin, which includes Turkey, Morocco, Suriname, Indonesia, and the Dutch Caribbean, and others outside of Europe. Level three breaks down others outside of Europe further into Other — Africa, Other — Asia, and Other — Americas and Oceania. Finally, level four recognizes individual countries. Each of these categories require more strict scrutiny in order to collect that data. Researchers must prove that there is a <a href="https://www.cbs.nl/en-gb/news/2022/07/cbs-introducing-new-population-classification-by-origin?">specific reason</a> that a study needs any data above level one in order to collect it, which is often difficult if there is a significant research gap or if the study is not specifically about how being in one of those categories affects public opinion. This manifests particularly when examining topics such as privilege, say in job applications). A white American individual is classified as a migrant from outside of Europe in a dataset, but clearly retains all of the social and political privileges that whiteness provides. On the other hand, a person of color who has lived in the Netherlands for six generations is categorized as a Dutch Native from the Netherlands, but still would face societal and political discrimination due to the color of their skin.</p><p>By not allowing for self-identified racial categorization, these categories are at best unrepresentative descriptions and therefore build failure into public opinion studies. The systemic denial of access to this kind of data prevents researchers from filling gaps in political science literature, and from representing realities that minority populations face every day. Colorblind approaches to research in a society that very much “sees” color is simply misrepresentative.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/536/1*D5GkwYQkRD8JqdPGbb2ePA.png" /></figure><p>There are some significant questions that this post is unable to tackle. In the EU as a whole, how might the GDPR be reformed in order to both protect privacy while simultaneously preventing discrimination? Furthermore, as the Netherlands (among other EU countries) sees <a href="https://www.dutchnews.nl/2025/10/the-big-election-issues-immigration-racism-and-discrimination/">immigration</a> becoming a more central topic in the organization of political parties, how might race become a more predictive factor? When might it be more valuable to collect racial data, and simultaneously, why could it become more problematic? Still, it is clear that a colorbind approach to public opinion polling is ineffective, because society itself is not colorblind — race is still one of the most impactful variables in individuals lives. It may be <a href="https://research.ceu.edu/en/publications/race-in-mainland-european-legal-analysis-towards-a-european-criti/">a social construct</a>, but it is one that has real world impacts, and therefore should be taken into account when attempting to represent the real world. Researchers have an obligation to present the realities of society, and all of its diversity. Until the EU is able to effectively restructure its approach to race, public opinion research will continue to be ineffective at representing all of its citizens.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e2df7112951c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Trump’s support increased the most in counties with the fewest transplants]]></title>
            <link>https://tufts-pol.medium.com/trumps-support-increased-the-most-in-counties-with-the-fewest-transplants-2cd319da0199?source=rss-e4754e5edb87------2</link>
            <guid isPermaLink="false">https://medium.com/p/2cd319da0199</guid>
            <category><![CDATA[public-opinion]]></category>
            <category><![CDATA[2024-elections]]></category>
            <category><![CDATA[voting]]></category>
            <dc:creator><![CDATA[Tufts Public Opinion Lab]]></dc:creator>
            <pubDate>Wed, 11 Feb 2026 19:24:13 GMT</pubDate>
            <atom:updated>2026-02-11T19:24:13.081Z</atom:updated>
            <content:encoded><![CDATA[<p><em>by Toby Winick (class of 2026)</em></p><p>According to U.S. Census data, <a href="https://www.census.gov/topics/population/migration/guidance/acs-1yr.html">over 11%</a> of the U.S. population moved to a new residence in 2024, and about a quarter of those movers went to a new state. Inflation, cost of living concerns, and an uncertain housing market are all driving many Americans to uproot their lives and begin elsewhere, something that’s easier than ever to do with the proliferation of remote work opportunities. These “push factors” are all issues that elected officials — many of whom represent communities with a greatly different demographic makeup than they did a few elections ago — <a href="https://www.ipsos.com/en-us/cost-living-concerns-will-loom-large-midterms">are expected to have the answers to</a>.</p><p>Americans continue to move to the Sun Belt in the greatest numbers. Texas, Florida, Georgia, Colorado, Arizona, and the Carolinas <a href="https://www.newsweek.com/colorado-only-blue-state-10-places-americans-moving-2036075">boasted</a> some of the strongest net migration numbers in 2023. Meanwhile, California, New York, Massachusetts, New Jersey, Louisiana, and New Hampshire were all amongst the states losing the most residents. The states that movers are choosing are largely Republican, with all but Colorado choosing Donald Trump in the 2024 Presidential Election.</p><p>To analyze the electoral impact that droves of new voters are making on these states, I turned to the U.S. Census’ geographic mobility measure for each county in the country. This allowed me to calculate the proportion of residents in each county that were not there 5 years ago, according to their 2019–2023 5-year estimates. I also used county-level elections data for the 2020 and 2024 presidential elections to calculate the change in party vote share during that time. I joined this data to county shapefiles with GIS and graphed the results.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/936/1*FSdOHtZtDDGp5Cgywt2Low.png" /></figure><p>This map displays these two factors together. Counties with a Republican shift above 6 percent are considered high values as the national popular vote shifted by that amount towards Donald Trump between 2020 and 2024. A significant value for share of new residents is one above the national county median of 5.7%.</p><p>Furthermore, any county that exhibited a shift towards Democrats at all is marked as significant, as these are the counties that reversed the national trend entirely. Although there is a large concentration again in Georgia and the Carolinas, many states in the West such as Colorado, Utah, Washington, and Oregon have concentrated areas with both many new residents and a leftward shift.</p><p>Compared to their Republican counterparts, these Democratic destination counties seem to be much more locally concentrated. There are a few solid dark blue “blocks” in Colorado, Kansas, and Georgia, which don’t have many light blue counties surrounding them — implying that Democratic shifts are primarily occurring in regions with high concentrations of movers. This suggests that Democratic vote shift is closely tied to an influx of new residents, while there isn’t a strong relationship between Republican vote shift and high-mobility counties.</p><p>To test the relationship between moving and voter turnout on a county level, I ran a local regression with these variables and every county, weighted (and sized on the graph) by county population.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/936/1*a3Kff5pjSNwH1mFviEefWQ.png" /></figure><p>Amongst counties with between 0–5% new residents, there is a very strong negative relationship between these variables: nearly all of the counties with the highest shift towards Trump fall in this category. However, after this point, the line is nearly horizontal, indicating that there is almost no relationship between vote share and new residents. Indeed, the line hovers around 6% on the right side of the graph, which was the national average shift.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/940/1*PcUlak48bRfAV-rGx7JeGA.png" /></figure><p>Looking at the counties that moved the most in either direction suggests that significant swings towards Trump were driven by his <a href="https://www.pewresearch.org/politics/2025/06/26/voting-patterns-in-the-2024-election/">significant strides with Hispanic/Latino voters</a>, and not by new movers. Regardless of size and location, all of the counties with the biggest Republican swings have a mover rate below the median, with many only around 2 or 3%. They also have a significant Hispanic population. Queens County, NY has the lowest Hispanic population of the 10 Republican-shifting counties, but Hispanic residents still make up <a href="https://datausa.io/profile/geo/queens-county-ny#demographics">over a quarter</a> (27.5%) of its population. Notable are Hidalgo County, TX (<a href="https://datausa.io/profile/geo/hidalgo-county-tx">92%</a> Hispanic), Passaic County, NJ (<a href="https://datausa.io/profile/geo/passaic-county-nj">39% Hispanic</a>), and Imperial County, CA (<a href="https://datausa.io/profile/geo/imperial-county-ca">86% Hispanic</a>). In every county except Queens, Hispanic is the largest ethnic group. For the counties in New York, New Jersey, and California, it is also possible that many Democratic voters chose to leave, joining the trend of net migration away from these states.</p><p>Conversely, nearly all of the counties with a large shift towards Democrats experienced a percentage of new residents above the national median of 5.7%. These counties are largely in destination states like Georgia, Texas, Colorado, and North Carolina. One possible reason is that many movers are Democrats. Demographics associated with an increase in mobility, such as being educated and younger, are also associated with Democratic identification. These qualities might allow voters to relocate somewhere for new work opportunities, or capitalize on remote work that isn’t available in trade professions with a less-educated workforce.</p><p>Despite this, Democrats still face a demographic hurdle. The areas that are getting the most new residents are largely Republican (and therefore would take very large population changes to shift overall partisanship), and the next electoral college realignment will likely add votes to Texas and Florida, for example, at the cost of Democratic strongholds in the Northeast. However, this analysis shows that at a local level, new movers can play a role in shaping their community’s politics, and this could extend to a state level.</p><p>Future research could explore the extent to which a community’s politics influence movers: the Sun Belt continues to attract Democratic migrants despite a regional Republican majority, though Democratic shifts occurred in historically liberal Colorado, Washington, Oregon, and even North Carolina, which is seen as a swing state (despite not electing a Democrat since 2008). The relationship between movers and electoral outcomes will continue to be an important one for lawmakers and redistricting.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2cd319da0199" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[“Queers for Palestine”: Ironic, or a representation of Queer Theory in action?]]></title>
            <link>https://tufts-pol.medium.com/queers-for-palestine-ironic-or-a-representation-of-queer-theory-in-action-152b8a1f0c7c?source=rss-e4754e5edb87------2</link>
            <guid isPermaLink="false">https://medium.com/p/152b8a1f0c7c</guid>
            <category><![CDATA[queer-theory]]></category>
            <category><![CDATA[lgbtq]]></category>
            <category><![CDATA[public-opinion]]></category>
            <category><![CDATA[palestine]]></category>
            <category><![CDATA[gaza]]></category>
            <dc:creator><![CDATA[Tufts Public Opinion Lab]]></dc:creator>
            <pubDate>Mon, 09 Feb 2026 14:13:46 GMT</pubDate>
            <atom:updated>2026-02-09T14:13:46.742Z</atom:updated>
            <content:encoded><![CDATA[<p><em>by Seona Maskara (class of 2026)</em></p><p>Since the beginning of the current Israel-Palestine conflict, groups organizing under the titles “Queers for Palestine” have faced <a href="https://www.queermajority.com/essays-all/queers-for-palestine-and-the-death-of-irony">criticism for being ironic</a>, as being queer is illegal in Palestine. However, queer groups, such as <a href="https://quitpalestine.org/">Queers Undermining Israeli Terrorism</a> (QUIT) and the <a href="https://www.instagram.com/p/C8yNQhMNFzA/?img_index=1">Dyke March</a> have highlighted that a part of being queer is fighting for liberation for all people, including those in Palestine, who live without self-determination. <a href="https://alqaws.org/news/Reflecting-on-Queerness-in-Times-of-Genocide?category_id=0">Queer people in Palestine</a> have added to the discourse, arguing that it would be impossible to fight for equality for queer people in Palestine without equality for Palestinians in general, <a href="https://www.hrw.org/news/2023/12/05/does-israels-treatment-palestinians-rise-level-apartheid">which they currently do not have</a>.</p><p>The idea that queerness in itself is a fight for liberation comes from <a href="https://www.lgbtqnation.com/2021/10/queer-theory-101-matters/">queer theory</a>. Queer theorists argue that an essential part of being queer is understanding systems of oppression such as homophobia, hetero-normativism, religious systems, and the patriarchy that prevent one from living life as a queer person. This lens is then expanded to be critical of systems of oppression in general by questioning why societal constructs exist and who is left out of them, as queer people themselves are left out of society due to the systems mentioned above. Queer theorists critique systems, including but not limited to, policing and incarceration, religion, imperialism, and colonialism;<a href="https://d.lib.msu.edu/utafiti/5"> systems under which the current state of Israel was formed</a>.</p><p>Under queer theory, supporting Palestine would not be hypocritical at all, but essential to one’s queer identity; and under queer theory, to be queer would be to question more systems, and therefore hold radical beliefs. This leads to the question: do queer people hold more ‘radical’ viewpoints than their liberal peers?</p><p>Overwhelmingly, the answer is yes. Queer Americans are more ‘radical’ across the board than straight liberal Americans in the 2024 Cooperative Election Study. I paid specific attention to issues that tend to divide liberals, such as support for Palestine, defunding the police, and systemic racism, but also to issues that tend to have high, but somewhat varying support, among liberals: immigration and Medicaid. Agreeing with these issues (or disagreeing with limits for Medicaid) is taken to be more “radically” liberal for the purposes of this post.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/936/1*0lEQ5tSYOkeKWphEpaz34w.png" /></figure><p>The graph visualizes support for Palestine and support to defund the police among Americans who identify as liberal. However, the pattern persisted among a range of issues; 66% of queer liberals strongly agreed that black people faced systemic barriers, compared to 51% of their straight/cisgender counterparts, 91% vs 82% agreed that undocumented immigrants should be granted legal status after three years, and 25% vs 36% agreed that a job should be required for Medicaid (something leftists argue against, but liberals are divided on).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/936/1*W7VJ5WYtbzEpxJ2QPV2sHQ.png" /><figcaption><em>The ‘other’ category in the graphs refers to those who said their sexuality was ‘other’, but excludes those who declined to answer the question; they are included in the ‘queer’ umbrella, seen to represent the “+” in LGBTQ+</em></figcaption></figure><p>​Across categories, gay men were the least radical, while transgender and genderqueer individuals were the most; this potentially points towards the idea that the more oppressed an identity is (<a href="https://www.acponline.org/advocacy/state-health-policy/attacks-on-gender-affirming-and-transgender-health-care">as currently transgender people face widespread attacks</a>), the more likely they are to use queer theory to examine society.</p><p>While the data suggests that queerness is an indicator of radical beliefs, the question remains if queerness is the key indicator in those beliefs. In the graph below, each point can be interpreted as, all other categories equal, the effect of that variable on likelihood to agree with the given prompt. For example, all other things equal, high religiosity makes respondents less likely to agree to provide military aid for Gaza.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/936/1*kaG5KZUd-i9oflauDsWNgg.png" /></figure><p>So, for both categories, being queer (as opposed to not being queer) indicates a 7% higher likelihood to agree with the policy proposals, all other factors controlled. The ‘race’ category uses ‘white’ (the highest proportion of respondents) as a control, indicating a slight decrease in support for military aid to Gaza for non-white respondents. The ‘religion’ category, similarly, uses ‘Protestant’ as a control, indicating that, in general, not being protestant has a slight decrease in support for both proposals. However, neither are incredible significant factors, especially when compared to queerness.</p><p>When controlling for factors including religion, age, race, and education across liberals, queerness was the most determining factor across the board, even more so than Judaism in determining support (or lack of) for Gaza, arguing against stereotypes of who and who does not support Palestine. The effect of queerness on supporting radical issues that divide liberals is not to be ignored; queerness has the greatest effect on predicting agreement to defund the police and support for Palestine.</p><p>Thus, maybe “Queers for Palestine” isn’t ironic. Examining systems of oppression and inequality is essential to queer theory, and based on the results of the poll, queer people are more likely to act in solidarity with other oppressed groups than their straight and heterosexual peers.Queerness is a uniquely strong predictor of radical beliefs; it seems that “Queers for Palestine” is indeed not ironic, but a valid representation of queer theory.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=152b8a1f0c7c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The X-perience Factor: What polls miss about Black experience with college and support for…]]></title>
            <link>https://tufts-pol.medium.com/the-x-perience-factor-what-polls-miss-about-black-experience-with-college-and-support-for-24bbee658091?source=rss-e4754e5edb87------2</link>
            <guid isPermaLink="false">https://medium.com/p/24bbee658091</guid>
            <category><![CDATA[college]]></category>
            <category><![CDATA[public-opinion]]></category>
            <category><![CDATA[black-people]]></category>
            <category><![CDATA[affirmative-action]]></category>
            <category><![CDATA[supreme-court]]></category>
            <dc:creator><![CDATA[Tufts Public Opinion Lab]]></dc:creator>
            <pubDate>Mon, 22 Dec 2025 16:54:44 GMT</pubDate>
            <atom:updated>2025-12-22T18:54:38.554Z</atom:updated>
            <content:encoded><![CDATA[<h3><strong>The X-perience Factor: What polls miss about Black experience with college and support for Affirmative Action</strong></h3><p><em>by Miles Kendrick (class of 2026) and Rachel Kuhn (class of 2026)</em></p><p>Polls can capture opinions, but they rarely capture experiences. And when it comes to issues like affirmative action, experience turns out to be the difference between Black Americans supporting the Supreme Court’s ruling and opposing it.</p><p>Colleges and universities began to consider race as a factor in admissions in the late 1960s. In <a href="https://edtrust.org/blog/a-brief-history-of-affirmative-action-and-the-assault-on-race-conscious-admissions/">response</a> to the continued systemic discrimination of racial and ethnic minorities in employment and education, affirmative action was proposed as a means of promoting nondiscrimination and mitigating entrenched inequalities in the education system. Since its first implementation, affirmative action has endured numerous unsuccessful challenges. Then, in 2022, the Supreme Court heard arguments in <em>Students for Fair Admissions vs. President and Fellows of Harvard College</em> and concluded that Harvard’s race-conscious admissions processes violated Title VI of the Civil Rights Act of 1964 by discriminating against Asian American and white applicants. In one fell swoop, affirmative action was ended, overturning over fifty years of judicial precedent.</p><p>Despite the disruptive nature of the decision and <a href="https://www.pewresearch.org/short-reads/2025/09/03/favorable-views-of-supreme-court-remain-near-historic-low/">continued unpopularity</a> of the Supreme Court, Americans by and large support the decision to overturn affirmative action; two in three <a href="https://news.gallup.com/poll/548528/post-affirmative-action-views-admissions-differ-race.aspx">say</a> the decision is “mostly a good thing.” Even more interestingly, although Black Americans are among the most <a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC11254878/">affected</a> by affirmative action, a slight majority agrees that the decision was a good one. Perhaps support for the decision can be explained by gender, class, or political ideology. Or maybe it’s worth our time to delve deeper into public opinion around this issue, looking at how the lived experiences of Black Americans affect their view on the Court’s decision.</p><p>The 2024 survey from the <a href="https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/X11EP6">Cooperative Election Study</a> asked Americans if they support or oppose the Court’s decision on affirmative action. We looked into how Black Americans, who may have benefited from affirmative action in the past, know someone who has, or remember a day in which educational or job discrimination was more salient, might view the Court’s decision. As it turns out, Black Americans with at least some college experience are more likely to disapprove of the decision than Black Americans without.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/936/1*wW3EWMKOyrHLRnTYl-yWvA.png" /><figcaption><em>Source: 2024 Cooperative Election Study national poll. Graph created by Miles Kendrick.</em></figcaption></figure><p>To begin, we estimated a regression model where we predicted whether Black Americans opposed versus supported the Supreme Court’s decision to overturn affirmative action. In the model we included a set of relevant predictors, namely education (split into those with and without college experience), age group (split into under 40 and 40+ groups), gender, family income, political identity, and religious importance. The x-axis reflects the effect of the y-axis variables on likelihood to support affirmative action; these are percentage point increases or decreases on the issue (i.e. an increase in .02 is equivalent to a 2 percentage point increase in support for affirmative action). To simplify an otherwise confusing measurement, opposition to the Court’s decision will be referred to as support of affirmative action.</p><p>This graph provides many key insights, some more surprising than others. Partisanship, for example, has an expected effect: Black Republicans are less likely to support affirmative action than Black Democrats. The relationship between age group and affirmative action, on the other hand, is more surprising: Black Americans aged 40 and older are 6 percent more likely to support affirmative action compared with those under 40. This phenomenon has been previously addressed in a 2024 <a href="http://google.com/search?q=black+opinion+on+affirmative+action&amp;rlz=1C5CHFA_enUS937US937&amp;oq=black&amp;gs_lcrp=EgZjaHJvbWUqBggBECMYJzIMCAAQRRg5GLEDGIAEMgYIARAjGCcyBggCEEUYOzIKCAMQLhixAxiABDIICAQQRRgnGDsyBggFEEUYPDIGCAYQRRg8MgYIBxBFGDzSAQgxODgzajBqNKgCALACAA&amp;sourceid=chrome&amp;ie=UTF-8">Gallup</a> poll, which attributed the divide to general life experience and generational perceptions of affirmative action.</p><p>Religiosity is also a strong predictor. The more important religion is to your life, the less likely you are to support affirmative action. Perhaps policies that supposedly treat people differently based on race can feel incompatible with the <a href="https://www.insightforliving.ca/read/articles/what-does-bible-teach-about-all-people-being-equal">idea</a> that everyone should be judged equally before God. Less discussed, however, is the dramatic positive correlation between education and support for affirmative action. Black people who experienced college are 19 percentage points more supportive of affirmative action than those who did not.</p><p>Perhaps this is unsurprising. We know, for example, that higher levels of education are highly <a href="https://www.pewresearch.org/politics/2016/04/26/a-wider-ideological-gap-between-more-and-less-educated-adults/">correlated</a> with more liberal ideologies in all demographics. But what if there was something specific about affirmative action that triggered a more dramatic discrepancy in support between college-educated and non-college-educated Black people than other liberal policies? To test this, we charted the differences in support of gun control, carbon dioxide emissions, and abortion, as well as the differences in party identification.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/936/1*pMkdpKr_rojEuiSwoisz9Q.png" /><figcaption><em>Source: 2024 Cooperative Election Study national poll. Graph created by Rachel Kuhn.</em></figcaption></figure><p>On the Supreme Court’s affirmative action decision, the divide between college experience versus no college experience is dramatic. Black Americans with some college experience disagree with the Court’s ruling 21 percentage points more than those without. This difference is larger than the gaps on gun control or emissions regulations — issues that typically produce some of the most predictable <a href="https://www.pewresearch.org/science/2015/07/01/americans-politics-and-science-issues/">education-based </a><a href="https://www.pewresearch.org/internet/2024/06/12/politics-across-platforms-acknowledgments/">splits in public opinion</a>. Around 55 percent of Black Americans with college experience disagree with the Court’s decision on affirmative action, whereas around 35 percent of Black Americans without college experience disagree with the decision. The magnitude of the difference in opinion between these groups suggests that perhaps positive experience with affirmative action on campus plays a central role in shaping how Black Americans feel about the Court’s ruling.</p><p>Our analysis supports the idea that experience in education does more than structure political attitudes in general; it appears to sharpen the perceived stakes of ending affirmative action specifically. Black Americans who have gone through some version of the admissions process and possibly experienced the benefits of a diverse student body firsthand disproportionately support affirmative action for this very reason.</p><p>What’s more, our findings reaffirm the importance of lived experience, something too often flattened in public opinion research. Pollsters and analysts routinely rely on demographic categories like race, age, religion, or ideology to explain complex political behavior. While these markers can reveal broad patterns, they cannot substitute for the textured and unique realities that shape how people think, decide, and act. Our analysis suggests a need for greater nuance in both the research process and the narratives built from it. We need researchers, journalists, and scholars to intentionally integrate lived experience into survey design, data interpretation, and writing. Doing so will deepen our understanding of why people support or oppose particular policies and how they make sense of political choices. Ultimately, we challenge those in the field, as well as our readers, to look beyond the data and toward the people behind it.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=24bbee658091" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[LGB Republicans look a lot like other Republicans and less like other LGB Americans]]></title>
            <link>https://tufts-pol.medium.com/lgb-republicans-look-a-lot-like-other-republicans-and-less-like-other-lgb-americans-d9f6b6939596?source=rss-e4754e5edb87------2</link>
            <guid isPermaLink="false">https://medium.com/p/d9f6b6939596</guid>
            <category><![CDATA[lgbtq]]></category>
            <category><![CDATA[political-parties]]></category>
            <category><![CDATA[republican-party]]></category>
            <category><![CDATA[sexual-identity]]></category>
            <category><![CDATA[public-opinion]]></category>
            <dc:creator><![CDATA[Tufts Public Opinion Lab]]></dc:creator>
            <pubDate>Mon, 27 Oct 2025 16:38:33 GMT</pubDate>
            <atom:updated>2025-10-27T16:38:33.200Z</atom:updated>
            <content:encoded><![CDATA[<p><em>by Toby Winick (class of ‘2026)</em></p><p>With control over the Supreme Court and a recent federal government trifecta, conservatives have taken the opportunity to deliver wins to their base on various “culture war” issues. The Supreme Court in <a href="https://www.npr.org/2022/06/24/1102305878/supreme-court-abortion-roe-v-wade-decision-overturn">Dobbs vs. Jackson</a> overturned precedent that protected abortion and subsequently issued a <a href="https://www.npr.org/2023/06/29/1181138066/affirmative-action-supreme-court-decision">ruling that effectively ended</a> affirmative action in college admissions. Following this, President Donald Trump signed executive orders <a href="https://www.federalregister.gov/documents/2025/02/03/2025-02178/prioritizing-military-excellence-and-readiness">banning transgender people</a> from military service and <a href="https://www.skadden.com/insights/publications/2025/02/the-informed-board/dei-under-siege">targeting Diversity, Equity, and Inclusion programs</a> in public and private sectors.</p><p>Despite this conservative stance on many social issues, the Trump administration has received <a href="https://www.nytimes.com/2025/08/26/style/gay-men-trump-administration-republicans.html">attention</a> for the significant number of gay men in its ranks. Secretary of the Treasury Scott Bessent, Kennedy Center President Richard Grenell, and Under Secretary of State Jacob Helberg are some of the highest-ranking gay men in the federal bureaucracy. The New York Times describes the administration’s “A-Gays” as “overwhelmingly white… not the type to be telling anyone their pronouns… and they aren’t the least bit offended that the leader of their party continues to stoke a moral panic about transgender people.” These gay men say that President Trump, unlike many Republican politicians, has a track record of being pro-gay.</p><p>However, they also find themselves at the intersection of contradicting cultural identities. The vast majority of <a href="https://news.gallup.com/poll/656708/lgbtq-identification-rises.aspx">Americans who identify as LGBT are not Republicans</a>, and only <a href="https://news.gallup.com/poll/646202/sex-relations-marriage-suppor">46% of Republicans support same-sex marriage (compared to the national average of 69%)</a>. Moreover, even if President Trump supports same-sex marriage, he has come under fire for <a href="https://www.nytimes.com/2025/06/18/well/lgbtq-988-suicide-prevention.html">cutting </a>funding for <a href="https://www.nytimes.com/2025/06/18/well/lgbtq-988-suicide-prevention.html">LGBTQ suicide prevention services</a>, ending a <a href="https://www.nytimes.com/2025/07/23/health/pepfar-shutdown.html">global AIDS relief</a> program, and cracking down on transgender <a href="https://www.whitehouse.gov/presidential-actions/2025/02/keeping-men-out-of-womens-sports/">athletes</a> and soldiers, all of which impact the broader LGBT community.</p><p>Using data from the 2024 Cooperative Election Study (CES), I analyzed the 5,799 LGB respondents to see if LGB Republicans looked like their peers. Taking inspiration from the Williams Institute’s <a href="https://williamsinstitute.law.ucla.edu/publications/lgb-party-affiliation/">brief</a>, I also analyzed differences in identity and sense of community between Democrats, Republicans, and Independents using data from <em>Generations</em>, a 2023 study of more than 1,500 lesbians, gay men, and bisexuals.</p><p>Almost everyone mentioned in the <a href="https://www.nytimes.com/2025/08/26/style/gay-men-trump-administration-republicans.html">New York Times article</a> was a white, gay man. I first wanted to see if this demographic trend held for LGB Republicans as a whole. The first graph shows the party identification for each race category in the sample.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/936/1*4nmRwI8G0uyv7PvpamC1Jg.png" /></figure><p>Breaking down the sample by race and gender shows that White men dominate the Republican LGB sample, making up almost half of LGB Republicans, nearly twice as much as White women. Republicans also have a lower proportion of respondents for every other group than Democrats, except for “Other” men, which includes Asian, Middle Eastern, and Native American. Still, these groups make up a minority of both parties, as they also make up a minority of voters. Independents contain the highest proportions for most minority groups, marking an interesting group of voters to target as Democrats have seen a <a href="https://www.pewresearch.org/politics/2025/06/26/voting-patterns-in-the-2024-election/">decline in support</a> amongst Black and Latino voters.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/936/1*aiGBX32Z8DSk4Hy5ign2IQ.png" /></figure><p>The next graph breaks down the sample based on each LGB individual’s specific sexual identity. Although the New York Times’ article focused mostly on gay men, bisexual men actually identify with the Republican party at a significantly higher rate (and even lower with Democrats). Bisexual women and gay men are the two largest categories, respectively, and offer similar profiles for party ID. Lesbians appear to be the most loyal to the Democratic Party, but are the smallest group aside from Bisexual non-binary/Other. Amongst all bisexual identities, there is a larger number of respondents outside the two major parties. It is worth noting that this trend persists despite the fact that bisexual women significantly outnumber bisexual men.</p><p>Overall, the CES data shows that the profile of LGB Republicans contains proportionally more White and male respondents especially bisexual males.</p><p>Although the data from the CES provides a large sample of LGB Americans to view demographically, <em>Generations</em> offers a more complete picture of how they view their sexual identity.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/936/1*Dqt5lHb8OZHahUD6_SMwNA.png" /></figure><p>The first question reflects the respondent’s internal image of their sexuality, while the second one reflects the role they feel it has in their daily lives. Interestingly, while 54.5% of LGB Republicans stated that their sexual orientation is central to their identity, this number falls by about 10% when asked if it is an important part of their life. It is possible that these respondents may not feel that they are as consciously representing their identity in their day to day lives. Democratic respondents exhibit the opposite effect, suggesting that they feel it affects their lives to an increased extent.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/936/1*Vrz4-hevETBhvH_T_GAhUQ.png" /></figure><p>When examining community-based questions, Republican respondents identify with the LGBT community significantly less and have a lower opinion of it. Considering the overall partisan lean of the LGBT community, it is not surprising that Republicans may feel excluded and not as proud overall. The last question touches on the concept of <a href="https://www.pewresearch.org/short-reads/2019/07/11/linked-fate-connectedness-americans/">linked fate</a>, the idea of a group consciousness creating political cohesion amongst voters of a certain identity. Democrats are almost twice as likely to resonate with this concept of linked fate, which may be because they comprise the largest voting bloc of LGBT Americans, allowing them to exert greater influence over the community’s culture.</p><p>This analysis also raises a “chicken or the egg” question about LGB Republicans. Is it the case that LGB Republicans identify as such because they do not feel as connected to their identity or to the LGBT community, or are LGB people who have these opinions more likely to become Republicans? Although the <em>Generations</em> survey does not contain questions about political identity, Republicans have both a lower mean (15.7) and median (14.5) response to the question “At what age did you first realize you were LGB?” when compared to Democrats (16.2 and 15). Perhaps these Republicans are more likely to experience political socialization after realizing they are LGB, which shapes their self perception.</p><p>It is also possible that in conservative environments, these Republicans face more external pressure to conform, making them more aware of their difference from their peers. These same conservative environments could influence political socialization while mitigating connection with the LGBT community. Future research could certainly examine the relationship between these two identities in terms of socialization.</p><p>Overall, the data reflect the interview responses gathered in the Times article, that queer Republicans feel less connected to their identity and to the LGBT community than Democrats or Independents. Moreover, their profile mirrors Republicans nationally, having the largest proportion of White and male respondents.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d9f6b6939596" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[¿Qué Está Pasando? Generational Assimilation & Latino America’s Rightward Shift]]></title>
            <link>https://tufts-pol.medium.com/qu%C3%A9-est%C3%A1-pasando-generational-assimilation-latino-americas-rightward-shift-8b3b431ab6a6?source=rss-e4754e5edb87------2</link>
            <guid isPermaLink="false">https://medium.com/p/8b3b431ab6a6</guid>
            <category><![CDATA[public-opinion]]></category>
            <category><![CDATA[latinos]]></category>
            <category><![CDATA[immigration]]></category>
            <category><![CDATA[2024-elections]]></category>
            <dc:creator><![CDATA[Tufts Public Opinion Lab]]></dc:creator>
            <pubDate>Fri, 17 Oct 2025 13:12:03 GMT</pubDate>
            <atom:updated>2025-10-17T13:12:03.452Z</atom:updated>
            <content:encoded><![CDATA[<p><em>by Rolando Ortega (class of ‘26)</em></p><blockquote><em>“Also, Latino men, despite the utter disrespect shown by Trump and his promise to deport some of your mixed-status families,… voted in a 55% majority to make the deportations happen. Y’all voted with Steven Miller and David Duke and against your own sisters… so you own everything that happens to your mixed-status families, and to your wives, sisters, and abuelas.”</em></blockquote><p>Two days after the 2024 presidential election, <a href="https://youtu.be/GridByQB8qQ">MSNBC host Joy Reid sent the message above to Latinos on her former show</a>.</p><p>In the wake of Donald Trump’s 2024 presidential victory, political analysts quickly turned to examine how different demographic groups voted — especially Latino voters. Despite being thought of as a Democratic-leaning constituency due to Republicans’ and Trump’s hostile rhetoric on immigration, <a href="https://pewrsr.ch/3Tbflih">Trump won a historic number of Latino votes</a> en route to his election victory over then-Vice President Kamala Harris. This added to <a href="https://sites.tufts.edu/cooperativeelectionstudy/2025/08/19/trump-built-a-coalition-in-2024-that-took-advantage-of-cracks-in-the-democratic-partys-coalition/">the gains he made among Latinos in the 2020 election</a>, even while losing to Joe Biden.</p><p>How could a candidate like President Trump add so many Latino voters to his coalition? Some commentators, like Reid, proposed that this shift is <a href="https://www.cleveland.com/reckon/2024/11/how-a-new-generation-of-latino-men-are-redefining-machismo-through-politics.html">due to machismo, or a sense of hypermasculinity, in the Latino community</a>, while others felt that Latinos behaved similarly to voters in other demographic groups <a href="https://www.nbcnews.com/news/latino/trump-economy-latino-vote-2024-election-rcna178951">out of anxiety over inflation and rising costs under Biden</a>. However, an overarching factor influencing this political shift is assimilation.</p><p>Assimilation is <a href="https://escholarship.org/content/qt1kx005x6/qt1kx005x6.pdf">a process where two groups become more similar over time</a>. This often occurs when generations of minority or outside groups increasingly adopt the culture, values, and norms of a dominant culture. Multiple factors, like interethnic marriage, emphasis on speaking English, and upward socioeconomic mobility, shape how people assimilate. Immigrant communities throughout the United States have experienced this process. For example, while once facing discrimination by the homegrown white population in the 19th and 20th centuries, European immigrants like the Irish and Italians <a href="https://www.jstor.org/stable/pdf/29737719.pdf">eventually shed their foreign identities and became part of the “white American”</a> fold. Consequently, <a href="https://watermark02.silverchair.com/82-2-481.pdf?token=AQECAHi208BE49Ooan9kkhW_Ercy7Dm3ZL_9Cf3qfKAc485ysgAAA10wggNZBgkqhkiG9w0BBwagggNKMIIDRgIBADCCAz8GCSqGSIb3DQEHATAeBglghkgBZQMEAS4wEQQMhQ9ePmVoXCcs-_djAgEQgIIDEK552C9OFxkDYFGDh-u9XtxZ47JAyIj157ek5BfYEE3oS1CsixxjoqcFh_RAfPgHtQE-b0zsKzcs4xPoLtF4AjyruJMCO-pKbbJv80N-fkxvlq_66vOdnZAUCbIVk0WJiffTW7k3mA535zq04gW4ho5-ky6Mlt1Llq-A2V2hPaNPjrM-jfqVQNoozjzWO4nrZ7pD7seKIv6mLs2f3SQ0hctAycT3I7xFMRKx2DcBSyesMh0jf3TQPyrZgxUfXNolAyOc21F9NP5L1OiKtJjyqAgHr0ieZsC4h8CDQVlDljlLvRQR8uODIlYNA0Wjo71gS1_1W5vQhChgGznp0KFzSSaukR5ABGPTNUJ4pwsvFQQk0A2IPuU9hQ6fyWaQtKPuwkAV2h04ul6ccrGhuyqXQit35F4ib5Wy-z1c_kFISDuZiIxtMRmpwIn8IM7POWN3eYlbZ5InmHRgZq8qxWJzKnfBRzu6U7l9paSVlWgMTTpq-_exGFNKkeyjVl0vOxzOa6gCN20x3bkyAiUFE5GA1IYsS5Osxl-nrY0GxWpvtbTCUKIEelcuBRWFrhy8MecI2g-qzHmyFpb8lvSc7kTrkNlikVEC83cALgaUVQj0nH2oHLszUbkKpp5QorK9WHuRyoTV0x8Gr_tPh8bMXV8ju_8Trxsa97EuWs20fUuaS-ysdxbIVZPpDmR7GykqX2zCMGplDUTbftG6Ixq-7Y8oztjQFsj9gN14GEK7qfiY71ZLhfiJRPYzfENoj8hMf0aMd1SaFh7fPIsnqVhHM_wDzXQ2uR2CKxCKZKnobfDWBBSB6cmb9A_-jCjCofQ63dXYUChwD0HD8sXOCTl_vEIIldu_7k7rsyAUy6K5AiJFFtpj8R6B60k6QPTU-POGIk5ET3dWh-va3LFRb3VCfR8LLbO_1Q7715t0Ue5hY9NtNMMYMcztn4z_FkPUk9rW_W1OXEM63cQXUlsBei8sPy8nsa4RwSjSO2V7PHY8qQaEEZtb_LtVjohMBSxV7PsazTn1hhaMYSbDyjCWcO-1FGJO2II">later generations of these groups gave more political support to Republicans and were more hostile toward other newer immigrants</a>.</p><p>The same phenomenon has been unfolding with Latino voters: later generations, further removed from the immigrant experience, naturally become more assimilated to the dominant white American identity. Hence, more Latinos have more negative views of newer immigrants and identify less with the “Latino” label. This pulls them away from policies Democrats often use to mobilize Latino voters, like immigration reform.</p><p>To explore how assimilation is affecting Latino voters, I turned to the most recent Cooperative Election Study (CES) and American National Election Study (ANES) surveys conducted during the 2024 election. Both asked respondents a series of questions on their views of immigrants and immigration policy, along with partisanship and ethnic identity.</p><p>Indeed, Latinos in both surveys demonstrated cross-generational shifts in views on immigration and identity. The CES asks respondents for both their racial and ethnic identity and, separately, whether they identify as Hispanic or Latino. According to the data, Latino identity weakens across generations: 87% of first-generation Latinos identified as Hispanic or Latino when asked for their race or ethnicity, compared to just 61% in the fourth generation. Conversely, white self-identification rises from 7% to 24%. Although a majority of Latinos consistently identify themselves as such, a clear decline emerges between the immigrant generation and later generations, whose families have lived in the U.S. longer and have undergone greater assimilation as a result.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/788/1*kEPcVcyyfIaIm39_KDHBbw.png" /></figure><p>The data also show that later-generation Latinos are less likely to take the pro-immigration side on several issues than first-generation Latinos–those who are immigrants themselves. For example, 81% of first-generation Latinos supported giving legal status to undocumented migrants. This decreases from generation to generation, dropping by 16 percentage points in the fourth generation, where only 65% of Latinos share this view. Similarly, while 52% of first-generation Latinos oppose denying asylum to immigrants who crossed the U.S.-Mexico border illegally, only 41% of fourth-generation Latinos expressed the same view. This was consistent on other issues, such as support for permanent residence status for immigrant children brought to the U.S. by their parents, also referred to as Dreamers, and opposition to President Trump’s border wall.</p><p>Notably, these differences between generations are significantly greater now than they were in 2020. At least 60% of Latinos in each of the generations took the more pro-immigration stance on all of these issues in 2020–overwhelming majorities across the board–and the largest gap between generations on any particular issue was just 7 percentage points. Latinos in later generations are, as they spend more time in the U.S., clearly taking more conservative stances on immigration policies than those who are immigrants themselves or the children of immigrants. As these Latinos identify less with the immigrant experience of their ancestors and more with standard “American” culture, perhaps they begin to adopt views on immigration that are more in line with protecting that dominant culture.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/790/1*3yuWR8_lC4wWevJ5U8x6Kg.png" /></figure><p>Similar generational divides among Latinos also appeared in the ANES data, which, in particular, asked several questions about how respondents perceived immigrants. First-generation Latinos are much more sympathetic toward newly arrived and illegal immigrants, as a 52% majority believe that illegal immigrants do not affect the crime rate in the United States. In comparison, an overwhelming 88% say that it is somewhat or not likely that immigrants take away jobs from Americans.</p><p>This differs significantly from the views of fourth-generation Latinos. 72% believe that illegal immigrants increase the crime rate–a 25-percentage-point increase from the first generation–while only 63% believe that immigrants are somewhat or not likely to take away jobs from Americans, representing a 25-point decline in pro-immigration sentiment. Interestingly, fourth-generation Latinos have even more negative attitudes towards immigrants than the American population overall, as 62% of Americans believe illegal immigrants increase the crime rate and 69% believe that immigrants take away their jobs.</p><p>Despite being grouped in the same ethnic category, generational status creates divisions among Latinos. Those in later generations see newer immigrants as an economic and social threat, reflective of the views of other native-born Americans. Like previous groups in immigrant communities, Latinos in the United States who may be more assimilated into the white-dominant American culture and removed from the immigrant experience are more likely to view newcomers in a negative light.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/718/1*no2w-9w4SVfBn-5QMm4KGg.png" /></figure><p>This week marked the end of Hispanic Heritage Month, a celebration of the richness and diversity of the Latino community in the United States. Yet, as the refrain goes, Latinos are not a monolith. The data suggest that Latinos exhibit generational differences on immigration issues and identity, shifting away from positions that, in the eyes of the political world, would reflect their ethnic label. The key question, as with past groups of immigrants, is whether these differences will continue to grow in the future, making 2024’s shifts more durable. Answering this is crucial not only because the American electorate is gradually diversifying, but also because a 54% majority of Latinos belonged to the third or fourth generations in 2024. Examining the history of immigration and assimilation in the United States, while considering how it may apply to migrant groups today, may help to inform how the political world engages with the Latino community and shape the future of each party’s electoral</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8b3b431ab6a6" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Trump built support in 2024 by taking advantage of cracks in the Democratic Party’s coalition]]></title>
            <link>https://tufts-pol.medium.com/trump-built-a-coalition-in-2024-that-took-advantage-of-cracks-in-the-democratic-partys-coalition-4c46b56826d9?source=rss-e4754e5edb87------2</link>
            <guid isPermaLink="false">https://medium.com/p/4c46b56826d9</guid>
            <category><![CDATA[2024-elections]]></category>
            <category><![CDATA[voting]]></category>
            <category><![CDATA[donald-trump]]></category>
            <category><![CDATA[kamala-harris]]></category>
            <dc:creator><![CDATA[Tufts Public Opinion Lab]]></dc:creator>
            <pubDate>Tue, 19 Aug 2025 15:02:00 GMT</pubDate>
            <atom:updated>2025-08-19T15:39:02.299Z</atom:updated>
            <content:encoded><![CDATA[<p><em>by Caroline Soler and Brian Schaffner</em></p><p><em>This post first appeared on the </em><a href="https://sites.tufts.edu/cooperativeelectionstudy/2025/08/20/trump-built-a-coalition-in-2024-that-took-advantage-of-cracks-in-the-democratic-partys-coalition/"><em>Cooperative Election Study blog</em></a><em>.</em></p><p>In 2024, Donald Trump ran for president as the Republican nominee for the third consecutive cycle. However, this time something new happened: Trump won the popular vote. This result wasn’t the result of winning over any single demographic; instead, it represented a broad-based shift in how Americans across racial, religious, and generational lines cast their ballots.</p><p>We authored a<a href="https://tinyurl.com/ces2024report"> new report</a> on the election using data from the Cooperative Election Study <a href="https://tischcollege.tufts.edu/research-faculty/research-centers/cooperative-election-study">(CES)</a> which interviewed 60,000 Americans, including over 36,000 validated voters and more than 23,000 non-voters. Crucially, the survey validates whether respondents actually voted by matching them to comprehensive voter files through TargetSmart, eliminating the common problem of people claiming to vote when they didn’t.</p><p>The report demonstrates that the shifts toward Trump in 2024 were widespread and sometimes quite dramatic. Black voters have been moving away from Democrats since Obama’s presidency and Hispanic voters also shifted rightward, continuing movement that has been ongoing since Trump’s first term. Religious voters across denominations — Catholics, Jewish voters, Protestants, and those with no particular faith — moved toward the GOP. Millennials, long considered reliably Democratic, voted more Republican than they had in any previous election. And perhaps most striking is this: for the first time in decades, non-voters preferred the Republican candidate over the Democratic. if every American had voted, Kamala Harris would have lost by an even greater margin.</p><p>In this post, we highlight three of the most striking shifts documented in our report: the movement across racial lines, the religious realignment, and the generational upheaval.</p><p><strong>Racial shifts</strong></p><p>Trump’s victory wasn’t just built on mobilizing his traditional base — indeed, some of the biggest shifts occurred among racial and ethnic groups that tend to vote strongly Democratic. The most striking changes occurred among Black and Hispanic voters, two racial groups that had been reliably Democratic for years.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*LVDy8Tir2k2pEf4W" /></figure><p>While Obama won 97% of the vote among Black voters in 2008, Harris’s share of the Black vote dropped to just 84% in 2024. This erosion has been happening steadily since Obama’s first term, and continued into 2024. The Black electorate has <a href="https://docs.google.com/document/d/e/2PACX-1vR3cPQ81X2YD1Rm5cQgfudFWPgiJCpXx6K2njV91htbJ4nLCY5r6kASroo4icpj6j59YmNzjoddeWhT/pub#h.t84xflx95pfh">fractured across various demographic cleavages</a>. A massive gap has emerged between born-again and non-born-again Black voters. In 2008 and 2012, both groups voted Democratic at similar rates, but by 2024, a 13-point gap emerged (with born-again Black voters at 77% Democratic support compared to 90% for non-born-again Black voters). The gender divide is even more dramatic, with Black women supporting Harris at 90% while support among Black men dropped to 75% (in 2008, both groups were above 95%).</p><p><a href="https://docs.google.com/document/d/e/2PACX-1vR3cPQ81X2YD1Rm5cQgfudFWPgiJCpXx6K2njV91htbJ4nLCY5r6kASroo4icpj6j59YmNzjoddeWhT/pub#h.tsbzgxa2gw9e">Hispanic voters</a> experienced an even more concentrated shift toward Republicans. Their 10-point decline since 2016 accelerated dramatically between 2020 and 2024, falling from 66% to just 58% Democratic support in a single cycle. This rightward movement was driven primarily by urban Hispanic communities, whose Democratic support plummeted from 77% in 2016 to just 60% in 2024, while rural and suburban Hispanic voters remained relatively stable.</p><p>What makes these shifts particularly consequential is that they’ve happened as America has become more diverse. The electorate has grown less white — from 77% in 2016 to 73% in 2024 — with Hispanic voters now matching Black voters at a 10% share of the electorate each. This demographic change was <a href="https://www.americanprogress.org/article/the-changing-face-of-americas-electorate/">supposed to benefit Democrats</a>, but Trump’s inroads with these groups has nullified any such advantage.</p><p><strong>Religious realignment</strong></p><p>Trump’s gains weren’t limited to racial groups — he also made inroads across America’s religious landscape. From Catholics to Jewish voters to Protestants, and even those with no particular religious affiliation, nearly every religious community shifted toward Republicans. Only atheists and agnostics remained consistently Democratic.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*BimEJAF-jpHtcgvQ" /></figure><p><a href="https://docs.google.com/document/d/e/2PACX-1vR3cPQ81X2YD1Rm5cQgfudFWPgiJCpXx6K2njV91htbJ4nLCY5r6kASroo4icpj6j59YmNzjoddeWhT/pub#h.98759ed6oxd7">Jewish voters</a>, historically one of the most reliable Democratic constituencies, declined from 75% Democratic support in 2016 to 63% in 2024. This shift was driven primarily by non-Reform denominations, with Conservative Jewish voters (who made up almost 1/4 of all Jewish voters in 2024) dropping 17 points since 2008 and Orthodox Jews experiencing a dramatic 37-point plunge from 53% in 2016 to just 16% Democratic support in 2024. Catholics experienced a sharp acceleration away from Democrats, falling from 47% in 2020 to 40% in 2024 after remaining relatively stable for over a decade. Protestants continued their slow but steady decline, moving from 60% Republican in 2008 to 69% in 2024.</p><p><a href="https://docs.google.com/document/d/e/2PACX-1vR3cPQ81X2YD1Rm5cQgfudFWPgiJCpXx6K2njV91htbJ4nLCY5r6kASroo4icpj6j59YmNzjoddeWhT/pub#h.x8djdlamvc15">Born-again</a> support dropped from 36% in 2008 to just 24% in 2024, with the decline happening across all racial groups. Hispanic born-again voters plummeted from 48% Democratic support in 2020 to 33% in 2024, while Black born-again Christians experienced a 20-point decline since 2008, and white born-again voters became 9% more Republican. Meanwhile, non-born-again voters remained stable around 61–62%, creating an ever-widening gap between these groups.</p><p>Those identifying as having “nothing in particular” for their religion were the second-largest religious group for the first time in 2024, making up 18% of the electorate. Combined with the doubling of the size of the atheist/agnostic voting bloc since 2008, this secularization could have benefited Democrats. Instead, the “nothing in particular” group shifted 3 points right since 2020, meaning that America’s growing secular population didn’t translate into Democratic gains.</p><p><strong>The Millennial surprise</strong></p><p>In 2024, Millennials finally reached parity with Gen X in terms of their share of the electorate, with each group representing about 24% of voters. Many thought that as Millennials became a larger share of voters, their reliably Democratic preferences would <a href="https://www.ft.com/content/c361e372-769e-45cd-a063-f5c0a7767cf4">give the party an advantage</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*SuyifxACRwBU0MBu" /></figure><p>Instead, <a href="https://docs.google.com/document/d/e/2PACX-1vR3cPQ81X2YD1Rm5cQgfudFWPgiJCpXx6K2njV91htbJ4nLCY5r6kASroo4icpj6j59YmNzjoddeWhT/pub#h.4b8b5eyzl9xf">Millennials voted more Republican</a> than they had in any previous election, dropping from 65% Democratic support in 2020 to just 56% in 2024. The decline was driven primarily by Hispanic Millennials, who experienced one of the steepest drops of any demographic group, falling from 76% Democratic support in 2020 to 55% in 2024. But the rightward movement wasn’t limited to one racial group — Millennial voters across all racial lines shifted toward Republicans.</p><p>Gen Z emerged as the lone generational group that moved leftward. But at just 10% of the electorate, their growing Democratic loyalty couldn’t offset the broader trends among other generational groups.</p><p><strong>What if there had been universal turnout?</strong></p><p>The scale of Trump’s coalition-building becomes even more apparent when examining what would have happened <a href="https://docs.google.com/document/d/e/2PACX-1vR3cPQ81X2YD1Rm5cQgfudFWPgiJCpXx6K2njV91htbJ4nLCY5r6kASroo4icpj6j59YmNzjoddeWhT/pub#h.z92q619dqvrx">if every American had voted</a>. The actual two-party vote share in 2024 was 50.8% Trump to 49.2% Harris. But if everyone had voted, Trump’s margin would have increased to 51.1% to 48.9%.</p><p>This is because, for the first time since 2008, <a href="https://docs.google.com/document/d/e/2PACX-1vR3cPQ81X2YD1Rm5cQgfudFWPgiJCpXx6K2njV91htbJ4nLCY5r6kASroo4icpj6j59YmNzjoddeWhT/pub#h.fc9ox9yep3wk">non-voters preferred the Republican candidate over the Democrat</a>. Among those who stayed home and expressed an interest in one of the two major party candidates, Trump led Harris 52% to 48% — a historic shift given that non-voters had consistently favored the Democrat from 2008 through 2020. The pattern held across most swing states, with non-voters in Pennsylvania, Michigan, Arizona, and North Carolina all preferring Trump. Democrats do not have modest turnout to blame for Trump’s victory.</p><p>The parallel shifts among both voters and non-voters reveal an uncomfortable truth for Democrats. Trump’s victory wasn’t a fluke, it was the culmination of a broad-based shift that was evident across most demographic groups across America.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4c46b56826d9" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>