“I would rather see my community banned than see my community slaughtered.”
American Muslim voter Suehaila Amen’s striking statement was featured in a July 2024 New York Times video feature (Stockton and Semple Reference Stockton and Semple2024), showcasing why she no longer supported President Joe Biden. Suehaila Amen’s statement alluding to tolerating Trump’s Executive Order 13769, commonly referred to as the “Muslim Ban,” over Biden’s current foreign policy stances with Israel is symbolic of the preferences of American Muslims in Michigan.Footnote 1 Since 2023, there has been a staggering exodus of support by Muslim voters for Biden in the 2024 presidential primaries. In 2020, 65% of Muslims reported voting for Biden; in 2024, his support had plummeted to 18% among likely Muslim voters. In many ways, this trend highlights a case of a potential shift in partisan loyalties in an era when social scientists have predicted that partisan loyalties are stable and consistent (Abramowitz Reference Abramowitz2014; Levendusky Reference Levendusky2009).Footnote 2 From the perspective of Suehaila Amen and other Muslim voices, no policy issue was more central to their voting interests than US foreign policy decisions that interlinked with the level of civilian casualties in Gaza. This finding was corroborated by a recent poll and studies of Arab Americans and American Muslims highlighting that Gaza is their key policy concern for the 2024 election (Hasan-Aslih et al Reference Hasan-Aslih, Fade, Sohad and Nura2024; Mogahed Reference Mogahed2024). These voters blame the current Biden administration for failing to secure a solution that would diminish civilian loss in Gaza (Khatib, McKee, and Yusuf Reference Khatib, McKee and Yusuf2024; Stockton and Semple Reference Stockton and Semple2024).
As a researcher who is keenly attentive to the political behavior of Arab and Muslim communities, my in-depth interviews with Muslims during this period alerted me to their growing dissatisfaction with the Biden administration.Footnote 3 Weeks before the February 2024 Michigan primary, I spoke with colleagues who specialize in American politics about whether the growing dissatisfaction of Muslim voters was something that election forecasters had carefully considered. The answer was: not really—the forecasters largely assumed Arab and Muslim voter dissatisfaction was not significant enough to influence national election outcomes. Weeks later, the role of Muslim voters in Michigan became one of the biggest headlines for the Democratic primaries this year: 100,000 Democrats voted “uncommitted,” opting to send a political message instead of selecting Joe Biden as the nominee. In Dearborn and Hamtramck, two Muslim-majority cities in Michigan, 60% of Democrats voted “uncommitted.” The movement gained national momentum, with the uncommitted movement inspiring similar campaigns in Minnesota, Wisconsin, and other key states: approximately 800,000 Democrats nationwide voted for a version of “uncommitted” on their ballots (Leingang Reference Leingang2024). In August 2024, 30 delegates represented the uncommitted vote at the Democratic National Convention (DNC), the largest number of unaffiliated delegates to the DNC since civil rights activist Fannie Lou Hamer’s Mississippi Freedom Democratic Party gained delegates at the 1964 DNC.Footnote 4 Their presence and engagement became a key talking point during the 2024 DNC during its first-ever convened panel on Palestine that was organized at the insistence of the uncommitted delegates (Bilal Reference Bilal2024).
Minority voters utilize their votes in Democratic primaries in important ways, and this trend is crucial to understand because it can signal their voter preferences more meaningfully than captured within the constraints of a general election (Smith Reference Smith2022; Wamble Reference Wamble2025). This article draws attention to the case of American Muslims to exemplify why election forecasters should consider minority populations as a variable in state-level forecasting models, even when they may be a numerical minority nationwide. Given their relatively small number, American Muslims are a challenging population for building an election forecasting analysis. Muslims comprise 3% to 5% of the American electorate; therefore, discourse about the importance of the Muslim vote has generally been dismissed, which reiterates their relatively low political power in American elections (Pew Research Center 2017; Sediqe Reference Sediqe2024). If we disaggregate and consider the role of minority voters such as Muslims at a state level, particularly in swing states, their relative power and influence shift and become pivotal to consider.
If we disaggregate and consider the role of minority voters such as Muslims at a state level, particularly in swing states, their relative power and influence shift and become pivotal to consider.
As Blumenthal noted in the 2014 special issue of PS: Political Science & Politics on election forecasting, “For what purpose?,” it is a vital part of understanding the American electorate (Blumenthal Reference Blumenthal2014; Campbell Reference Campbell2012; Lewis-Beck Reference Lewis-Beck2005; Lewis-Beck and Stegmaier Reference Lewis-Beck and Stegmaier2014). For those scholars who are invested in election forecasting, if their purpose is to accurately capture the dynamics of election results—particularly for presidential races—considering minority voters more carefully will become more important with the shifting population and increased political acumen of minority voters. The 2024 presidential primary elections highlight the shortcomings of taking for granted minority voters and their role in shifting election outcomes.Footnote 5 It heeds the call of political scientists who specialize in minority political behavior and advocate for centering minority voters, given the influential role they have in American politics (Hutchings and Valentino Reference Hutchings and Valentino2004; McClain et al. Reference McClain, Ayee, Means, Reyes-Barriéntez and Sediqe2016). Election forecasting would benefit from better recognizing the pivotal role that minority voters can play in swing states and considering different strategies to understand and gauge their involvement. The following section explores the case of American Muslim voters in Michigan to underscore the political significance of this strategy. The discussion includes suggestions for analytical strategies to account for this within election-forecasting models.
THE CASE OF AMERICAN MUSLIM VOTERS IN MICHIGAN
Arab and Muslim organizers in Michigan have been at the heart of the Uncommitted National Movement. Michigan is home to a vibrant Christian and Muslim Arab community and a Muslim community inclusive of Arab, Black, and South Asian voters. While their interests may diverge, Arab and Muslim voters’ interests have converged this electoral year, given their increasing concern about the civilian death toll in Gaza. Palestinian Christians are a vibrant part of American life and have expressed concerns like Palestinian Muslims. For American Muslims of other cultural backgrounds, their social ties to Palestinians—coupled with a sense of connection to Muslims transnationally—have made this issue important to them (Sediqe Reference Sediqe2019; Shryock and Lin Reference Shryock and Lin2009).
Historically, American Muslims have relatively limited political power, but the tide is shifting in the state of Michigan. Michigan is home to the first Muslim-majority city council; the Democrats’ House Majority Leader (Abraham Aiyash) was Muslim; the mayor of one of the largest tourist destinations in Michigan is Muslim (Abdullah Hammoud); and the state elected the first Arab American Muslim to the US Congress (Rashida Tlaib) (Sediqe Reference Sediqe, Brown and Gershon2023). This growing political representation in Michigan has become a key feature of the Democratic Party and it signals a stable voter base for Michigan Democrats. This is meaningful, given how important Michigan was for the Democrats in the 2020 presidential election: the Biden ticket won the state by approximately 154,000 votes. Currently, there are an estimated 242,000 registered Muslim voters in the state (Emgage Michigan 2024).
Nevertheless, the escalating humanitarian crisis in Gaza has personally impacted Muslim Americans. Michigan families like the Abu-Shaban family lost 42 family members in one single airstrike (Rosales, Gallagher, and Sayers Reference Rosales, Gallagher and Sayers2023). Citizens and Muslim elected officials have called increasingly for more engaged diplomatic measures by the Biden administration. Abdullah Hammoud, the first Muslim mayor of the City of Dearborn—home to one of the largest per-capita cities of Arabs in the country—penned an op-ed for the New York Times titled, “I’m the Mayor of Dearborn, Michigan, and My City Feels Betrayed.” Hammoud (Reference Hammoud2024) wrote the following:
What compounds the constant fear and mourning is a visceral sense of betrayal. In the past three federal elections, Arab American voters in Michigan have become a crucial and dependable voting bloc for the Democratic Party, and we were part of the wave that delivered for Joe Biden four years ago….Until just a few months ago, I firmly believed that Joe Biden was one of the most consequential and transformative presidents…but no amount of landmark legislation can outweigh the more than 100,000 people killed, wounded, or missing in Gaza. The scales of justice will not allow it.
Following Hammoud’s editorial, campaigns such as the Listen to Michigan Campaign escalated efforts to have Democrats vote “uncommitted” in the Democratic primaries, thereby sending a signal to Democrats to take the issue of Gaza more seriously. The campaign’s goal of having 10,000 Democrats vote uncommitted on the Democratic primary ballot was greatly exceeded: approximately 100,000 Democrats selected the “uncommitted” option in the Michigan primaries, signaling their growing unhappiness with the Democratic Party.
This was a marked change in American Muslim support for President Biden. In the 2020 Collaborative Multiracial Post-Election Survey (CMPS), the sampled Muslims (N=592) reported having generally favorable feelings toward President Biden: 32.43% reported very favorable feelings and 31.76% somewhat favorable feelings. In a December 2023 nationwide survey of American Muslims, a sample of 200 respondents indicated that in 2020, 58% had voted for President Biden, which mirrors support levels found by the 2020 CMPS. This shifted for Muslims in 2023 and 2024 as a result of the escalation of conflict in Gaza. In both the December 2023 survey and a July 2024 survey (N=150), 67% of Muslim respondents stated that they would vote for “someone else” and not the Democratic Party nominee in the 2024 presidential election.Footnote 6 In their open-ended answers on the December 2023 survey, respondents shared sentiments such as, “I’m outraged that, as an American, the taxes that I pay are being used to fund a war that I don’t support.”Footnote 7 This is consistent with the most recent large-scale sample of American Muslim opinion, fielded by the Institute for Social Policy and Understanding in July 2024 (N=888).Footnote 8
Currently, Gaza remains the top policy issue for 71% of surveyed American Muslims who stated that it is their main concern in the upcoming national election (Mogahed Reference Mogahed2024). For minority voters, group-specific interests matter more than may be expected (Sanchez Reference Sanchez2008). Whereas election forecasters emphasize the role of economics in predicting vote change, the policy priorities of minority communities shaped by unique sociopolitical histories in the United States underscore the need to integrate external sociopolitical factors into their forecasting models for greater accuracy. Considering minority communities other than American Muslims, the dynamic history of Black voters in the United States has influenced their policy priorities as well (Walton Reference Walton1985). In a recent 2024 polling of policy priorities for Black voters, respondents noted that improving education, addressing problems related to the poor, and issues around race are salient policy concerns. This is distinctive when compared to the policy preferences of white voters—who were surveyed in the same study—and did not perceive these issues as priorities for their electoral concerns (Cox Reference Cox2024).
Whereas election forecasters emphasize the role of economics in predicting vote change, the policy priorities of minority communities shaped by unique sociopolitical histories in the United States underscore the need to integrate external sociopolitical factors into their forecasting models for greater accuracy.
HOW ELECTIONS FORECASTERS CAN ACCOUNT FOR MINORITY VOTERS’ INTERESTS
How do election forecasters resolve this challenge? Having acknowledged that minority communities with strong group interests can be a powerful voting bloc, the case of Michigan highlights how a state-level forecast may be a meaningful way to incorporate minority voters into forecast models. Centering the Electoral College vote as the dependent variable of interest, the analysis of election forecasters focusing on state-level forecasts provides a meaningful type of forecasting model that could incorporate minorities’ approval ratings (Campbell Reference Campbell1992; DeSart and Holbrook Reference DeSart and Holbrook2003; Jérôme et al. Reference Jérôme, Jérôme, Mongrain and Nadeau2021; Rosenstone Reference Rosenstone1983). Jérôme et al. (Reference Jérôme, Jérôme, Mongrain and Nadeau2021, 77) pointed out that “The Electoral College makes the geographical dimension of presidential elections particularly salient for the forecasting community.”
The importance of minority voters as a voting bloc that can alter electoral outcomes becomes more evident when their vote is considered in states that had slim margins of victory in recent elections. Considering the battleground states in 2020, for example, states including Arizona, Georgia, Wisconsin, Nevada, Pennsylvania, and Michigan would benefit from including minority-voter approval ratings in state-level forecasting models. Table 1 presents the 2020 margin of victory for four of the six key swing states as an example. In the case of Michigan, the margin of victory in 2020 was 154,188 votes. Considering that there are an estimated 242,000 Muslim voters and that 145,620 Muslims were estimated to turn out, the margin is small enough that their low approval rating for the Biden administration could influence the results. In Pennsylvania, the number of registered Muslim voters (i.e., 167,618) is more than double the margin of victory from 2020 (i.e., 81,660).
Table 1 2020 Swing States and the Margin of Victory by Overall Votes and Registered Minority Votes
In Wisconsin, the “uninstructed delegation” (i.e., the state’s iteration of the “uncommitted” vote) won 8.3% of the total votes (i.e., 48,162)—which is more than double the 20,682 votes that the Biden administration won Wisconsin with in 2020. In Dane County, where Madison is located, this percentage increased significantly, with 12.5% casting uninstructed ballots. Wisconsin had 15,142 Muslim voters in 2020 and now has 21,122 registered Muslim voters (Karnopp and Fowlkes Reference Karnopp and Fowlkes2024). In a context in which the Electoral College vote remains the primary dependent variable for success, forecasters would benefit from considering the number of registered minority voters in their analyses. This becomes most clear in the case of Georgia in 2020, where 12,670 votes determined the margin of victory and where 61,000 Muslims turned out to vote—almost five times more than the slim margin of victory in the state.
If we disaggregate and consider the role of minority voters such as Muslims at a state level, particularly in swing states, their relative power and influence shift and become pivotal to consider.
Including minority voters in election-forecasting models has precedence in earlier forecasting models. Rosenstone’s (Reference Rosenstone1983) forecast model for the 1960 presidential vote included the Roman Catholic population at the state level because of the prominence of John F. Kennedy’s identity as a Roman Catholic. His Roman Catholic identity became salient in the electoral campaign, given the anti-Catholic sentiment that pervaded the discourse in that period (Carty Reference Carty2001). Estimations indicate that 80% of Roman Catholics voted for Kennedy, underscoring the importance of Rosenstone’s forecast model (Schneider Reference Schneider2005).
In addition, Rosenstone’s (Reference Rosenstone1983) forecasts for the 1952 and 1968 elections included a variable accounting for the “mismanagement of war,” accounting for the anti-war sentiment prevalent in those periods as well. Recent polling of Muslims in key swing states (i.e., Michigan, Pennsylvania, and Georgia) identified that the key policy issue affecting their vote is the Biden administration’s handling of the war in Gaza (Mogahed Reference Mogahed2024). In 2020, the key policy issue reported by American Muslims was healthcare (19%), whereas in 2024, 71% reported that Gaza is their key policy priority.Footnote 9 In the 2024 elections, state-level models would be strengthened by including the percentage of American Muslims in key swing states, as well as their current opinions about US support for the military conflict in Gaza. Moreover, given the anti-war social movement’s heightened engagement in 2024, with approximately 16,435 protests nationwide involving 1.62 million protesters, the anti-war protests have included voters across cross-cutting cleavages and ethno-racial backgrounds (Ulfelder Reference Ulfelder2024). A variable comparable to Rosenstone’s (Reference Rosenstone1983) “mismanagement of war” would be useful to include for all registered voters in 2024 to measure the general American public’s support for continuing to fund military weapons to Israel.
Moreover, Campbell’s (Reference Campbell1992) inclusion of regional variables in state-level models captured shifting racial attitudes and their impact on candidates running for office. His inclusion of regional-specific variables to specific states provided additional precedence and examples of how forecasters may consider minority voters’ interests in specific state-level models.Footnote 10 This relates to a larger question of when specific variables should be included in forecasting models (Campbell Reference Campbell1992). In analyzing the discourse surrounding election forecasting, preeminent Black politics expert Hanes Walton alluded to the temporal phenomenon of considering when race matters for elections (Walton et al. Reference Walton, Allen, Puckett, Deskins and Starks2010). In Walton’s analysis, the variable was of greater significance when a Black candidate was running for office (e.g., Shirley Chisolm in 1972 and Barack Obama in 2008), as in the current case with Democrat Kamala Harris as the first Black and Indian American presidential candidate. These strategies underscore how state-level models can consider the key policy preferences of minority voters in swing states, as well as related policies (e.g., racial attitudes when a candidate is from a nonwhite background), in forecasting models when specific policy issues or candidates make them more salient.
CONCLUSION
Election forecasting is a unique space in which public discourse and academic discourse converge. In the case of American Muslims, there has been meaningful public discourse surrounding their influence in electoral outcomes but limited insight in academic discourse in determining their political behavior and inclusion in forecasting models. By highlighting the shift in dissatisfaction with the Democratic Party’s presidential ticket among American Muslims, this analysis explores how policy concerns have affected their voting preferences. The intricacies of Muslim voters would be easy to ignore if the impacts were not so consequential. As the case of Michigan demonstrates, a small minority can strategically organize power and influence election results. In this case, it can cost the presidential election and challenge election forecasting if they are ignored in states where they have an influence.
As researchers of election forecasting move forward, consideration of specialized populations must occur. Including minority voters in state-level analysis is one pathway forward that can better capture these sentiments in election forecasts. In many ways, the high dissatisfaction among Arab and Muslim voters served as a “canary in the coal mine,” as witnessed by President Biden ending his reelection campaign only months before the election. Election forecasters must take greater heed in examining the potential of minority voters in swing states. They need to better understand how growing levels of distrust by segments of voting-eligible citizens can have a profound impact on voter turnout and on which presidential ticket is successful. Reflecting on the case of Muslims exemplifies how this process may unfold for other minority voting blocs.
In many ways, the high dissatisfaction among Arab and Muslim voters served as a “canary in the coal mine,” as witnessed by President Biden ending his reelection campaign only months before the election.
This trend may be seen in other minority voters and could have a positive outcome. For example, Black American voters experienced a substantial surge in voter-registration rates in the summer of 2020 in response to the murder of George Floyd and the national conversation that elevated the #BlackLivesMatter movement (Frey Reference Frey2021). The fact that 87% of Black voters ultimately voted for President Biden highlights how pivotal a minority voting bloc can be. This is particularly true in states with a slim margin of victory, as in Georgia in 2020, which has a sizeable and influential Black voter base (Budiman Noe-Bustamante Reference Budiman and Noe-Bustamante2020). This example encourages discussion of how minority voting blocs and their role should be considered beyond the case of American Muslims. It underscores the need for deliberative thinking by scholars who are invested in election forecasting to build a more curated strategy involving more community-engaged research when engaging minority voters, particularly in swing states. The Democratic Party’s shift from incumbent Joe Biden to Kamala Harris at the top of the presidential ticket presents many challenges for election forecasters. In the case of minority voters, the fact that Harris is the first Black and South Asian woman to run for president highlights the need to consider issues of race in current forecasting models. Whereas the inclusion of Muslims and Arabs in state-level models such as Michigan is necessary, the inclusion of Black and Asian American and Pacific Islander voter approval ratings also increases in relative importance for the presidential ticket. The case of American Muslims provides an important point not only for who is on the ballot but also when the social movement strategies of minority voters should become more serious considerations for political scientists.
CONFLICTS OF INTEREST
The authors declare no ethical issues or conflicts of interest in this research.
Open access
