Note: JPIA selects articles on the merits of each piece's academic rigor and contribution to policy discourse. Our editors and editorial board do not endorse any opinions expressed therein.

By Jose Luis Gandara

Abstract

I investigate the effects of the race of the perceived beneficiaries of an affordable housing development on white homeowners’ support for the project, using an online survey experiment with 520 participants. I find that priming respondents to believe a nearby proposed project’s residents will likely be Black significantly increases opposition compared to the white prime. However, the effect is moderated by respondents’ racial attitudes, such that self-reported racially sensitive individuals instead become more supportive when led to believe a project’s residents will be Black. Despite racial cues increasing opposition, respondents do not express different concerns with development in a racialized context. These results suggest that race is a central factor driving attitudes toward affordable housing; however, racially motivated public commenters mask their concerns behind those ostensibly unrelated to race. Policymakers concerned with advancing equity while addressing the housing crisis may reconsider public comment’s role informing them of the public’s preferences toward development.

Introduction

California’s Senate Bill 35 has proven incredibly effective in spurring housing development, having led to over 18,000 housing units entering the development pipeline from 2018 to 2022 (Manji and Finnigan 2023). Perhaps surprisingly, it does not change zoning laws or subsidize new projects. Instead, the law allows certain developments to avoid one of the greatest obstacles to residential construction: discretionary approval processes. The legislation allows qualifying housing projects in cities behind on production goals to receive ministerial approval, rather than wait for approval from the local government. Many jurisdictions within California and across the United States require proposed residential developments to be approved through a discretionary approval process, where in addition to complying with zoning and other regulations, projects must win approval from local government officials. Frequently, these approval processes are characterized by long, uncertain timelines that discourage development. In San Francisco, for example, projects take 1,000 days on average to be approved (Landes 2022). Decisions to approve or reject a project are frequently made at public meetings which invite public comment, and approvals may sometimes even be appealed by individual citizens and litigated further.

During public comment, arguments against development frequently cite its potential impacts, for example, increased traffic, strain on local resources, or the project’s aesthetics defying the neighborhood’s character. While opponents often object to the impacts of having more neighbors, rarely do arguments take issue with who an affordable development’s residents might be. However, an expansive literature illustrating the centrality of race in the formation of policy preferences gives reason to believe concerns about race may nevertheless be present in public meeting rooms (Hutchings and Valentino 2004). The absence of racial considerations during public comment may be further explained by scholarship on social desirability bias. Are attitudes regarding affordable housing in part a reflection of the perceived beneficiaries of affordable housing policies? If opposition to development is driven in part by race, do opponents self-censor those motivations?

This paper seeks to investigate whether white homeowners’ support for a proposed affordable development is affected by the race of the residents they expect it to house. I experimentally test this relationship with an online survey of white homeowners living in the United States. With racially discriminatory motivations potentially driving opposition to affordable development, policymakers concerned with equity may do well to reconsider the role of public comment as a source of information of the community’s preferences—or reconsider discretionary approval processes altogether.

The paper will be as follows. Section 2 describes the role of public comment and discretionary approval processes in the housing crisis through a case study of California. Section 3 reviews the existing literature on racial attitudes and the formation of policy preferences, addressing residential integration and housing policy in particular, as well as the effects of social desirability bias. In Section 4, I develop the theoretical motivations informing my expectations for this study before describing the survey design and summary statistics. Section 5 outlines the empirical methodology analyzing the survey data, and Section 6 explores the results and limitations of my analyses. In Section 7, I conclude by discussing the implications for policymakers seeking to address the housing affordability crisis while advancing equity goals.

Background

California Case Study: The Housing Affordability Crisis and Discretionary Approvals

Communities in California, alongside many others across the United States, are experiencing a housing affordability crisis. Median rent in the state has increased 38 percent since 2000, while the median renter’s income has only grown by seven percent (California Affordable Housing Needs Report 2023). Lower-income residents bear the brunt of this affordability crisis: In the Bay Area, 70 percent of extremely low-income households are severely cost-burdened—spending more than 50 percent of their income on rent—compared to only one percent of moderate-income households (Karlinsky 2023). Black and Latino households are particularly likely to be burdened by the cost of housing, with 60 percent and 55 percent of each group spending more than 30 percent of their income on rent compared to 42 percent of their white counterparts (Housing Burden n.d.).

While exacerbated by growing income inequality, the housing affordability crisis is primarily caused by housing supply failing to keep pace with growing demand (Erdmann 2023). From 2010 to 2020, California added 3.2 times more people than housing units (McGhee, Paluch, and Hsieh 2021). The mismatch is even more dramatic in the Bay Area, which added 658,000 jobs and only 140,000 housing units from 2011 to 2017 (Karlinsky and Wang 2021). Insufficient housing supply in urban job centers forces higher-income newcomers to compete for housing with lower-income residents, who may relocate to farther-flung suburbs and exurbs and face increasingly large commutes. To address undersupply and account for future population growth, the California Department of Housing and Community Development (HCD) projects that the state must build more than 2.5 million homes over the next eight years, with at least one million units set at prices affordable to lower-income households (A Home For Every Californian: 2022 Statewide Housing Plan 2022). Such a dramatic increase in the housing stock would require an unprecedented rate of residential construction not seen since at least 1960 (Construction Permits n.d.). To accomplish this monumental task, the barriers obstructing new housing construction must be understood and addressed. 

A primary obstacle to building housing in California is the entitlement process, during which a proposed development must receive permits and approval from a planning commission or other governing body overseeing the jurisdiction where the construction will take place. A national survey of residential developers found permitting and the approval process to be their principally cited regulatory barrier, with 70 percent of single-family and 75 percent of multifamily builders reporting it as a major concern. In cities with a population of at least one million, 81 percent of developers identify the approval process as a top concern (Colton and Ahluwalia 2019). While developers can navigate objective building standards to produce a proposal that complies with zoning and other regulations, public officials may nevertheless deny a project if convinced by community opposition expressed during public comment periods.

Public Commenters and Their Influence on Approvals

Such participatory mechanisms allow for small but highly engaged groups of residents hostile to development to slow, obstruct, or outright reject housing developments. These groups are often colloquially referred to as NIMBYs (Not In My Back Yard). About 50 percent of developers in a California survey reported that the local planning commissions, city councils, and county boards that must approve their projects have been persuaded “a lot” or “a great deal” by public opposition arguments to new housing development (Burinskiy, Green, and Takahashi 2021). While opportunities for public input are designed to foster democratic decision-making and address community concerns, not all residents use them equally. An analysis of public comment at approval meetings in Massachusetts found that residents who give public comment are far more likely to be homeowners, white, and over the age of 50 than the general population, and that they generally speak in opposition to the proposed development (Einstein, Glick, and Palmer 2020). A recent study by Alexander Sahn (2024), which analyzes over 40,000 public comments delivered to the San Francisco Planning Commission over the past two decades, confirms this unrepresentativeness. Additionally, Sahn finds that white commenters are uniquely influential compared to all other commenters, with their preferences better predicting commission votes. When public officials rely on public comment to inform the discretionary housing approval process, they risk making decisions based on a potentially distorted understanding of the public’s preferences. Comments may be unrepresentative of the broader community. Public comments may also not directly reflect the true opinions of commenters—each commenter is able to craft their message to be a persuasive argument, rather than a direct articulation of their honest concerns.

Some observers of housing approval meetings around the Bay Area have criticized NIMBY complaints as dishonest, concealing opposition motivated by race and class (Perigo 2020). Some opposition arguments certainly appear to be frivolous attempts to weaponize procedural steps against development: for example, opponents of a student housing development in Berkeley recently blocked the project on the grounds that the environmental impact report did not consider the noise from student residents to be a form of pollution (Hernandez and Apodaca 2023). The immense influence of public comment within discretionary approval processes invites deeper study. To what extent do racial concerns motivate opposition to affordable housing?

Literature Review

Based on the literature, I develop the theoretical case that racial attitudes and assumptions around the perceived race of those who live in affordable housing can motivate white opposition to nearby affordable development, despite racial arguments being absent during public comment. My research question addresses a critical gap in the literature to illustrate that public opposition to affordable housing development is motivated in part by underlying racial concerns that go unmentioned during public comment.

Racial Attitudes in the Formation of Policy Preferences

Contemporary social scientists continue to understand racial prejudice in a manner consistent with Allport’s early definition of prejudice as “antipathy based on a faulty or inflexible generalization” (Allport 1954). This definition contains two elements: a negative feeling toward another group and a stereotype about that group. However, scholars have had to adjust their methods of measuring racial prejudice among whites as its expressions have continuously changed in the decades since the Civil Rights Movement. Explicit expressions of racial prejudice have declined rapidly: in 1945, 55 percent of white respondents to a National Opinion Research Center survey answered that white people should be given the first chance at any job (Schuman et al. 1997), but by 1972, 97 percent of white respondents to the General Social Survey (GSS) answered that Black people should have an equal chance for employment (Quillian 2006). However, when the 1990 GSS asked respondents to rate racial and ethnic groups on seven-point scales of hard working/lazy, violence prone/not violence prone, and intelligent/unintelligent, white respondents rated whites an average of more than one point higher than they did for Black and Hispanic people (Smith 1991). Racial stereotypes and affect may persist, but whites generally avoid expressing support for policies and practices whose discriminatory nature is explicitly clear. 

Attempting to make sense of this phenomenon, several theories of a contemporary “new racism” have emerged, including symbolic racism (Kinder and Sears 1981), modern racism (McConahay 1986), ideological refinement (Jackman and Muha 1984), and laissez-faire racism (Bobo et al. 1997). While their specific interpretations vary, “all perspectives emphasize that new prejudice involves a combination of racial and ostensibly nonracial attitudes” joining conservative values, particularly individualism, with negative affect towards minorities (Quillian 2006). This new racism represents a worldview that racial discrimination is no longer a problem after the progress of the Civil Rights Movement, that current racial inequalities are the result of individuals’ unwillingness to work hard, and that government assistance for these individuals is ineffective and unfair (Hutchings and Valentino 2004, p. 390). 

Studies find that white racial attitudes reflecting this perspective predict support for a host of issues, including those ostensibly unrelated to race such as crime and punishment as well as welfare and other redistributive programs (Kinder and Sanders 1996; Hurwitz and Peffley 1997; Gilens 2009). Some critics contend that the new racism thesis neglects that its measures of racial attitudes are confounded with aspects of political ideology unrelated to race (Sniderman, Crosby, and Howell 2000). Many of the policies that may ameliorate racial inequality, like welfare, require greater government intervention or restrict behavior in ways that could be opposed on ideological grounds rather than racial ones. 

The political values of whites and the way policies are framed can also significantly affect how racialized policies are evaluated. For example, framing a policy to highlight its universal benefits rather than the benefits for specific groups can significantly increase white support (Sniderman and Carmines 1997). Nevertheless, many studies demonstrate that racial attitudes strongly influence support for racialized public policies like affirmative action and welfare, even when controlling for nonracial political values (Hutchings and Valentino 2004, p. 392). In addition to negative attitudes, racial attitudes motivated by concern and a desire to ameliorate injustice can shape policy preferences. Racial sympathy, theorized by Chudy (2021) as “white distress over Black misfortune,” can motivate whites to support policies that they believe will benefit Black Americans and oppose those thought to harm them. For racially sympathetic individuals, making Black Americans salient as a policy’s perceived beneficiaries can increase support, rather than drive opposition.

Race and Preferences on Residential Integration and Housing Policy

Racial attitudes also affect preferences toward housing policies and attitudes on residential integration. Scholars have largely considered three competing theories to explain opposition to residential integration: that it arises from whites’ assumptions about interethnic class differences, that racial groups share mutual preferences to live alongside in-group members, and that it results from whites’ hostile attitudes toward out-group minorities. Leveraging a large multiethnic survey in Los Angeles, Bobo and Zubrinsky (1996) find little support for the former two theories but identify racial prejudice as a significant contributor driving residential integration attitudes. Indeed, examinations of focus groups and surveys find that perceptions of low-income people and minorities are strong determinants of affordable housing opposition (Tighe 2009). By design, affordable housing is restricted to low-income residents; furthermore, racial income inequality causes racial minorities to be potentially more likely than whites to qualify for affordable housing in a given area. Therefore, respondents asked about affordable housing likely already have assumptions about a development’s future residents and attitudes about these groups will condition support for affordable housing more broadly.

Accordingly, a survey of white and non-white suburban residents found that white support for “affordable housing” was significantly less than white support for “lifecycle housing,” while the effect of framing was nonsignificant for non-whites (Goetz 2008). These results indicate that affordable housing is not an uncharged policy in the minds of many white respondents; it is a policy which brings to mind certain assumptions about its perceived beneficiaries. These assumptions are reflected in the concerns levied by opponents. A survey of affordable housing developers found that safety and crime are the primary concerns cited by residents opposed to affordable housing developments (Scally and Tighe 2015). Neither of these concerns explicitly mention race. However, both reflect stereotypes associated with Black Americans and low-income people. 

Operationalizing Race in Considerations of Policy 

While there is a general consensus among scholars that racial attitudes can affect whites’ policy preferences, there are several competing theories explaining this relationship with important implications for studies seeking to operationalize race. Kinder and Sears (1981), who developed the theory of symbolic racism, posit that racial attitudes are influential through a psychological process that begins with anti-Black affect acquired in childhood. Policies that are thought of as primarily benefiting Black Americans are assessed through this negative lens, informing opposition. In contrast to this individual-level explanation, group-oriented theorists including Blumer (1958), Bobo (1983), and Sidanius and Pratto (2001) argue that perceptions of conflict between the interests of racial groups drive attitudes toward racialized policies. In particular, theories of racial threat hold that whites’ racial attitudes and policy preferences are responsive to the size of minority populations and their proximity (Tolbert and Grummel 2003). Both views may complement one another in explaining the relationship between race and policy preferences: racial cues can elicit negative stereotypes about the perceived beneficiaries of a policy, as can increasing the salience of conflicting group interests. But racial sympathy, operating as a distinct but related attitude to the negative attitudes just described, motivates support for policies when their perceived beneficiaries are Black (Chudy 2021). Making Black Americans salient as a policy’s beneficiaries may cause two effects in opposite directions, increasing opposition for whites with racial animus while driving support among their racially sympathetic counterparts.

With respect to affordable housing policy, merely reinforcing the assumption that an affordable development’s residents will likely be Black can trigger both processes. On the one hand, leading white respondents to believe that an affordable development will likely have Black residents can elicit anti-Black stereotypes, which in turn decreases support for the proposed project. Respondents who endorse negative stereotypes about laziness and deservingness of public assistance may oppose the policy because it benefits the stigmatized out-group. Additionally, priming respondents to believe a project will house Black residents may cause them to consider implications on the racial makeup of their community. Concerns of an out-group’s incursion into the neighborhood may activate feelings of racial threat, motivating opposition to the project out of a desire to maintain distance from the out-group. For racially sympathetic individuals, bringing Black Americans to mind as the likely residents may increase support with the intention of providing opportunity and ameliorating past injustices. 

Social Desirability and Self-Censorship of Racist Beliefs

If white opinion on affordable housing is shaped by race, why are these considerations absent during public comment? The declining social acceptability of expressing racist beliefs has discouraged individuals from openly articulating racial motivations. Pressured to conform to social norms, individuals frequently seek to portray themselves in a positive light, demonstrating a social desirability bias that can obscure researchers’ collection of self-reported data (Nederhof 1985). Demonstrating this bias, experiments in racial attitude surveys find that whites express less liberal racial attitudes and racial policy preferences as their privacy increases (Krysan 1998). During a housing approval meeting that invites public comment, opponents find themselves in a highly public setting alongside their neighbors and community leaders. Without anonymity and knowing that their comments may be remembered by people in their community, opponents may feel pressured to self-censor potentially racist beliefs in order to remain socially desirable. If opponents of affordable housing development are motivated by race, they would be reasonably expected to feel reluctant to openly express so.  

Rather than explicitly citing racial concerns, opponents may be more likely to mask their opposition with race-neutral reasons. Crandall and Eshleman’s (2003) justification-suppression model holds that prejudices are restrained from being expressed by norms and values that suppress them and are only expressed when released by justifications. Racially motivated opponents of affordable housing may grasp at justifications ostensibly unrelated to race—including concerns about increased traffic or competition for parking, the project’s aesthetics, adverse environmental impacts, or strain on schools and public services. However, this phenomenon of masking racial motivations behind nonracial concerns has not yet been demonstrated by scholars. With social desirability bias obstructing researchers’ understanding of respondents’ true feelings, it is difficult to demonstrate that public comments may misrepresent actual concerns to public officials. But by triggering racially motivated opposition to development without respondents’ knowledge, an experimental design which asks respondents to express their concerns about a project may catch the self-censorship in action. Does manipulating the race of the perceived beneficiaries of an affordable development change the concerns respondents articulate about the project? 

Theory, Data, and Empirical Approach

Theory and Argument

Affordable housing is a racialized policy issue. It is designed to benefit low-income individuals, a population that is disproportionately composed of racial minorities. Like with other racialized issue areas, whites’ preferences regarding affordable housing are shaped by their racial attitudes. That is, what a respondent thinks about racial minorities contributes to their evaluation of an affordable development. When a project is proposed nearby, additional feelings of racial threat may be triggered in response to the implications for one’s neighborhood demographics. The relationship between race and support for affordable housing is, therefore, likely stronger when considering a specific local project rather than housing policies more broadly.

With affordable housing racialized in this way, whites with negative stereotypes toward Black Americans may be already more likely to oppose affordable development. Making these stereotypes more salient will further increase opposition among those with negative racial attitudes. Thus, I expect priming respondents to believe the perceived beneficiaries of an affordable development will be Black will increase opposition, and that respondents’ racial attitudes will moderate this effect. However, racially motivated opponents will not necessarily be forthright about their concerns when asked. The pressure to remain socially desirable by conforming to norms stigmatizing racial prejudice drives opponents to self-censor their beliefs and attribute their concerns to reasons ostensibly unrelated to race. Therefore, I expect no statistical difference between conditions regarding respondents’ concerns with the development.

Survey Design

To test the expectations regarding race and its relation to white Americans’ support for affordable housing, I administered a survey on Prolific in March 2024 to a sample of 580 white homeowners in the United States. Prolific is a participant recruitment platform that has yielded reliable results representative of targeted populations (Coppock and McClellan 2019). I limited my sample to white homeowners because compared to minorities and renters, these individuals are more likely to appear in approval meetings (Einstein, Glick, and Palmer 2020) and exercise greater influence over approval decisions (Sahn 2024). Restricting the sample to whites also increased the likelihood of detecting smaller effects, as I would need to otherwise control for additional racial groups who could respond differently to racial cues. Due to resource constraints, I could not ensure that Prolific collected a geographically representative sample of respondents. Participants were compensated $2 to complete a roughly 10-minute survey. Further discussion of the sample selection is described in Section 4.4.

At the start of the survey, participants were randomly assigned into two conditions: White Prime and Black Prime. After first responding to a standard set of demographic questions, including age, income, education level, partisanship, political ideology, and racial attitudes, respondents answered a series of questions about their neighborhood. Respondents were asked whether they live in a single-family home, to estimate the percentage of homes in their neighborhood that are single-family, and to report the level of income and racial diversity in their neighborhood. In addition to providing data for potential controls and randomization checks, these questions prompted respondents to think about the details of their own neighborhood. Such priming may help respondents provide a more realistic response in the following section of the survey, where they were asked to imagine a hypothetical scenario. Respondents’ perceptions of their neighborhood may not always align with objective reality. However, because opinions on local development are shaped more by perception than by fact, these subjective impressions are relevant to understanding homeowner attitudes toward new housing.

Respondents were then asked to imagine that a developer who plans to build an affordable housing development a half-mile from their home has left a pamphlet on their doorstep to inform them of the project. Respondents reviewed the pamphlet, which contained information about the number of units in the project, their affordability level, the development’s added community amenities, and potential environmental impacts. Notably, the pamphlet featured an image of a family. The pamphlet was identical except that the family was white in the White Prime condition and Black in the Black Prime condition. This simple racial prime was intended to reinforce the perception that the development’s residents would likely be Black; accordingly, I also refer to the Black Prime condition as the treatment. After correctly answering a comprehension check, respondents answered questions about their support for the project and concerns with the development. On a seven-point Likert scale, respondents expressed how much they agreed with various concerns, including impacts on traffic and parking, the strain on local resources, the project’s aesthetics, and the explicitly racialized reason, “I am concerned about changing the racial composition of my neighborhood.” A full list of survey questions is available in Appendix C.

Measuring Racial Attitudes

Racial attitudes condition how participants respond to racial cues as they evaluate a proposed affordable housing development. Racial animus may prompt respondents to increase opposition when the project’s perceived beneficiaries are Black; conversely, racial sympathy may drive an increase in support in response to the priming. Meanwhile, the motivation to remain socially desirable may encourage respondents to provide what they consider to be a politically correct answer. Asking participants panels of questions capturing their racial attitudes on each of these dimensions is time-consuming and risks revealing the intent of the study or otherwise altering responses. To best capture these attitudes together, I ask respondents how much they agree with the statement that “It is important to me that I be seen by others as a racially sensitive person." To identify whether the effect of racial cues on support for affordable housing is moderated by racial attitudes, this question efficiently sorts respondents into the relevant groups for analysis. Racially sensitive participants may respond to racial cues with greater support because of a genuine desire to ameliorate injustice; however, they may also believe supporting a project housing Black Americans is the more socially desirable position. Conversely, racially insensitive individuals may respond with heightened opposition.

Sample Selection, Randomization, and Descriptive Statistics

A total of 580 participants responded to the survey from March 7 to March 17, 2024. After eliminating non-homeowners, non-whites, and those who failed the attention check, the sample contains 522 respondents. Summary statistics of the entire sample are listed in Table 1. The sample skews older, with the average respondent being 47 years old, and women are slightly overrepresented at 58 percent. A greater share of the sample self-identifies as a Democrat (57 percent) than a Republican (32 percent), although on average, respondents’ political ideology leans only slightly left of center. Just over half (59 percent) of respondents agree that it is important that they be regarded as racially sensitive, while 17 percent disagree. Notably, 77 percent of respondents believe there is a lack of affordable housing in their area.

To ensure participants were randomly assigned into each condition, I performed a series of t-tests comparing the mean values of various demographics and other control variables of interest across conditions. The full results are shown in Appendix B as Table B.1. I found no statistically significant differences between conditions, although women made up a slightly greater share of the treatment group. With randomization having successfully created two statistically indistinguishable groups, I do not include demographic controls in my regression models.

The regression models leveraged in the analysis use racial sensitivity as an independent variable, interacting with the condition to affect the outcome variables of interest—support and agreement with various concerns. However, it is possible that other variables correlated with racial sensitivity have independent and confounding effects. Among potential controls, political ideology is most closely correlated with racial sensitivity, with a one unit increase in conservatism associated with a 0.36 unit decrease in racial sensitivity. Additionally, a simple regression analyzing political ideology and racial attitudes’ association with support finds that both have independent, statistically significant relationships. Therefore, my regression models include political ideology as a control to isolate the effect of racial sensitivity, the condition, and their interaction.

Empirical Methodology

To answer the first part of my research question, whether manipulating the race of an affordable development’s perceived beneficiaries affects attitudes toward the project, I first perform a series of t-tests as described in Appendix A. These t-tests compare mean differences between the treatment and control on each outcome measure, including support for the development and agreement with various concerns about the project. For each outcome variable I then perform linear regression analyses to identify whether exposure to racial cues differentially affects participants on account of their racial attitudes. I do so by interacting the effect of exposure to the treatment with the effect of racial sensitivity, while also controlling for political ideology.

Linear Regression Model

I estimate the effect of racial priming on housing attitudes using the following regression model:

Outcome_i = β₀ + β₁Tx_i + β₂Sensitivity_i + β₃Tx_i x Sensitivity_i + β₄Controls_i + e_i

where Outcome_i represents one of two dependent variables: support for the development and level of agreement with stated concerns. Tx_i is a treatment dummy (1 = Black Prime condition, 0 = White Prime condition). Sensitivity_i represents how strongly a respondent agrees or disagrees that it is important to them to be seen by others as racially sensitive, measured on a seven-point scale where 1 indicates strong disagreement and 7 indicates strong agreement. Perhaps most relevant to my findings, Tx_i x Sensitivity_i  is the interaction term between receiving the treatment and racial attitudes, illustrating how the effect of receiving the treatment is different across levels of racial sensitivity. 

In this model, Controls_i  represents a seven-point scale measuring self-reported political ideology, where 1 signifies extreme liberals and 7 signifies extreme conservatives. The random assignment of the treatment eliminates the need to include other demographic controls. However, I suspect that racial attitudes and political ideology are highly correlated, and I also expect that both measures have independent relationships with support for affordable housing. To account for the effects of political ideology, I include it as a control variable. In the following subsections, I discuss this model as applied to each outcome variable and provide an overview of my predictions.

Support for Development

I first use this regression model to examine the effect of exposure to racial cues on support for the affordable development, as moderated by racial attitudes. The dependent variable Support_i represents a seven-point scale where one signifies strong opposition and seven signifies strong support for the affordable housing development. 

I hypothesize that racial priming will decrease support among respondents with low racial sensitivity (β₁ < 0) by reinforcing the perception that a project’s beneficiaries will likely be Black. For these respondents, raising the salience of Black people as potential neighbors will activate racial stereotypes or feelings of group threat, lowering support for the project. 

Conversely, I expect that individuals who value appearing racially sensitive may be more supportive of affordable housing regardless of racial priming (β₂ > 0). Existing scholarship has established that calling income-restricted development “affordable housing” elicits differential effects on support between white and non-white respondents (Goetz 2008). Therefore, racially sensitive individuals may be inclined to lend greater support to a policy they believe generally supports Black Americans, even absent racial priming. Finally, I predict β₃ > 0, meaning that racial priming will amplify support among respondents high in racial sensitivity, as they may be more motivated to express pro-housing attitudes when Black beneficiaries are salient.

Concerns with Development

I use the regression model to analyze the treatment and racial attitudes effects on respondents’ agreement with frequently-cited concerns to development. The dependent variables measure agreement on a seven-point scale with concerns including traffic and parking, project design, strain on resource, property values, crime, and changing neighborhood racial composition. 

I predict the treatment will have statistically insignificant effects for all concerns as I expect any racialized motivations to be distributed across the concerns without a distinguishable pattern. Further, I expect self-censorship motivated by social desirability bias to prevent any statistically significant effects from the treatment RacialComp_i,, which is explicitly racialized.

I also predict Sensitivity_i will be statistically insignificant for racially benign concerns including traffic, design, and resources, as these issues do not connect to race in the minds of respondents. Similarly, I expect Sensitivity_i will be statistically insignificant for implicitly racialized concerns including property values and crime. However, for the explicitly racialized concern of changing neighborhood racial demographics, I expect to see an effect, as more racially sensitive individuals would be less likely to agree with a concern that so clearly reflects an aversive view of Black Americans, even without the racial prime bringing Black Americans to mind.

Lastly, I expect the interaction term between racial sensitivity and receiving the treatment to be statistically insignificant for all concerns. While I predict racial cues will increase opposition among those with negative racial attitudes, I believe these respondents will avoid attributing their opposition to a plainly racially discriminatory concern. Therefore, they will cite concerns at similar rates as their counterparts who do not receive the racial prime.

Results

T-Tests Comparing Differences in Means

A series of t-tests comparing differences in means across conditions finds no statistically significant differences in any of the outcome variables. In both conditions, the average respondent is slightly supportive of the project, scoring about 4.5 on a seven-point scale where 7 is strongly supportive. On average, respondents in both conditions are not very likely to give public comment, scoring about 2.3 on a 4-point scale where 4 is very likely. Respondents neither agree nor disagree with most concerns on average; however, most respondents disagree with the concern about neighborhood racial composition. Notably, only 65 respondents, or 12.3 percent of the sample, agree at least slightly with the concern that the development might change their neighborhood’s racial composition. These respondents are split almost evenly across conditions, with 31 in the control and 34 in the treatment. A full comparison is available in Table 2.

These results indicate that analyzing the treatment alone does not reveal a clear effect on respondent’s attitudes toward affordable housing. That is, perceiving the beneficiaries of an affordable housing development to be Black does not spur a noticeable change in support for the project or concern regarding it. However, it is possible that the apparent lack of movement across the sample of respondents is the result of differential and countervailing effects within it. Exposure to racial cues in the experimental condition may cause respondents with negative racial attitudes to become more opposed to the development, while driving support among racially sensitive individuals. To investigate the potential moderating effect of racial attitudes, I continue my analysis with the linear regression model previously described.

Support for Development

Analyzing the interaction effect of racial attitudes and exposure to racial cues on support for the development yields interesting results generally confirming my hypotheses. Full regression results are available in Table 3. 

Priming respondents to perceive an affordable housing development’s residents will likely be Black has differential effects on support on account of racial attitudes. A one-unit increase in racial sensitivity is associated with an additional 0.17-unit increase in support for the development when its perceived beneficiaries are Black. This effect is statistically significant at the 10 percent level (p = 0.075), and it may have been identified with greater precision and certainty with a larger sample size. This result suggests racially sensitive individuals grow more supportive of a project when they are led to believe it will house Black residents, while respondents with negative racial attitudes become more opposed. A visual representation is shown in Figure 1. For the least racially sensitive individuals, support decreases in response to the treatment; but for the most racially sensitive, support increases. The differential effect of racial attitudes is illustrated further by respondents who self-report neutral racial sensitivity. For these indifferent individuals, there is no change in support upon receiving the treatment.

Further confirming my hypotheses, receiving the treatment is associated with a 0.821-unit decrease in support among the least racially sensitive respondents (p = 0.1). While I do not find a statistically significant association between racial attitudes and support, the coefficient’s positive sign aligns with my expectations. A one-unit increase in racial sensitivity is associated with a 0.076-unit increase in support among respondents who do not receive the racial prime (p = 0.28), with the 95 percent confidence interval suggesting mostly positive effects ranging from -0.062 to 0.213. A larger sample size may have yielded statistically significant results.

Notably, political ideology as a control variable has a large and statistically significant association with support. A one-unit increase in conservatism is associated with a 0.311-unit decrease in support (p = 0.000), suggesting that attitudes toward affordable housing are closely related to political ideology. Running the regression model without including political ideology as a control yields much larger and more significant results for the effects of treatment, racial sensitivity, and their interaction term, as shown in Table A.2. The stark contrast affirms my decision to include political ideology as a control variable to account for its confounding effects with racial attitudes on support and other outcome variables.

Concerns with Development

Receiving the treatment did not change respondents’ concerns about the project, regardless of racial attitudes. These results, depicted in Table 5, support my hypothesis that racially motivated opponents avoid explicitly articulating their concerns. However, there are additional insights in these measures of concern. Racial attitudes were not associated with any concerns, except two: design and racial composition. Racially sensitive individuals are more inclined to levy concerns with a project’s design, and less so anxieties about neighborhood racial demographics. Whether this reflects an honest difference in concern or a pattern of racially self-aware deflection is unclear. One might expect that increasing racial sensitivity is associated with lower racial demographic concerns, but why would racially sensitive individuals be more concerned about a project’s design? The racial sensitivity measure captures how considerate respondents are to how they appear on race. While many racially sensitive individuals are also racially sympathetic, some may have negative attitudes that they seek to hide to remain socially desirable. These individuals may have substantively similar concerns as others, but, aware of the stigma of appearing racist, they articulate their opposition through the most innocuous, racially benign concern available: design. Further research may find a list experiment useful in disentangling the potential social desirability bias in respondents’ answers.

Limitations

The study has several notable limitations. As an exercise in imagination, this design does not ensure that participants provide responses that reflect how they would respond to a real-life scenario. While I demonstrate that racial priming has short-term causal effects on housing attitudes, it is unclear whether these results are generalizable to actual housing approval meetings. For example, racially sensitive respondents, who are also disproportionately Democrats and politically liberal, may respond to racial cues in a survey with greater support because of their political identity as someone who supports racially liberal policies. It is unclear whether this support persists when respondents are presented with the actual possibility of change in their neighborhood. 

Many scholars encounter difficulty producing externally valid causal effects when studying the impact of race on policy preferences. The most reliable method, a randomized controlled trial in the field, is frequently difficult to design, prohibitively expensive, and too impractical to implement—although still sometimes possible. For example, Ryan Enos’ (2014) study investigating the effect of intergroup contact on exclusionary immigration attitudes leverages a clever design featuring Spanish-speaking confederates placed at specific Massachusetts commuter trains during rush hour. Rather than run an online survey instrument, I could have delivered pamphlets with a similar racial prime to households; however, I would have either needed the developers’ consent if the pamphlets advertised actual developments or navigated the ethical considerations of misleading residents about a fictional project.

Additionally, the estimated effect sizes of receiving the treatment are a potential overestimate of the actual effect of race. The White Prime condition cannot be considered a neutral control condition, as viewing a white family on the pamphlet could trigger in-group favoritism effects motivating increased support of the development. Additionally, showing a white family could correct respondents’ assumptions about the race of affordable housing residents. Regardless of the potential impacts, the difference between receiving treatment and control conditions is not merely that the racial priming leads respondents to perceive the beneficiaries of the project as Black. In the control condition, respondents potentially perceive them to be white. Whether mere exposure to an image of a white family is sufficient to overpower racial and class stereotypes about who lives in affordable housing is an empirical question inviting deeper study. Regardless, the estimated effects in my analyses could represent a combination of factors beyond the effect of perceiving. Aware of this limitation as I designed my survey, I chose to include a white family in the control condition, as showing no family at all would have caused the pamphlets to be even more nonidentical. It is possible that the image of a family reminds respondents of a more favorably viewed beneficiary from development and increases support for the project.

Finally, my measure for racial sensitivity collapses two dimensions of racial attitudes, limiting interpretability. White opinion towards minority groups has both affective components, measured by symbolic racism, racial resentment, and other scales, as well as a positive component, conceptualized by Jennifer Chudy as racial sympathy. These two dimensions are likely closely correlated, with whites who feel greater motivation to address racial injustice also tending to reject harmful stereotypes and other negative feelings toward minorities. Therefore, I can, with some confidence, interpret the moderating effects of racial sensitivity as suggesting differences between racially prejudiced and sympathetic individuals. However, including a measure of racial animus would have provided an opportunity to make a more robust argument about the differential effects of racial cues on the formation of policy preferences on account of racial attitudes.

Discussion and Policy Implications

The Centrality of Race in the Housing Debate

My findings suggest that the debate over affordable housing operates within a racialized context. Priming respondents to perceive that the beneficiaries of an affordable development will be Black affects their level of support for the project differently depending on their racial attitudes. Individuals who value appearing racially sensitive become more supportive of a project when they perceive that its beneficiaries will be Black. In contrast, less racially sensitive participants increase their opposition in response to racial cues. However, while making Black people salient in evaluations of an affordable project can motivate opposition among prejudiced whites, it does not change the concerns articulated to explain that opposition. Racially motivated respondents are just as likely as their counterparts in a non-racialized context to cite concerns about neighborhood racial demographics—which is to say, not very much. The bias to remain socially desirable encourages respondents to avoid expressing racially prejudiced views.

Although this study uses a national sample of white homeowners, its findings are particularly relevant to California and the Bay Area, where high housing costs, progressive political identity, and the legacy of exclusionary zoning create a uniquely charged political environment around new development. Homeowners in the Bay Area may be more likely than those elsewhere in the United States to be aware of the racial implications of housing policy and its rhetoric, making them especially attuned to how they frame their opposition. Furthermore, the multiracial makeup of San Francisco and surrounding communities complicates the interpretability of these findings, as the sentiments of homeowners from various racial and ethnic groups toward a diverse set of potential beneficiaries are relevant to understanding the nature of housing opposition. For example, Asian Americans comprise a plurality of residents on much of the city’s west side, and approximately 40 percent of homeowners in San Francisco are Asian (U.S. Census 2024). The relationship between Asian Americans’ racial attitudes and support for affordable housing is incredibly consequential for understanding housing politics in San Francisco. And with only about 6 percent of San Franciscans identifying as Black, homeowners may be more likely to perceive other racial groups, like Latinos, as the potential beneficiaries of affordable housing.

Additionally, the complex breakdown of political ideology across racial groups in San Francisco further complicates how racial attitudes interact with housing politics. While majority-white precincts in San Francisco tend to be highly progressive, Asian-majority precincts tend to be significantly more moderate. The median Progressive Voter Index (PVI) score among majority-white precincts is 66 out of 100, with 100 being the most politically progressive, while the median PVI score in majority-Asian precincts is 32 (Sumida 2024). These ideological differences suggest that racial attitudes toward housing development in San Francisco may not follow the same patterns identified in this study of white homeowners. Homeowners in different racial and political communities may respond to racialized housing debates in distinct ways, shaping both support and opposition to new development. And while racial attitudes shape housing support nationwide, in a context like San Francisco—where racial and economic inequality are highly visible, and the political culture may often call to progressive and inclusive values—homeowners may be even more motivated to conceal racial opposition behind ostensibly race-neutral concerns.

While exposure to racial cues does not change respondents’ stated concerns with development, the effect of race in shaping positive expectations of a project’s impacts is unclear. Although public comment frequently features arguments against development, a growing, grassroots “Yes in My Back Yard” (YIMBY) movement seeks to popularize arguments in support, including in San Francisco (Dougherty 2020). During approval meetings, activists and supportive community members increasingly deliver comments highlighting the potential benefits of development. These arguments are numerous: chiefly, growth in the housing stock and decreased housing costs; increased economic activity and tax revenue; lowered net carbon emissions and preserved greenspace; and the advancement of equity goals, for example. Future research should investigate how agreement with these supportive arguments is changed when affordable housing is framed as particularly benefitting Black Americans or other racial minority groups, and whether racial attitudes also moderate this effect. Perhaps racially prejudiced individuals respond to racial cues by closing themselves off from accepting that there may be benefits to development.

Policy Implications

The arguments individuals raise with development are not necessarily honest reflections of the concerns that motivate their opposition. This survey experiment demonstrates that even when the perceived beneficiaries’ race is the only differentiating factor between conditions, respondents do not change their stated reasons for why they might be concerned with development. That is, racially motivated opponents nevertheless mask their reasoning behind other concerns. 

Planning commissions and other local decision-making bodies rely on public comment to inform them of their community’s preferences as they evaluate whether to approve a project. Opposition undermines the accuracy and validity of public comments when racial motivations go unreflected in the arguments opponents make. This misinformation risks several troubling consequences. First, officials and developers may waste time and resources attempting to appease concerns not at the root of residents’ issues. Second, reliance on public comment as the sole source of the community’s preferences may produce systemic biases in approval decisions. Leaders concerned with advancing equity goals, particularly in ameliorating the legacy of residential segregation, may do well to reconsider how they solicit public opinion. Reforms aimed at collecting a more honest and objective measure of residents’ views and concerns may better inform public officials as they seek to mitigate adverse impacts on a neighborhood while advancing solutions to the housing crisis. Perhaps streamlining the development process by replacing adversarial hearings with objective design standards and review processes would do well to eliminate the potential for racial bias altogether.

Policymakers may also reconsider when to solicit public opinion. Rather than allowing public input at the final approval stage for individual housing developments—where mobilized opposition can delay or derail projects—policymakers may consider shifting public engagement earlier in the process, such as during the development of zoning regulations and design standards. Moving debates away from specific projects and toward broader housing policies can help address community concerns proactively while reducing the risk that individual developments become flashpoints for obstruction.

In San Francisco, for example, city leaders may consider expanding ministerial approval for a greater set of developments. While state laws like SB 423 have already streamlined approvals for some projects, they include many requirements that can drive up costs and limit housing typologies, including labor standards and affordability requirements. Expanding ministerial approval for more housing types would limit the ability of racially motivated opposition to delay projects under the guise of neutral concerns. City leaders may also consider reforming the public input process to reduce bias. Live public comment can disproportionately amplify the voices of long-time homeowners while excluding lower-income residents, renters, and younger populations. Replacing or supplementing traditional public hearings with structured surveys, representative panels, or deliberative polling could provide a more accurate and inclusive picture of community sentiment.

These reforms would not eliminate opposition to development but would reduce the extent to which racial bias—explicit or concealed—distorts San Francisco’s housing approval process. A more equitable approach to public engagement could shift the balance of power away from exclusionary interests and toward solutions that meaningfully address the city’s housing crisis.

*This article was edited by Mera Cronbaugh (Princeton University) and Conway Reinders (Princeton University). 

About the Author 

Jose Luis Gandara is a Master of Public Policy student at Stanford University, where he also earned his BA in Public Policy with honors. His research focuses on housing policy, public opinion, and racial attitudes, including published work on multiracial democracy co-authored with Dr. Hakeem Jefferson. Jose has worked on housing policy at the local, state, and federal levels, including with U.S. Senator Maria Cantwell and State Senator Scott Wiener. He also has experience with a San Jose-based nonprofit affordable housing developer. Jose is passionate about advancing housing reforms that increase opportunity and reduce inequality, and he looks forward to engaging in impactful work in California as his program nears completion.

Return to Top

AI Statement

I used ChatGPT to ideate and generate feedback throughout this project. However, I did not use any generative AI to draft final copy of my writing, or to write the code used to analyze my data.

February 24, 2025
Jose Luis Gandara

References

California Department of Housing and Community Development. 2022. “A Home for Every Californian: 2022 Statewide Housing Plan,” March. https://storymaps.arcgis.com/stories/94729ab1648d43b1811c1698a748c136.

Allport, Gordon W. 1954. The Nature of Prejudice. Addison-Wesley Publishing Company.

Blumer, Herbert. 1958. “Race Prejudice as a Sense of Group Position.” Pacific Sociological Review 1 (1): 3–7. https://doi.org/10.2307/1388607.

Bobo, Lawrence. 1983. “Whites’ Opposition to Busing: Symbolic Racism or Realistic Group Conflict?” Journal of Personality and Social Psychology 45 (6), December, 1196. https://doi.org/10.1037/0022-3514.45.6.1196.

Bobo, Lawrence, James R. Kluegel, and Ryan A. Smith. 1997. “Laissez-Faire Racism: The Crystallization of a Kinder, Gentler, Antiblack Ideology.” In Racial Attitudes in the 1990s: Continuity and Change, edited by Steven A. Tuch and Jack K. Martin. Praeger. 15–42. 

Bobo, Lawrence, and Camille L. Zubrinsky. 1996. “Attitudes on Residential Integration: Perceived Status Differences, Mere in-Group Preference, or Racial Prejudice?” Social Forces 74 (3), March, 883–909. https://doi.org/10.2307/2580385.

Burinskiy, Evgeny, Richard Green, and Lois Takahashi. 2021. “The Biggest and Smallest Barriers to California Housing Development.” USC Lusk Center for Real Estate and USC Price School of Public Policy, February. https://lusk.usc.edu/sites/default/files/attachments/USC%20Price%20CHF%20Final%20Report%2003.01.21.pdf.

California Housing Partnership. 2023. “California Affordable Housing Needs Report 2023.” March. https://chpc.net/wp-content/uploads/2023/03/HNR_CA_CHPC-Master2023-FINAL.pdf.

Chudy, Jennifer. 2021. “Racial Sympathy and Its Political Consequences.” The Journal of Politics 83 (1), January, 122-136. https://doi.org/10.1086/708953.

Colton, Kent, and Gopal Ahluwalia. 2019. “A Home Builder Perspective on Housing Affordability and Construction Innovation.” Joint Center for Housing Studies, Harvard University, July. https://www.jchs.harvard.edu/sites/default/files/harvard-jchs-home-builder-perspective-on-housing-affordability-colton-2019.pdf.

California Department of Finance. n.d. “Construction Permits.” Accessed May 24, 2024. https://dof.ca.gov/forecasting/economics/economic-indicators/construction-permits/.

Coppock, Alexander, and Oliver A. McClellan. 2019. “Validating the Demographic, Political, Psychological, and Experimental Results Obtained from a New Source of Online Survey Respondents.” Research & Politics 6 (1). https://doi.org/10.1177/2053168018822174.

Crandall, Christian S., and Amy Eshleman. 2003. “A Justification-Suppression Model of the Expression and Experience of Prejudice.” Psychological Bulletin 129 (3), May, 414. https://doi.org/10.1037/0033-2909.129.3.414.

Dougherty, Conor. 2020. “California’s Housing Crisis: How a Bureaucrat Pushed to Build - The New York Times.” New York Times. February 13. https://www.nytimes.com/2020/02/13/business/economy/housing-crisis-conor-dougherty-golden-gates.html?searchResultPosition=2.

Einstein, Katherine Levine, David M. Glick, and Maxwell Palmer. 2020. “Neighborhood Defenders: Participatory Politics and America’s Housing Crisis.” Political Science Quarterly 135 (2): 281–312. https://doi.org/10.1002/polq.13035.

Enos, Ryan D. 2014. “Causal Effect of Intergroup Contact on Exclusionary Attitudes.” Proceedings of the National Academy of Sciences 111 (10), February, 3699–3704. https://doi.org/10.1073/pnas.1317670111.

Erdmann, Kevin. 2023. “Home Price Trends Point to a Worsening Lack of Supply.” Mercatus Center. May 30. https://www.mercatus.org/research/research-papers/home-price-trends-point-worsening-lack-supply.

Gilens, Martin. 2009. Why Americans Hate Welfare: Race, Media, and the Politics of Antipoverty Policy. University of Chicago Press.

Goetz, Edward G. 2008. “Words Matter: The Importance of Issue Framing and the Case of Affordable Housing.” Journal of the American Planning Association 74 (2): 222–29. https://doi.org/10.1080/01944360802010251.

Hernandez, Jennifer, and Robert Apodaca. 2023. “How CEQA Made UC Berkeley Students an Environmental Threat.” Los Angeles Times. March 2. https://www.latimes.com/opinion/story/2023-03-02/california-uc-berkeley-ceqa-housing-environment.

Bay Area Equity Atlas. n.d. “Housing Burden.” Accessed May 24, 2024. https://bayareaequityatlas.org/indicators/housing-burden?breakdown=by-race-ethnicity.

Hurwitz, Jon, and Mark Peffley. 1997. “Public Perceptions of Race and Crime: The Role of Racial Stereotypes.” American Journal of Political Science 41 (2), April, 375–401. https://doi.org/10.2307/2111769.

———. 2005. “Explaining the Great Racial Divide: Perceptions of Fairness in the US Criminal Justice System.” The Journal of Politics 67 (3), August, 762–83. https://doi.org/10.1111/j.1468-2508.2005.00338.x

Hutchings, Vincent L., and Nicholas A. Valentino. 2004. “The Centrality of Race in American Politics.” Annual Review of Political Science 7, June, 383–408. https://doi.org/10.1146/annurev.polisci.7.012003.104859.

Jackman, Mary R., and Michael J. Muha. 1984. “Education and Intergroup Attitudes: Moral Enlightenment, Superficial Democratic Commitment, or Ideological Refinement?” American Sociological Review 6, 751–69. https://doi.org/10.2307/2095528.

Karlinsky, Sarah. 2023. “Losing Ground: What the Bay Area’s Housing Crisis Means for Middle-Income Households and Racial Inequality.” SPUR. March. https://www.spur.org/sites/default/files/2023-03/SPUR_Losing_Ground_0.pdf.

Karlinsky, Sarah, and Krisy Wang. 2021. “What It Will Really Take to Create an Affordable Bay Area.” SPUR. April. https://www.spur.org/sites/default/files/2021-05/SPUR_What_It_Will_Really_Take_To_Create_An_Affordable_Bay_Area_Report.pdf.

Kinder, Donald R., and Lynn M. Sanders. 1996. Divided by Color: Racial Politics and Democratic Ideals. University of Chicago Press.

Kinder, Donald R., and David O. Sears. 1981. “Prejudice and Politics: Symbolic Racism versus Racial Threats to the Good Life.” Journal of Personality and Social Psychology 40 (3): 414–431. https://doi.org/10.1037/0022-3514.40.3.414.

Krysan, Maria. 1998. “Privacy and the Expression of White Racial Attitudes: A Comparison across Three Contexts.” Public Opinion Quarterly 62 (4), February, 506–44. https://academic.oup.com/poq/article/62/4/506/1940482.

Landes, Emily. 2022. “SF Has the Longest Entitlement, Permitting Approval Timeline in the State.” The Real Deal. June 14. https://therealdeal.com/sanfrancisco/2022/06/14/sf-housing-bottleneck-resi-entitlements-permits-take-longest-in-state/.

Manji, Shazia, and Ryan Finnigan. 2023. “Streamlining Multifamily Housing Production in California: Progress Implementing SB 35.” Terner Center for Housing Innovation. August. https://ternercenter.berkeley.edu/research-and-policy/sb-35-evaluation/.

McConahay, John B. 1986. “Modern Racism, Ambivalence, and the Modern Racism Scale.” In Prejudice, discrimination, and racism, edited by John F. Dovidio & Samuel L. Gaertner. Academic Press, 91-125.

McGhee, Eric, Jennifer Paluch, and Vicki Hsieh. 2021. “New Housing Fails to Make Up for Decades of Undersupply.” Public Policy Institute of California. December 3. https://www.ppic.org/blog/new-housing-fails-to-make-up-for-decades-of-undersupply/.

Nederhof, Anton J. 1985. “Methods of Coping with Social Desirability Bias: A Review.” European Journal of Social Psychology 15 (3): 263–80. https://doi.org/10.1002/ejsp.2420150303.

Peffley, Mark, Jon Hurwitz, and Paul M. Sniderman. 1997. “Racial Stereotypes and Whites’ Political Views of Blacks in the Context of Welfare and Crime.” American Journal of Political Science 41 (1), January, 30–60. https://www.jstor.org/stable/2111708?seq=1.

Perigo, Sasha. 2020. “Who Are the San Francisco-Bay Area’s NIMBYs—and What Do They Want?” Curbed San Francisco. February 20. https://sf.curbed.com/2020/2/20/21122662/san-francisco-bay-area-nimbys-history-nimby-development.

Quillian, Lincoln. 2006. “New Approaches to Understanding Racial Prejudice and Discrimination.” Annual Review of Sociology 32, 299–328. https://doi.org/10.1146/annurev.soc.32.061604.123132.

Rezal, Adriana. 2022. “Here’s Why Austin and Seattle Are Building Way More Housing than San Francisco.” San Francisco Chronicle. August 1. https://www.sfchronicle.com/sf/article/housing-tech-hub-building-17339487.php.

Sahn, Alexander. 2024. “Public Comment and Public Policy.” American Journal of Political Science, 1–16. https://doi.org/10.1111/ajps.12900.

California Department of Housing and Community Development, Housing Policy Development Division. 2023. “San Francisco Housing Policy and Practice Review.” October. https://www.hcd.ca.gov/sites/default/files/docs/policy-and-research/plan-report/sf-housing-policy-and-practice-review.pdf.

Scally, Corianne Payton, and J. Rosie Tighe. 2015. “Democracy in Action?: NIMBY as Impediment to Equitable Affordable Housing Siting.” Housing Studies 30 (5), May, 749–69. http://dx.doi.org/10.4324/9780429299377-30.

Schuman, Howard, Charlotte Steeh, Lawrence Bobo, and Maria Krysan. 1997. Racial Attitudes in America: Trends and Interpretations. Harvard University Press.

Sidanius, Jim, and Felicia Pratto. 2001. Social Dominance: An Intergroup Theory of Social Hierarchy and Oppression. Cambridge University Press.

Smith, Tom W. 1991. “Ethnic Images.” GSS Topical Report No. 19. National Opinion Research Center University of Chicago. https://gss.norc.org/Documents/reports/topical-reports/TR19.pdf.

Sniderman, Paul M., and Edward G. Carmines. 1997. “Reaching beyond Race.” PS: Political Science & Politics 30 (3), September, 466–71. https://doi.org/10.2307/4201.

Sniderman, Paul M., Gretchen C. Crosby, and William G. Howell. 2000. “The Politics of Race.” In Racialized Politics: The Debate about Racism in America, edited by David O. Sears, Jim Sidanius, and Lawrence Bobo. University of Chicago Press, 236–79.

Tighe, Jenna Lee. 2009. Public Perceptions of Affordable Housing: How Race and Class Stereotyping Influence Views. The University of Texas at Austin. https://repositories.lib.utexas.edu/handle/2152/10637#:~:text=The%20focus%20group%20and%20survey,determinants%20of%20affordable%20housing%20opposition.

Tolbert, Caroline J., and John A. Grummel. 2003. “Revisiting the Racial Threat Hypothesis: White Voter Support for California’s Proposition 209.” State Politics & Policy Quarterly, 183–202. https://www.jstor.org/stable/40421487.

Appendix

Appendix A: T-Tests Comparing Differences in Means

To initially test whether exposure to racial cues affects respondents’ attitudes toward affordable housing, I perform t-tests comparing differences in means between conditions on support for the development and agreement with various concerns. A two-sample t-test is used to determine whether the means of two populations are equal, with the null hypothesis being that the difference in group means is zero. It is defined as:

where N₁ and N₂ are the sample sizes, Y₁ and Y₂ are the sample means, and s₁² and s₂² are the sample variances. At level of significance, one can reject the null hypothesis that the two means are equal if:

where t_1-α/2,v is the critical value of the t-distribution with v degrees of freedom.

I do not expect exposure to the treatment to create a statistically significant difference in means across the two conditions for either outcome variables of interest. Because racial attitudes will moderate the effect of a racial prime on attitudes toward affordable development, receiving treatment will increase some respondents’ support while decreasing it for others. I do not expect t-tests to reject the null hypothesis for either outcome variable.

Appendix B: Additional Tables

Table B.1: Comparison of demographic variables across conditions, with t-test

Table B.2: Effects of condition and racial attitudes on support for development, without ideology control

Appendix C: Survey Questions

In this study, you will be asked your views on some political issues and some questions about yourself.

What is your race? Check all that apply.

1. White/Caucasian 
2. Black/African American 
3. Hispanic/Latino 
4. Asian 
5. Native American 
6. Other

Do you own your home?

1. Yes
2. No

What is your age?

What is your gender?

1. Male
2. Female
3. Neither of these apply
4. I prefer not to say

Are you Hispanic or Latino?

1. Yes
2. No

For statistical purposes only, we have a question about your income. Last year, what was your total family income from all sources, before taxes?

1. Less than $15,000
2. $15,000 - $24,999
3. $25,000 - $34,999 
4. $35,000 - $49,999 
5. $50,000 – $74,999 
6. $75,000 – $99,999 
7. $100,000 - $119,999 
8. $120,000 - $149,999 
9. $150,000 - $199,999
10. $200,000 or greater

What is the highest level of education you have completed?

1. Less than high school
2. High school/GED
3. Some college
4. 2 year college degree
5. 4 year college degree
6. Masters degree
7. Doctoral degree
8. Professional degree (JD, MD)

In general, do you consider yourself liberal or conservative? Please answer on the following scale.

1. Extremely liberal
2.  
3.  
4.  
5.  
6.  
7. Extremely conservative

When it comes to SOCIAL policy, do you consider yourself liberal or conservative? Please answer on the following scale.

1. Extremely liberal
2.  
3.  
4.  
5.  
6.  
7. Extremely conservative

When it comes to ECONOMIC policy, do you consider yourself liberal or conservative? Please answer on the following scale.

1. Extremely liberal
2.  
3.  
4.  
5.  
6.  
7. Extremely conservative

Do you typically consider yourself a Democrat, Republican, or something else?

1. Democrat
2. Republican
3. Independent
4. Something else (please specify)
5. No preference
6. Don’t know

Would you call yourself a Democrat, or a not very strong Democrat?

1. Strong Democrat
2. Not very strong Democrat
3. Don’t know

Would you call yourself a Republican, or a not very strong Republican?

1. Strong Republican
2. Not very strong Republican
3. Don’t know

Do you think of yourself as closer to the Republican or Democratic Party?

1. Closer to Republican
2. Neither
3. Closer to Democratic
4. Don’t know

Think of the ladder below as representing where people stand in the United States. At the top of the ladder are the people who are the best off—those with the most money, the most education, and the most respected jobs. At the bottom are the people who are the worst off—with the least money, least education, and the least respected jobs or no job. The higher up you are on this ladder, the closer you are to the people at the very top and the lower you are, the closer you are to the people at the very bottom. Where would you place yourself on this ladder?

1. 10 – Top rung
2. 9
3. 8
4. 7
5. 6
6. 5
7. 4
8. 3
9. 2
10. 1– Bottom rung

Please indicate how strongly you agree or disagree with the following statement: “It is important to me that I be seen by others as a racially sensitive person.”

1. Strongly disagree
2. Disagree
3. Slightly disagree
4. Neither agree nor disagree
5. Slightly agree
6. Agree
7. Strongly agree

For statistical purposes only, we have a question about your zip code. What is your zip code?

What type of residence do you live in?

1. Single-family home (a standalone residential structure designed to house one household)
2. Duplex, triplex, or similar unit in a residential structure designed to house a few households
3. Apartment in a multi-family building
4. Condominium
5. Townhouse
6. Other (please specify): _____

Approximately what percentage of homes in your neighborhood are single-family homes? A single-family home is standalone residential structure designed to house one household.

1. 0-24%
2. 25-49%
3. 50-74%
4. 75-100%

Would you describe most households in your neighborhood as:

1. Low-income
2. Middle-income
3. High-income

How racially diverse would you say your neighborhood is?

4. Not at all
5. Slightly
6. Somewhat
7. Very
8. Extremely

Which of the following best describes the area you currently live in?

1. Large city
2. Suburb near a large city
3. Small city or town
4. Rural

Thank you for sharing about yourself and your neighborhood. Now, we are going to ask you to think about your metro region more broadly. 
 
When we mention the term "metro region," we're referring to the broader area that surrounds and includes your city or town. This isn't just limited to the city center or downtown area but extends to the suburbs, smaller towns, and communities that are closely connected to the main urban area through commuting patterns, economic ties, and shared services.

Now, please think about the availability of affordable housing in your metro region. Affordable housing refers to housing units that low- and moderate-income households can pay for while still having money left over for other essential living expenses. Typically, housing is considered affordable if rent and utilities cost no more than 30 percent of a household's monthly income.

Please indicate your level of agreement with the following statement: “There is a lack of affordable housing in my metro region.”

1. Strongly disagree
2. Disagree
3. Slightly disagree
4. Neither agree nor disagree
5. Slightly agree
6. Agree
7. Strongly agree

Please read the following scenario carefully. You will be asked questions about it later in the survey. This scenario is hypothetical. However, try your best to think about how you would respond if it was real.

Imagine that you have received a pamphlet from a developer who is seeking to construct a multi-family affordable housing development in your neighborhood. The proposed project site is located half a mile away from your home. Before the developer can begin construction, they must obtain permits from local government officials who can approve or reject the project. This decision will be made at an upcoming public meeting where community members like you are invited to share your concerns. Your feedback is valuable and can affect how local officials view the project.

Next, we will show you a sample pamphlet from the developer of the proposed project. Please review it carefully. You will be asked questions about it later in the survey.

What services will be offered in the building?

1. Childcare
2. Job training
3. Medical clinic
4. Senior support services

Please try to answer the following questions thoughtfully and honestly to reflect how you would act if this scenario were real.

Would you oppose or support the proposed affordable housing development in your neighborhood?

1. Strongly oppose
2. Oppose
3. Slightly oppose
4. Neither support nor oppose
5. Slightly support
6. Support
7. Strongly support

If the development was proposed in your neighborhood, how likely would you be to provide public comment at the local government meeting where it will be discussed? Your public comment may influence how public officials view the project.

1. Very likely
2. Somewhat likely
3. Not very likely
4. Not at all likely

When it comes to affordable housing, people may call to mind many different considerations. In evaluating the proposed development you read about, how much did each of the following potential consequences come to mind? Please respond with how much you agree with each of the following statements.

“I am concerned about increased traffic and/or competition for street parking in my neighborhood.”

1. Strongly disagree
2. Disagree
3. Slightly disagree
4. Neither agree nor disagree
5. Slightly agree
6. Agree
7. Strongly agree

"I am concerned about the development’s design or aesthetics fitting in with my neighborhood."

1. Strongly disagree
2. Disagree
3. Slightly disagree
4. Neither agree nor disagree
5. Slightly agree
6. Agree
7. Strongly agree

"I am concerned about the strain on local resources like schools in my neighborhood."

1. Strongly disagree
2. Disagree
3. Slightly disagree
4. Neither agree nor disagree
5. Slightly agree
6. Agree
7. Strongly agree

"I am concerned about decreased property values in my neighborhood."

1. Strongly disagree
2. Disagree
3. Slightly disagree
4. Neither agree nor disagree
5. Slightly agree
6. Agree
7. Strongly agree

"I am concerned about changing the racial composition of my neighborhood."

1. Strongly disagree
2. Disagree
3. Slightly disagree
4. Neither agree nor disagree
5. Slightly agree
6. Agree
7. Strongly agree

"I am concerned about increased crime in my neighborhood."

1. Strongly disagree
2. Disagree
3. Slightly disagree
4. Neither agree nor disagree
5. Slightly agree
6. Agree
7. Strongly agree

In your own words, please describe why you would support or oppose this project.