Bringing the Person Back In:Boundaries, Misperceptions, and the Measurement of

Racial Context

Cara Wong∗ Jake Bowers Tarah Williams Katherine Drake

April 30, 2011

Abstract

If “All politics is local,” as Tip O’Neill famously stated, then studying politicsrequires studying place. Yet place, as defined in many studies of “context effects”throughout the social sciences, is often so vague as to hinder the development ofunderstandings about how place and politics interact. In this paper, we borrowfrom Parsons and Shils to offer a formal conceptualization of “context,” for thepurposes of using “context” to learn more about individual level political outcomes.Our conceptualization, and a recognition of the statistical Modifiable Areal UnitProblem, lead us to a new measurement strategy: We propose a map-based measureto capture how ordinary people use information about their environments to makedecisions about politics. Respondents draw their contexts on maps — deciding theboundaries and limits of the relevant environment — and describe their perceptionsof the demographic make-up of these contexts. The evidence is clear: people’spseudoenvironments do not resemble governmental administrative units in shapeor content. By “bringing the person back in” to the measurement of context, we areable to marry psychological theories of information processing with sociologicaltheories of racial threat. And, our measure of context allows us to sidestep theModifiable Areal Unit Problem, a major stumbling block in research on contextthat prevents scholars from knowing whether they have substantive findings orsimply statistical artifacts.

∗Assistant Professor, Dept of Political Science, University of Illinois @ Urbana-Champaign(carawong@illinois.edu). Paper prepared for the 2011 meeting of the Chicago Area BehaviorWorkshop. Acknowledgements: We thank Andrea Benjamin, Alana Hackshaw, and Laura Potter for theirexcellent research assistance on the survey, and to the Interdisciplinary Institute at the University ofMichigan for funding the pilot study. Thanks for excellent suggestions, critiques, and comments fromcolleagues at Illinois, Michigan, and Harvard, and many anonymous reviewers for the political science,geography, and sociology panels of the National Science Foundation.

1

The places where people live can affect them in a myriad of ways because context can beboth objective and subjective, and the outcomes can be physical or psychological (Matei,Ball-Rokeach and Qiu, 2001; Sampson, Raudenbush and Earls, 1997; Sampson, Morenoffand Gannon-Rowley, 2002; Shinn and Toohey, 2003; Gans, 1962; Anderson, 1976; Wilson,1987; Jacobs, 1961).1 For example, in the case of pollution, environmental allergens inthe air breathed by individuals have a direct physical effect on their lungs and health(Weinmayr, 2007; Custovic et al., 1998; Braun-Fahrlander, 2002). This “context effect”requires no knowledge on the part of its residents about the environment or its contents;one does not have to see and recognize smog to be affected by it. Of course, if individualsare aware of the pollution, then they can take actions to avoid or counteract the negativehealth outcomes. Similarly, living in a poor area can minimize people’s ability to engagein politics due to limited access to transportation, information sources, and politicians’attention, all without entering into these residents’ explicit cost-benefit analyses aboutwhether to participate in politics; they may not realize political resources are missingfrom their locales, but the contexts in which they live can still have an impact on theirbehavior.

People’s surroundings can also have less direct effects, particularly when the outcomesare psychological in nature. For many important political phenomena, people mustbe cognizant of their environment for it to have an impact. Living in a poor area, forexample, can lead individuals to feel threatened, not just to vote less often due toscarce transportation. A preponderance in one’s community of high school dropouts,individuals living below the poverty line, and broken windows can lead to fears ofcrime or dropping property values, along with a greater sense of alienation and politicalinefficacy (Sampson, Morenoff and Gannon-Rowley, 2002). Similarly, living in an areawith many racial outgroup members can lead to fear for one’s personal safety, job,identity, or political voice (Blalock, 1967; Taylor, 1998, 2000; Stein, Post and Rinden,2000). These perceived threats are the main concern of political scientists focused onracial context.2 However, one cannot argue that these environmental characteristics

1In this paper, we set as aside explicit considerations of personal networks of acquaintances and focus onthe perceptual mechanisms of impersonal interactions only, although both are types of context (Huckfeldt,1979; Burbank, 1997; Putnam, 1966). We do discuss how some measures of impersonal interactions may,in fact, reflect interpersonal mechanisms.

2Empirical studies have shown that places where blacks are a larger proportion of the population havegreater white voter registration, voting for racist candidates, and opposition to black leaders such as JesseJackson (Giles and Buckner, 1993; Heer, 1959; Key, 1949; Prysby, 1989; Voss, 1996; Wright Jr, 1977; Gilesand Buckner, 1993). Racial context and the feelings of threat that it engenders are used to explain whiteracial attitudes and prejudice (Bobo and Hutchings, 1996; Cain, Citrin and Wong, 2000; Fossett and Kiecolt,1989; Giles, 1977; Glaser, 1994; Huckfeldt and Sprague, 1993; Kinder and Mendelberg, 1995; Kinder andSanders, 1996; Oliver and Mendelberg, 2000; Oliver, 2010; Quillian, 1996), political intolerance (Shamirand Sullivan, 1983; Feldman and Stenner, 1994), social capital (Alesina and La Ferrara, 2000), group and

1

lead to feelings of threat without someone — elites, if not also ordinary residents —being conscious and cognizant of the context (Lazarsfeld, Berelson and Gaudet, 1968).After all, fear is an emotional response to a perceived threat, and the brain acts as thefilter for recognizing outgroup members as “pollutants” or threats in the environment.

Just as the outcomes of context can be both physical and psychological (e.g., asthma andattitudes), context itself can be objective — defined by bodies of water, walls and gates,and lines on maps — as well as subjective (Wong, 2010). Political science measurementof racial context tends to emphasize the former, which has social science origins datingback more than half a century. When Lewin wrote that behavior is a function of theperson and the situations involved, the “situation” was treated as a fixed container ofsorts (Lewin, 1936) (See also (Tingsten, 1937)). When we conceptualize the environmentas a container (of recognizable and commonly shared dimensions), then it is sufficientto pick the size of the container, measure its contents, and test for any relationshipsbetween said content and our outcome of interest, whether that is a propensity todevelop asthma, interact with outgroup members, or be a robbery victim (but see(Lefebvre, 1991)). The researcher will still need to justify the choice of the containersize for statistical and theoretical reasons discussed later. However, if “context” isconceptualized as a container, it is unnecessary to ask residents if they are aware of thecontainer and its contents.

However, while one’s environment may serve as a physical container within whichindividuals exist and interact, context can also refer to the environment in whichpeople believe they live; Lippmann’s “pseudo-environments” are the reality with whichpeople engage, regardless of how accurate or distorted these perceptions (Lippmann,1991). While it is a truism in social science research that race is a social construct,we want to emphasize that people’s racial contexts are also social and psychologicalconstructions. Researchers have learned much from studies of phenomena like whiteflight, tipping points, and racial discrimination in housing, home mortgages, car sales,insurance provision, and medical care, all of which emphasize the importance of howindividuals perceive, construct, and evaluate the race and ethnicity of those they observeand with whom they interact (Yinger, 1997; Turner and Skidmore, 1999; Ayres and

partisan identity among whites (Giles and Evans, 1985; Giles and Hertz, 1994), provision of public goods(Alesina, Baqir and Easterly, 1999), roll call votes in Congress for liberal legislation (Jacobs and Tope, 2007),attitudes toward immigration (Cain, Citrin and Wong, 2000; Bobo and Hutchings, 1994), voting on ballotinitiatives (Campbell, Wong and Citrin, 2006; Glaser, 2002), the adoption of state felon disenfranchisementprovisions (Behrens, Uggen and Manza, 2003), and the occurrence of lynchings and intergroup conflictbetween whites and blacks (Olzak, 1994; Olzak, Shanahan and McEneaney, 1996; Tolnay, Beck and Massey,1988; Tolnay and Beck, 1995). The perception of racial threat also has been used to explain black intoleranceof white racists and political mobilization (Bowers, 1997; Green and Waxman, 1987). Scholars have alsofocused on the effects of race in conjunction with those of the economy, fiscal spending, partisanship,religion, education, employment rates, and family income on individual attitudes and behaviors (see, forexamples, (Campbell, Wong and Citrin, 2006; Campbell, 2008; Oliver and Mendelberg, 2000; Quillian,1995; Branton and Jones, 2005; Cho, Darmofal and Baer, 2007; Taylor, 1998)).

2

Siegelman, 1995; Wissoker, 1997; Schulman et al., 1999; Ridley, Bayton and Outtz, 1989).3

Recognition of these multiple types of outcomes and perceptions of context is importantbecause it highlights the different mechanisms by which context can affect politics —adding frustration aggression displacement theory to realistic group conflict theory,for example, by emphasizing how “perceived” shortfalls can matter as much as, if notmore than, “objective” deprivations (LeVine and Campbell, 1971).

In this article, first we work to conceptualize “context” as a pseudoenvironment, notjust a container. Then, we present a new measurement tool that brings the individualback into the study of racial threat. Our new measure of context helps us answerboth questions of measurement — what kind of geographic units ought to be mostrelevant for individual attitudes and behavior? — as well as questions of mechanism— how does place get into the heads of people and how do people represent placesin their minds? When we consider the conceptualization of context in the case of apsychological outcome, it is clear that including the individual’s frame of reference isnecessary for valid measurement. The operationalization of context needs to capturethe many different ways that the macro-social demography of a community is linked tothe political psychology of its residents. Therefore, we ask respondents to draw theircontexts on maps — wherever they see the boundaries of communities — and thendescribe the content of these contexts. With our new measures of context, we examinewhether people’s pseudoenvironments reflect the “objective” context across multipleunits of analysis and discuss how far common practice is from our conceptualization ofcontext. We also discuss why levels of misinformation or misperception may vary atdifferent geographies. We are thus able to bring together the public opinion scholarshipon political information with that of racial threat.

Our results help deepen and extend our knowledge about how context matters inpolitics, showing that governmental administrative units do not define the environmentsrelevant to individuals, and that people do not “see” what the Census “sees” in theseenvironments. In other words, Census numbers should not be used as proxies for whatordinary Americans perceive in their contexts. By measuring the environment thatis personally relevant to the individual, we can also help researchers avoid a serious,intractable methodological problem — the modifiable areal unit problem (MAUP) —that could explain the inconsistency of results in research on racial threat and contact.

3In this paper, we use the following terms interchangeably: “whites” and “non-hispanic whites,”“blacks” and “African-Americans,” and “Latinos” and “Hispanics.”

3

1 Conceptualization of Context

While political scientists have written of the importance to politics of people’s environ-ments for almost a century, “context” is very rarely defined in contemporary articles,perhaps because its ordinary language usage is that of the researchers’. It seems almostsuperfluous to state explicitly that “context” is the place, area, or environment in whichpeople live. Other definitions or synonyms used include the following: “shared livingspace,” “circumstances of their residential spaces,” “social surroundings”, “local politi-cal arenas,” “communities,” and “social milieu.” Agnew provides a longer definition,explaining that “context refers to the hierarchical (and nonhierarchical) ’funnelling’ ofstimuli across geographical scales or levels to produce effects on politics and politicalbehavior” ((Agnew, 1996) 132). However, stopping with these general definitions isdangerous, because one is unable to discern what does not count as context; they servemore as what Adcock and Collier would call the “background concept” rather thanthe “systematized concept” (Adcock and Collier, 2001). After all, according to thesegeneral definitions, the earth and the Milky Way galaxy could count as context, evenif political scientists rarely work on that scale. This lack of specificity leaves scholarswith little guidance as they move to the stage of operationalization, making the dangerof operationalism — equating a concept with its measures — very vivid (Blalock andBlalock, 1968). For example, context simply becomes percent black or ethnic diversity ina county. Explicitly conceptualizing “context” will help both with understanding themechanisms by which contexts affect individuals’ political judgments and the measuresbest suited to capture its core.

The concept of context should convey what unit of analysis might be most appropriate.Fortuitously, there is no need to reinvent the wheel, nor is it necessary to develop anidiosyncratic conceptualization of context. Instead, we can use a classic social sciencedefinition that arose in the era of the localized studies of Lazarsfeld, Berelson, and theircolleagues. In Toward a General Theory of Action, Parsons and Shils explain that there arethree components in the frame of reference of their theory of action: “actors, a situationof action, and the orientation of the actor to that situation” ((Parsons and Shils, 1951),56). The situation is

that part of the external world which means something to the actor whosebehavior is being analyzed. It is only part of the whole realm of objects thatmight be seen. Specifically, it is that part to which the actor is oriented andin which the actor acts (56).

The orientation of the actor to the situation is “the set of cognitions, catheses, plans,and relevant standards which relates the actor to the situation” (56). We believe thatParsons and Shils’s “situation of action” is the context to which scholars refer when they

4

talk about context having an effect on individuals’ political attitudes and actions. The“orientation of the actor to that situation” may include the varied mechanisms by whichcontexts can affect individuals’ politics. Note that this concept encompasses containers;a container is just a situation toward which all people are assumed uniformly oriented,both in terms of the size and boundaries of the contextual unit and its contents.

Using the Parsons and Shils definition does not preclude the instances in which objectivecharacteristics of a fixed container affect individuals’ attitudes and actions. However,given the extensive research in political psychology about the public’s information pro-cessing and levels of political knowledge, the study of context also needs to include thepossibility of addressing the following questions: (1) what are the perceived boundariesof the context? and (2) What is the perceived content of the context? In this article, weallow individuals to define their own relevant contexts using maps, and we discusshow bringing psychological theories of information processing to bear on sociologicaltheories of racial threat fit both our conceptualization and measurement of context.

2 How Big is “Context”?

Political scientists almost always justify their choice of contextual unit on theoreticalgrounds. For example, because individuals rarely live and work in the same census tract,and the effects of racial context in one’s life overall — not just one’s neighborhood —may have political effects, county may be the ideal contextual unit of analysis despite itsextreme heterogeneity (Branton and Jones, 2005). Alternatively, because racial contextmay be understood as the interaction between ingroup and outgroup members, the unitof analysis should be small — like the neighborhood — in order to capture actual contactbetween individuals and significant social relationships. Or, because the outgroup isdefined at the national level (i.e. immigrants), the appropriate context is the country(Quillian, 1995). Some scholars argue that the contextual unit should vary dependingon the type of threat under consideration: a more personal threat is experienced at thelocal neighborhood level, whereas a political threat is experienced at the level of the city.Political elites at the national level may also prime a local context that is, or is becomingmore, diverse (Hopkins, 2010). Scholars also argue that multiple contexts — macro-and micro-environments — may interact and should be considered simultaneously; thisraises the additional question of which of the many possible combinations of contextsto consider (precincts and electoral districts (Liu, 2001) or counties and sections of states(Key, 1949), for example.). As Oliver and Mendelberg explain, “Identifying a context’sboundaries is essential for understanding its potential effects” ((Oliver and Mendelberg,2000), 577) How can we adjudicate between these theoretical arguments to determinethe appropriate unit of analysis?

5

While the different theoretical arguments justifying area size are reasonable primafacie, the primary reason for why a particular unit is chosen in the vast majority ofstudies is that of practical necessity (Glaser, 1994). Because few studies were designedto address questions of racial context, scholars are constrained by the units availablein a given dataset. Most studies, for example, do not provide an easy way to identifyrespondents at more than one geographic level; while the rationale is to protect theanonymity of survey respondents, one by-product is that scholars are often limited toasserting that the contextual unit they use is ideal rather than being able to test andprove it. And unfortunately, the many studies of racial context have not converged to aconsensus about the ideal contextual unit. In their article, Cho and Baer provide anexcellent summary of many examples of the research on racial context; across 34 studies,the geographic units chosen for the articles vary from prison cells to census tracts tocountries (Cho and Baer, 2010). And, even when the same contextual unit of analysis ischosen, Cho and Baer show that the results can support contact theory, threat theory,or even both simultaneously.

2.1 The MAUP

In response to the cacophony of units and results, Cho and Baer recommend that morethan one unit should be used in analyses, partly as a robustness check, but mostly to beaware of the potential conflicting results that can arise from the “modifiable areal unitproblem” (MAUP). The MAUP is a phenomenon well-documented by geographers,whereby relationships between variables at one level can change (and even flip signs)when studied at a different level of aggregation; the areal units chosen are “modifiable”or arbitrary (Bhat and Guo, 2004; Anselin, 1988; Openshaw and Taylor, 1979; Yuleand Kendall, 1950). Cho and Manski describe the MAUP in terms more familiar topolitical scientists as “a problem that is isomorphic to the ecological inference problem”(Cho and Manski (2008), 548) (see also Achen and Shively (1995)). Gotway and Youngargue that the ecological inference problem is a special case of the MAUP, while Kingargues the converse (Gotway and Young, 2002; King, 1997). The MAUP is actuallycomposed of 2 problems concerning scale and aggregation. Gehlke and Biehl (1934)found that correlation coefficients vary with the size of the unit under study — withlarger units leading to larger coefficients, even when there is no greater correlation atthe individual level across units — showing that scale can affect statistical inference(Gehlke and Biehl, 1934).4 Openshaw and Taylor (1979) looked at the relationshipbetween the percentage of Republican voters and the percentage of elderly voters inIowa; they compared the correlations in the state’s 99 counties as well as in all possible

4They showed these patterns hold whether one was studying relationships between juvenile delin-quency and monthly rental costs, or between farm product values and numbers of farmers.

6

combinations of these counties into larger districts. They were able to find correlationsranging from -.97 to +.99, with no clear pattern between the spacial characteristicsof the districts and the variation in the coefficients (Openshaw and Taylor, 1979). Inother words, results can vary depending on how units are aggregated or zoned, evenwhen they are the same scale or size. The effects of the MAUP are less severe if theaggregation of units is done in a noncontiguous or spatially random way. However,given that tracts are composed of contiguous block groups, counties are composed ofcontiguous tracts, and so on, the aggregation effect of the MAUP may be even moresevere in political science research on racial context. Further research has shown thatthe MAUP affects multivariate regression, Poisson regression, multilevel models, spatialinteraction models, and spatial autocorrelation statistics, along with simpler statisticslike the mean, variance and correlation coefficient (Gotway and Young, 2002). As King(1996) notes, “The conclusions of these studies are almost always extreme pessimismthat leads the authors to question the vercity of almost all empirical analyses (163).”Scholars have proposed complex statistical ways for addressing the MAUP, but each hasits own limitations, set of assumptions, and critics (King, 1997); the simple statisticalstrategy commonly proposed is to undertake analysis at multiple scales or zones. The“real fix” is to know the correct model linking individuals to aggregates.

Not many political scientists have looked at contextual effects at multiple levels, andalmost all explain discrepancies across levels in substantive terms, rather than noting theMAUP (Oliver and Mendelberg, 2000; Cho and Baer, 2010; Baybeck, 2006; Mackuen andBrown, 1987). For example, an argument could be made that racial diversity only has athreatening effect at the metropolitan level and not at the neighborhood level becausethreat is only experienced at the larger aggregate levels (Oliver, 2010). When scholarsdo test their findings across levels, it is often a confirmatory exercise that the resultshold across levels (See Hopkins (2007), for example). Such confirmations are excellentfor assessing the robustness of results, but there is little discussion in the literatureas to why effects might be consistent across some levels, if not all.5 An equally goodexplanation for differences in coefficients across scales is the MAUP. We have no way ofknowing whether these differences arise because of the explained substantive rationalesor whether the differences are statistical artifacts of the MAUP with no substantiveimport. And, even if taking into account the effects of nearest neighboring contextualunits shows the robustness of one’s results to different such choices — as advocatedby Cho and Baer — there is still no way to adjudicate between one scholar’s argumentthat census tract is the best unit of analysis and another’s that city is superior if bothconclusions are fairly robust to specifications that recognize potential influences of

5Baybeck is an exception; he looks at the negative consequences of incongruence in effects across context,comparing respondents from homogeneous neighborhoods within diverse cities with respondents fromdiverse neighborhoods in diverse city (Baybeck, 2006).

7

neighboring tracts or cities, respectively. Of course, for certain outcomes, it is easierto justify the choice of contextual unit. After all, when scholars are interested in votechoice for a member of Congress, Congressional district is obviously the best unit, evenif percent Democrat is not a statistic that is scale invariant (King, 1996). Openshawhas argued that the “simplest solution” to the MAUP has been to ignore it (Openshaw,1984); he explains that

[i]t is also fortuitous that ad hoc zoning systems often produce plausibleresults despite the neglect afforded to the careful definition of areal entities.However, it should be noted that the general absence of comparative studiesmay have helped to disguise the extent to which zone-dependent regularitiesare being uncovered (31).

Cho and Baer have provided this comparison for political science research, probably tothe discomfort of many of the scholars in this field. All who study racial context shouldbe aware that the spectre of the MAUP hangs over everyone’s heads; none is immunefrom the possibility that substantive interpretations may simply be statistical artifacts.

2.2 Our Proposal: Sidestepping the MAUP

While simple statistical solutions to the MAUP are elusive and ignoring it is problematic,there is a geographical solution to the MAUP, or at least a way of sidestepping theproblem that we propose here. The choice of unit should not be haphazard, andit should have geographical meaning. Census data, for example, are reported forarbitrary and modifiable areal units (blocks, tracts, places, etc.), following criteriadetermined by political and logistical considerations rather than because they are“natural areas” with intrinsic geographical meaning (Hatt, 1946). If we think of theParsons and Shils conceptualization of context and the questions that motivate scholarsof racial threat, the real question of interest is how people react to their surroundings.Such a conceptualization naturally emphasizes new possibilities for operationalizingcontext. After all, given this definition, what is the ideal context to be studied? Parsons’and Shils’ “situation of action” should be measured at the level that individuals “see”,and it should be allowed to vary by individuals. Some people, for example, may thinkof their county when thinking of politics, while others picture their nation; the “externalworld” that matters may also vary from policy to policy.

While a boundary drawn by the Census or a state political party may contain mean-ingful places for some people, it is not designed with any given individual in mind.Relationships between the characteristics of such geographies and individual attitudesand behaviors may be hard to interpret because of the heterogeneity in the psychologicalrelationships between people and place: different people living in the same Census tract

8

probably understand their “local community” or “neighborhood” in different ways.6

Even the definition of context as “shared living space” simply assumes that the “space”is indeed “shared” (with concomitant questions about the meaning of “shared”). Whathas been ignored is whether people — who could even be next-door neighbors — varyin the geographic unit relevant to their political and racial attitudes.

In the new measures we develop and test, we ask individuals to define their ownrelevant contexts; rather than assume that administrative units are places with meaningto individuals, we ask our respondents to draw their contexts on maps. Then, the levelof interest does not have to be less aggregated than the level of data available: we careabout the context to which people are reacting, and we can gather individual level dataon both the context seen (i.e., its size and boundaries) and the reactions to it. In this case,the individual units are persons, not modifiable areal units. With this new measure,we are also able to answer the question of how the boundaries of people’s communitiesmatch up to administrative units used in most previous racial context work in politicalscience. If we think context is a “situation for action” and want to understand thepolitical outcomes of individual orientations to the situation, then we need to measurecontext at the level of the individual. If we do so, we also sidestep the MAUP: wewill know that our statistical summaries reflect differences across individuals ratherthan artifacts of aggregation and scale. This is a big advance because the MAUP is not“checkable” or fixable without strong auxiliary, statistical assumptions that often havelittle to do with the substance about which the analyst is an expert.

Given our conceptualization of context and past research on public opinion, we suspectthat geographic units drawn by the Census or political parties are not the contexts thatordinary citizens “see.”

3 What is the Content of the Context?

If context matters for individual-level psychological phenomena like threat, ethnocen-trism, and policy preferences, it must matter via perceptions. That people observe andunderstand their contexts is assumed by most research on this topic in political science,but the standard practice is to use Census data — at whatever level is available in a

6In other fields of research, scholars have already shown that institutionally-defined boundaries maynot be the most relevant for a wide range of outcomes of interest — including teenage pregnancy rates,collective efficacy, school failure, lead hazards, or crime — since physical structures, space allocation,social networks, businesses, and organizations can all play important roles in determining communityboundaries (Buck, 2001; Gauvin et al., 2007; Galster, 2001; Cho and Choi, 2005; Lee, 1968; Golledge andSpector, 1978; Aitken and Prosser, 1990; Nicotera, 2007; Lebel, Pampalon and Villeneuve, 2007; Shinn andToohey, 2003; Grannis, 1998; Colabianchi et al., 2007; Sastry, Pebley and Zonta, 2002; Nicotera, 2007; Wangand Burris, 1997; Coulton et al., 2001).

9

dataset — as the measure of diversity, and the interpretation is that individuals observequalities of their locales, and such observations will affect their opinions and actions.

3.1 Context as Threatening

The racial threat hypothesis was developed most rigorously as an explanation forintergroup conflict and discrimination by Hubert Blalock (Blalock, 1967). In Toward aTheory of Minority-Group Relations he carefully presented a general theoretical frameworkfor understanding phenomena such as the violence and discrimination often aimed atblacks in the southern United States. Figure 1 shows a schematic version of Blalock’spower threat theory: dotted lines represent the causal linkages he envisioned. Thesolid line that directly connects “percent nonwhite” to “motivation to discriminate” isnot in Blalock’s original figure. We have added it here to represent what past researchhas actually tested and what much past research has found — a positive relationshipbetween the size of a minority population (in an areal unit) and anti-minority attitudesand behaviors among members of a majority group.

Fear of Competition Fear of Power Threat

Motivation to Discriminate

Percent Nonwhite

Discriminatory Behavior

Inequalities

Figure 1: Blalock’s Power Threat Theory (Blalock (1967), 29). Percent nonwhite in a locale causesmotivation to discriminate (and thus discriminatory behavior) via “fear of competition” and“fear of power threat.”

When Key wrote his seminal research on the effect of black belts in Southern Politics, hedid not have access to public opinion data to test at the individual level his argumentsabout racial threat; he had to rely on aggregate level vote returns and populationstatistics to infer psychological processes of threat (Key, 1949). This follows the solidblack line in Figure 1. While we now have easy access to survey respondents, scholarsnevertheless have continued the tradition of using census figures to serve as proxiesfor beliefs about the racial context.7 Quillian provides one of the most straightforwardstatements of this practice when he notes:

7In a rare exception, Semyonov et al. (Semyonov, Raijman and Gorodzeisky, 2008) find a connectionbetween inflated perceptions of the size of the foreign population in countries increase negative attitudestoward foreigners. However, immigrant context often does not have the same effects as racial context(Hopkins, 2010).

10

Like past studies, I include percentage black as a group-level cause of prej-udice and accept the interpretation that percentage black increases racialhostility because it is related to perceived threat African-Americans pose towhites.

He continues with a footnote: “As far as I am aware, group threat is the only explana-tion in the literature of why percentage black should influence white racial attitudes”((Quillian, 1996), 821).

Much of the political science research on racial context assumes that beliefs are thesame as the facts, merging Census data with survey questions to determine whetherracial context leads to feelings of threat or lessens feelings of ethnocentrism. However,since Southern Politics, scholars of public opinion have also learned a great deal aboutwhy Census numbers may not be good measures of people’s pseudoenvironments.

3.2 Information (and Misinformation) about Context

The American public’s level of political information and knowledge is often surprisinglylow (Converse, 1964; Bartels, 1996; Carpini and Keeter, 1997). Ordinary citizens alsohave difficulty understanding the concepts involved in statistics, and they often makeincorrect inferences based on personal experiences or recent, salient events ude to theavailability heuristic and other perceptual biases (Ross, 1978; Tversky and Kahneman,1974). When it comes to the racial/ethnic make-up of the U.S., for example, surveys haveshown that Americans greatly overestimate the numbers of blacks, Jews, and Hispanicsin the country (Gladwell, 1995; Nadeau, Niemi and Levine, 1993; Nadeau and Niemi,1995; Smith, 2001; Wong, 2007), and researchers have begun to look at the extent andcauses of Americans’ innumeracy about minority groups (Highton and Wolfinger, 1991;Nadeau, Niemi and Levine, 1993; Sigelman and Niemi, 2001; Wong, 2007). Furthermore,we know that individuals’ misperceptions influence policy preferences. For example,Americans who overestimate the numbers of poor blacks in the U.S. are more likely tooppose welfare programs, and ignorance or misperception about political facts moregenerally can affect both values and policy decisions (Gilens, 2000, 2005; Hochschild,2001; Kuklinski et al., 2000; Gaines et al., 2008). Scholars studying the effects of theeconomy on voting also find that perceptions of the state of the economy can differfrom the true statistics, and that the former can be more important than the latterin influencing vote choices (Hetherington, 1996; Kramer, 1983; Kinder and Kiewiet,1981; Kinder, Adams and Gronke, 1989). To add a further complication, simply givingpeople factual information does not “correct” their policy opinions, and it may evenbackfire to bolster the effects of misinformation (Nyhan and Reifler, 2010); only whenmisperceptions are corrected in a way that changes the interpretations that people give

11

to their (mis)information are their policy preferences changed (Kuklinski et al., 2000).

And, even when Americans’ perceptions and beliefs are fairly accurate reflections ofreality, their interpretations can vary greatly. For example, Gaines et al. find that whileDemocrats and Republicans agreed (and were fairly accurate) about the number ofwar casualties in the current war in Iraq, they differed in how they interpreted thosenumbers and how those interpretations were related to their policy preferences aboutthe war (Gaines et al., 2008). That partisans share perceptions but not interpretations ofthe facts and opposed policy preferences has been found in other policy areas as well(Gerber and Huber, 2008; Bartels, 2002). The issue of misperception or misinformationhas largely been missing in the research on racial context (see Alba et al. 2005, althoughtheir work is presented as research on misperceptions rather than the effect of racialcontext). One strand of work in criminology has been an exception, noting the effect ofmisperceptions on fear of crime. One could argue that crime in a neighborhood makespeople afraid of crime. But, Chiricos et al. also show that actual levels of crime have aremarkably weak link to perceptions of crime (Chiricos, Hogan and Gertz, 1997; Chiricos,McEntire and Gertz, 2001). And other research has shown that the percentage of blacksliving in an area can lead whites to perceive higher crime rates (Quillian and Pager,2001; Covington and Taylor, 1991; Merry, 1981). In other words, because of negativestereotypes of blacks, whites who live in areas with more blacks may believe there ismore crime nearby and therefore fear crime more, whether or not crime is a problem intheir area.

Inaccurate perception of racial demographics can be compounded by other mispercep-tions of “that part of the external world which means something to the actor” ((Parsonsand Shils, 1951), 56). While some scholars have equated “context” with racial/ethnicdiversity, the content of a context obviously extends beyond this one dimension. Forexample, research on racial threat has on occasion also incorporated the effects of so-cioeconomic and partisan context along with racial context, and the findings indicatethat such attributes can interact or that socioeconomic status or partisan context maybe even more relevant an influence on people’s racial attitudes than the racial/ethnicdiversity of where people live, individuals’ personal income, or their own partisanidentity (Gay, 2004; Branton and Jones, 2005; Oliver and Mendelberg, 2000; Campbell,Wong and Citrin, 2006; Orbell and Sherrill, 1969).

However, the same problem of possible misperceptions arises. For example, the researchon the effect of the economy on vote choice has supplied us with ample evidence that“it’s the economy, stupid” should be changed to reflect that people’s perceptions of theeconomy are what, in fact, spur votes (Ansolabehere, Meredith and Snowberg, 2010;Duch, Palmer and Anderson, 2000; Conover, Feldman and Knight, 1987).8 While broken

8While there has been some research on the level of knowledge of the racial make-up or unemployment

12

windows are easy to observe, other indicators of the average socioeconomic status ofone’s neighbors can be less visible than race/ethnicity; unless someone is wearing an oldcollege sweatshirt, for example, it is difficult to discern at a distance if one’s neighboris a college graduate or not. Similarly, partisanship is not something most peoplewear on their sleeve (or car or front yard); while individuals are often egocentricallybiased to think other people think like them — and therefore assume a homogeneityof partisanship in their surroundings — a Republican living in a liberal college townalso might think of himself under siege. Using our new maps-based measurementof context, we can show the similarities and differences of misperceptions about theracial make-up, socioeconomic well-being, and partisanship of one’s surroundings.One cannot assume that merging multiple contextual variables — like percent black orpercent unemployed in the census tract — to one’s survey has the same implicationsfor causal mechanisms simply because the variables come from the same source, likethe U.S. Census or other governmental bureaucracies.

4 Application of the New Measurement

To illustrate the utility of our proposed conceptualization and measurement strategy,we use data from a pilot survey that we conducted of 62 individuals living in one countyin the state of Michigan in 2004. The in-person survey included an innovative map-drawing measure of context and the research design allows us to bring the individualback into the conceptualization and measurement of context. Despite the limitedsize and location of the pilot, our findings about misperceptions of racial context, forexample, are almost identical to ones found in a national survey conducted during thesame time period. While the particular maps drawn by our respondents are not meantto be generalizable to the nation, our measure of context is broadly applicable, as arethe questions our measure raises about standard practices. The sample design of oursurvey also allows us to suggest improvements for future studies of context.

4.1 Survey Design

There are several requirements for a sample design for any survey that would providedata to address and expand upon theories of the psychological reactions of individualsto characteristics of their geographies. First, any design that provides a persuasive testof these theories must sample enough of both geographies and individuals to enablestatistical tests at both levels (of context and of individual) to detect substantivelymeaningful effects. Second, such a design must allow researchers to locate individuals

level in the country, there is much less research on local context and beliefs at smaller units of geography.

13

accurately within geographic areas. Third, any sample design that analysts desire touse to distinguish between effects of contexts and the effects of individuals must samplemultiple individuals per context. And, a design that is to be used to estimate the effectsof contextual units must sample as many contextual units as possible, stratifying onkey independent variables.9 In the sample design of our pilot study, we met all fourrequirements. We use Census block groups as the basis of our sampling because blockgroups are the smallest administrative unit for which racial data are available (andlargely unscrambled) and could be used as building blocks for ascertaining contextualeffects at more aggregated levels. After stratifying — and thereby targeting whites andblacks living as both majorities and minorities in their block groups — and samplingbased on non-rural block groups in the county, we purchased a white pages sample ofnames, addresses, and phone numbers.10 The white respondents lived in block groupsthat ranged from 0 to 74 percent black, and the black respondents lived in block groupsthat ranged from 3 to 97 percent black. In other words, we were able to interviewrespondents who lived in very diverse racial contexts.

To create a measure of personally relevant places that matches our conceptualization ofcontext, we develop a map-drawing addition to a traditional political science survey.11

So, in addition to answering a number of survey questions, the respondents were alsoasked to refer to a few maps. They were first shown two maps — one centered on theblock group in which their house was located and one encompassing all of the countyfrom which the block groups were sampled — and were asked to draw on either mapthe area that made up their “local community.”12 The approximate location of theirhouse was indicated on each map, and the maps included both block group boundariesas well as nearby streets. In order to ascertain respondents’ perceptions of the content

9One problem with racial context research using data like the NES is that conclusions are often madeas though respondents lived in counties ranging from 0 to 100 percent black, well beyond the scope of thedata. Given that the NES and other national surveys work assiduously to ensure that their samples arerepresentative of the nation, there are in fact relatively few whites, for example, in their samples that livein majority-minority counties.

10While we would like to have interviewed equal numbers of white and black respondents, our finalsample contained 41 whites and 21 African Americans living within 25 different block groups.

11We build on the work on mental mapping pioneered by Lynch and which has continued in geographyand sociology (Matei, Ball-Rokeach and Qiu, 2001; Lynch, 1973; Grannis, 1998; Coulton et al., 2001; Coultonand Korbin, 2007; Garling and Golledge, 2000; Tversky, 2000; Svendsen, Campbell and Fisher, 2008;Svendsen and Campbell, 2008). See (Weston and Handy, 2004) for a review of the mental mappingliterature in urban planning and geography.

12The map-drawing task was one of the first in the survey, so the respondents were not primed to thinkabout particular issues or communities by other survey questions. We also provided no definition of“local community,” both to avoid the respondents second-guessing what the researchers might want, andalso to avoid cueing any particular level of aggregation. Granted, the largest map that could be drawnwas at the county level, although respondents could also respond that they did not think of their “localcommunity” in this way. While it woud have been ideal to know how perceptions of their communitieschanged, depending on the issue, for a pilot study we were interested in getting a more general senseof how people would define their local community. The terminology “local community” was chosento emphasize a politically and personally relevant grouping of people (a “community”) that is spatiallyinterdependent (“local”).

14

of their contexts, we then asked about perceptions of the proportion of blacks andwhites, Democrats and Republicans, and unemployed living in that community.13 Later,they were shown a map with their block group highlighted, and were again asked todescribe the racial, partisan, and economic breakdown of the block group. They werealso asked about their perceptions of these groups at the national level. Thus, we wereable to gauge perceptions of multiple politically-relevant characteristics of multiplegeographic contexts, and because some of the contexts were shared, we could makecomparisons both within and across individuals of two different racial groups.

4.2 Does the Government Define Pseudoenvironments?

We wanted to allow individuals to define for themselves (using maps) what they meanby “community,” rather than assume that how context affects political judgments mustoccur within administrative units defined by government officials. While we weredesigning the map-drawing questions, we had contradictory predictions. On the onehand, we were skeptical that people knew where the boundaries of any government-designated units in which they resided lay. For example, there was no reason whya respondent even had to include her own residence in her perception of her “localcommunity.” And, unlike the drawing of Congressional districts, we did not restrictrespondents to communities of compact or contiguous areas. On the other hand,we thought (wrongly, in this case) that respondents might be guided by the cuesprovided and follow the bold lines presented on the map, which designated blockgroup boundaries.

Almost 2/3 of the respondents (n=36) chose to draw their community on the smallermaps centered around their block group (see Figure 2 for examples), while the restchose the larger one (23) (see Figure 3 for examples).14 Regardless of the size of their“local community,” none of the respondents’ drawings followed block group lines neatly.A couple of the maps had non-contiguous areas marked, some marked areas as small asa single street, while others encompassed the entire large map of the two cities withinthe county. No two maps were identical, even for residents living in the same blockgroup, and they ranged in size from smaller than a block to larger than two cities

13While there are many possible measures of economic context, we chose to use percent unemployed.The average education level of residents in an area may be a more reliable measure (Huckfeldt 1986)than unemployment, but we believe that if perceptions are driven in part by observational learning, thenresidents’ employment status may be more visible to others than their diplomas. Ansolabehere et al. alsofound that information about unemployment affects political outlooks, whereas knowledge of gas prices —another possible indicator of economic context — does not.

14We were concerned that respondents might find the map-drawing task overly difficult, but our worriesappeared to be unfounded. One respondent did not draw on the maps, explaining that she did not thinkof her “local community” in these ways. Two respondents who were visually impaired did not draw onthe maps, although they were able to describe the major streets and landmarks that defined the borders oftheir local communities. All other respondents were able to draw their local communities on the maps.

15

together.15 There is no clear pattern by race, income, or education as to what size ofcommunity was drawn.16 A few respondents volunteered information about whatthey were thinking as they drew their maps. For example, one mentioned that she wasexcluding the low-income housing nearby her house as she drew an irregular shape.Another said that he was including his home, where he shopped, and where he worked,as he drew his community of non-contiguous polygons.

Figure 2: Four examples of “Local Communities” drawn on “small” maps.

We need to keep in the mind that even though the size of the communities drawn inthe pilot varied widely, we believe the range would have been even greater if there hadbeen more than two choices of maps offered to respondents. Given social psychologicalresearch about anchoring, it is very likely that people chose communities that would fiton one of these 2 maps, and that if they were not so restricted to these particular spatialresolutions (e.g. were given an interactive tool), the size of the community may havebeen even larger for some respondents, extending beyond the county in which they

15About a quarter of the communities drawn were smaller than a block group. Each hand-drawn mapwas traced into ARCGIS, where we could determine the size of the community and how it related toCensus block groups.

16There is also no simple relationship between size of community drawn and ethnocentrism. Whilewe believed that more ethnocentric individuals would choose a smaller community — limiting theirinclusion of outgroup members in the community, for example — there is no clear pattern across a rangeof racial attitudes that this is the case. These analyses used polygon(s) as “size” of community for theseassessments.

16

Figure 3: Four examples of “Local Communities” drawn on “big” maps.

live.17 Nevertheless, even given the limited scope of the pen and paper maps, it is clearthat people’s “local communities” are (1) different from governmental administrativeunits and (2) idiosyncratic and not commonly-shared visions.

Later in the survey, respondents were shown the “small” map again (like those inFigure 2), this time with their block group highlighted in yellow. They were told thatthe borders on the map showed the boundaries of what the Census Bureau defines as a“block group.” Respondents were then asked whether they thought the highlightedarea captured what they thought of as their neighborhood. Sixty-one percent agreedthat their block group was indeed a reasonable representation of their neighborhood.This percentage may be biased upward for two reasons: suggestibility — with a crediblegovernment institution as the source — and acquiescence bias. Even taking into accountthat the response could be inflated, almost 4 out of 10 respondents did not think of theirblock group as a close approximation of their neighborhood. For those who disagreed,we do not know whether they believe their neighborhood would be smaller or larger.Other research has also shown that people’s perceptions of their neighborhood vary

17We are currently developing such an online module using Google Maps. While the pilot study’s penand paper costs were high because of the need for in-person interviews, the computer interface will allowus to make this measurement tool feasible and cost-effective across a fuller range of resolutions. This toolwill help scholars interested in the boundaries of communities, immigration, intergroup conflict, andpublic policy to see how ordinary people understand and change their understandings fo their communityin response to changes in their environments. It will be able to be integrated with other online or mobilesurveys.

17

greatly. Sastry et al., for example, find that a third of their sample of Los Angelesresidents described their neighborhood as their block or street, and almost a quarterdescribed it as several blocks or streets (Sastry, Pebley and Zonta, 2002). In contrast,Huckfeldt, Ross et al., and Tienda argue census tracts are the best approximation ofneighborhood (CITE). And, Kwan finds that few people regard neighborhoods as smallas their census tract (CITE). This all serves as evidence that “neighborhood” is no morea commonly shared container than is “community,” and that the wide variance acrossrespondents’ maps is not because “community” is a particularly idiosyncratic term.

Our findings suggest that people’s definitions of “local community” vary a great dealin size, and that people do not tend to equate their neighborhood with their “localcommunity”; otherwise, we would have expected to see many more drawings thatcorresponded with the block group boundaries on their small maps. In other words,the assumption made in operationalizing context that people observe and define theircommunities in similar ways — particularly along lines defined by government agencies— is not met. The “situation of action” is not a uniform container. Furthermore, byallowing individuals to define their own meaningful contexts, we are also able tosidestep the problems caused by the MAUP.

We want to stress again Parsons’ and Shils’ definition of the situation of action; it is“that part of the external world which means something to the actor.” Individuals varya great deal in the part of the external world that is relevant to them, and this diversityin salient boundaries should be taken into account in our analyses of contextual effects.It may be helpful to think of a medical analogy: even if subjects in a study are all giventhe same dose of a drug, they may differ greatly in how well they absorb the medicine;their reactions will reflect the amount absorbed rather than the amount ingested. Inoperationalizing context as the situation of action defined by the actor herself, we moreclosely approximate the medicine absorbed. When we find contextual effects usingconvenient, institutionally-drawn boundaries, it is impossible to determine if theseeffects are overly conservative or liberal, since they most likely vary a great deal acrossindividuals, across subgroups, and across geographic units. While equating contextwith containers drawn by government bureaucracies is a logistically expedient way tooperationalize context, only by asking respondents to define the boundaries of theirown contexts can we measure the concept of context to capture psychological outcomesand also avoid the MAUP.

4.3 Do You See What I See?

A common assumption in the research on racial context is that people’s beliefs alignwith the facts about the demographic make-up of the areas in which they live. We are

18

interested in whether individuals indeed are accurate when it comes to describing theircontexts. Using the map-drawing measure of context, we are able to examine whetherpeople’s pseudoenvionments are distorted in racial/ethnic, partisan, or economic terms.Our respondents were asked to describe the percentages of blacks and whites, un-employed, Democrats, and Republicans in their block group, their self-drawn “localcommunity,” and the nation as a whole. Thus, we can compare people’s perceptionsacross multiple levels with the Census reports for the same geographic levels, and alsoacross individuals for the same geographic unit. For the analyses presented here, thearea we use as a point of comparison for respondents’ “local community” is the averageof any block group that is included in their drawing. For example, if a “community”was drawn to overlap with three different block groups, then the “objective” point ofcomparison for racial context used is the Census-reported percentages of whites andblacks for those three block groups together. We considered a couple other alterna-tive measures, but none seemed superior. Using only block groups that were entirelyenclosed within a “community” is problematic because some communities did notinclude a single entire block group. We also considered calculating the fraction of thearea of a block group contained within a “community” and using it to include only thatproportion in the demographics; however, this measurement assumes that individualsare evenly distributed across a block group, which is clearly incorrect.

We want to compare perceptions of context across different levels of aggregation becausesimilarities or differences across levels will give us information about the mechanismby which facts about one’s environment become beliefs that illustrate one’s pseudoen-vironments. If levels of misinformation or accuracy are constant, regardless of the sizeof the context, then we could infer that how one learns about one’s surroundings maybe similar, whether one is picturing one’s block, county, or state. However, if levels ofmisperception vary, then we may infer that sources of information about where peoplelive may also vary, or that people process information differently, depending on thesize of the context they are envisioning. For similar reasons of gaining more insightinto the mechanism linking objective and subjective contexts, we also want to exploreperceptions of different attributes of the same unit. Are people better informed aboutthe racial make-up, for example, than the partisan make-up of their block group? Ifso, do they learn about the different qualities of the places where they live from variedsources?

We hypothesized that respondents would be more accurate in their perceptions of theirlocal community than the nation, given their personal experiences over a sustainedperiod of time in the area. However, we were wary of the dangers of availabilityheuristics, which could lead someone to exaggerate the percentage of outgroups in thelocal area if he saw one or two outgroup members on a regular basis. However, the

19

same problem could arise for national perceptions, since the presence of an AfricanAmerican news anchor and actors on television shows, for example, could lead a whiterespondent to misperceive the proportion of blacks in the nation as well. We alsoexpected that respondents would be more accurate about their racial contexts thanabout the levels of unemployment or percentages of partisans in their environments. Aswe mentioned earlier, the race of a neighbor is oftentimes more visible at a brief glancethan that neighbor’s partisanship or job status. And, while an outsider’s observationof race will often not accuately reflect an individual’s self-identification, we imaginethat unless someone observes their neighbor putting up Obama signs or staying homeduring working hours, their perceptions of their neighbor’s partisan identification oremployment status will be equally fraught with error. We start by examining perceptionsof racial context before turning to socioeconomic and partisan context.

Figure 4 shows two scatterplots, comparing respondents’ perceptions of blacks andwhites in their block groups with the Census percentages for the respective groups in thesame areas. As can be seen from the plot on the left, the fitted smoothed curve is flatterthan a 45-degree line, indicating that respondents’ perceptions of the percentage ofblacks living in their block group is greater than that reported by the Census. Conversely,their estimates of the number of whites are smaller than the objective numbers. Even ifthe situation were a container, orientations toward the container are not uniform andare not accurate. Lippmann’s pseudoenvironment, in other words, is not identical tothe world described by Census numbers. On average, the proportion of whites wasunderestimated by 5 percentage points and that of blacks overestimated by 8 percentagepoints.

●

●●

●

●

●

●●

●

●

●

●

●

●

●

●

●●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

0.0 0.2 0.4 0.6 0.8 1.0

0.0

0.2

0.4

0.6

0.8

1.0

% Black Perceived in Block Group

Cen

sus

%Bl

ack

in B

lock

Gro

up

●

●

●

●

●●

● ●

●

●

●

●

●

●

●

●

●

●

●

● ●

●

●

●

●

●

●

●

●

●

●

●

●● ●

●●

●

●

●

●

●

●

●

● ●

●

●

●

●

●●

●

●

0.0 0.2 0.4 0.6 0.8 1.0

0.0

0.2

0.4

0.6

0.8

1.0

% White Perceived in Block Group

Cen

sus

%W

hite

in B

lock

Gro

up

Figure 4: Relationship between Objective Racial Context and Perceptions at the Block Group.

Because of the sampling design, we are also able to look at the perceptions of whiteand black respondents living in the same block group; while one could hypothesizethat neighbors may have shared spatial experiences and similar perceptions of the

20

neighborhood in which they live (Stipak and Hensler, 1983), one might also hypothesizethat black and white respondents have different experiences living in the same area(Kwan, 2000). Scholars have found, for example, that while both whites and blacksare willing to live in integrated neighborhoods, their definitions of “integration” aremarkedly different (Clark, 1991). Blacks and whites also rely on different sources ofmedia for news and entertainment. Comparing white and black respondents who livewithin the same block group, we find that blacks’ estimate of the proportion of blacksliving in their block group was, on average, 12 percentage points lower than their whiteneighbors’ estimates, and their estimate of the proportion of whites was 13 percenthigher than that of their white neighbors. Within-racial-group average differencesrange from 1 to 5 percent. However, whites and blacks both tend to misperceive theirracial contexts and in the same directions: they overestimate the percentage of blacksand underestimate the percentage of whites.

Perceptions of racial context at the community level are similar to those at the blockgroup level, with some variation. Figure 5 shows that at this geographic level, theline for percent black appears even flatter and the percent white steeper than thosein Figure 4, indicating that perceptions of the numbers of blacks and whites are evenmore distorted when people evaluate their own “ community” compared to blockgroup perceptions. On average, respondents overestimated the proportion of blacks by15 percentage points and underestimate that of whites by 13 percentage points. Themisperceptions are even greater at the national level: respondents overestimate theproportion of blacks by 17 percent and underestimate the proportion of whites by 16percent.18 Furthermore, while black and white neighbors have different pictures ofthe same neighborhood in which they live, black and white respondents in our surveyshare similar visions of the nation; blacks’ estimate of the proportion of blacks living inthe nation was only 3 percentage points lower on average than whites’ estimates.

Another way to think about these differences is presented in Table 1. The table presentsthe mean percentage of whites and blacks perceived in the nation compared to the 2000Census figures, and the same comparison for the respondents’ “local community” andfor their block group. The means are calculated for the differences between subjectiveand objective context at the level of the individual respondent. While the percentages ofblacks are consistently overestimated and that of whites underestimated, an interestingpattern emerges from comparing these different geographic units: perceptions of racialcontext become more accurate at more localized levels. In other words, while respon-dents tended to overestimate the percentage of blacks in the nation by 18 percentage

18These results are replicated almost identically in the 2004 General Social Survey. In the GSS, respon-dents’ estimates of the proportion of whites and blacks were 59 percent and 31 percent, respectively (Wong,2007). In our pilot, the corresponding numbers were 58 percent and 30 percent. This similarity lendssupport to the idea that these levels of misperception in our pilot are not unique to our sample.

21

●

●

●

● ●

●●

● ●

●

●

●●

●

●

●

●

●●

●●

●

●

●

●

●●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●●

●

●

●

0.0 0.2 0.4 0.6 0.8 1.0

0.0

0.2

0.4

0.6

0.8

1.0

% Black Perceived in Community

Cen

sus

%Bl

ack

in C

omm

unity

●

●

●

●

●●

●●●

●

●

●

●

●

●

●

●●

● ●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●●

●

●

●

●●

●

●

●

●

●●

0.0 0.2 0.4 0.6 0.8 1.0

0.0

0.2

0.4

0.6

0.8

1.0

% White Perceived in Community

Cen

sus

%W

hite

in C

omm

unity

Figure 5: Relationship between Objective Racial Context and Perceptions at the CommunityLevel.

points — thinking the nation is 30 percent black instead of 12 percent — the overesti-mate at the block group level was only 9 percentage points. Of course, it is possible thatordinary citizens do not have a sophisticated understanding of percentages, providinganswers that sum to more than 100. The gap between objective and subjective context,in that case, could result from a simple lack of mathematical understanding. In order todetermine whether this is the case (and test the reliability of our map-based measure ofcontext), we included in our survey multiple measures of respondents’ perceptions oftheir self-defined communities. In addition to being asked what percentage of blacks,for example, lived in their “local community,” they were asked (later in the survey) tolook again at the map they drew and say whether their community was “mostly white,”“mostly black,” “half and half,” or “a mixture.” While Americans may not always givepercentages that add up to 100 percent (Wong, 2007), our respondents do display anunderstanding of percentages generally: for example, for those who described theircommunity as “mostly white,” the estimates for percent black in their communityranged from 1 to 50 percent, and averaged 64 percent white and 25 percent black. Forthose who thought their areas were “mostly black,” they perceived their communitiesto be anywhere from 60 to 95 percent black, with an average of 75 percent black and 19percent white. In other words, the percentages were not just random guesses; they seemto map reasonably onto common sense understandings of quantity and proportion.It should be noted that in the block groups sampled, the racial distribution is largelyblack and white. While there are small numbers of other minority groups present, itwould be unsurprising if our respondents perceived only a biracial neighborhood.

It appears that people are more misinformed about their racial contexts at larger geo-graphic units compared to smaller ones. This difference in misperception perhaps couldbe explained by different sources of information about each geographic level, withpersonal experience more accurately informing perceptions at the localized levels. We

22

National Local Community Block GroupBlacks Subjective % 30 41 42Blacks Objective % 12 26 35

Blacks Average Difference 17 15 8Whites Subjective % 59 50 49Whites Objective % 75 63 53

Whites Average Difference -16 -13 -5

Table 1: Average Racial Perceptions Across Contexts

1

Table 1: Misperceptions of Racial Context Across Geographic Levels

will come back to this possibility after examining perceptions of economic and partisancontext across these same three levels. The story of growing levels of misperception asthe pseudoenvironment increases in size becomes more complicated when we look atthese different types of context.

We start with respondents’ beliefs about the economic and partisan contexts of theirneighborhoods. Figure 6 and Figure 7 present plots for perceptions of the percentunemployment, Democrat, and Republican at the block group level, compared withCensus figures and vote returns for 2004. In contrast to the racial content of one’scontext, the relatively flat line in Figure 6 shows that respondents are even more misin-formed about the level of unemployment in their block group than the percentage ofblacks and whites, with very little relationship between objective and subjective eco-nomic contexts.19 On average, respondents overestimated the percentage of residentsunemployed in their block group by 16 percent, although the misperceptions rangedfrom an overestimate of 82 percent to an underestimate of 10 percent. Respondentswith a college education tended to be more accurate than those with less education,although no clear pattern emerged for differences based on income. Republicans andmen tended to be more accurate than Democrats and women. Comparing residents ofthe same block group, whites tended to overestimate the percent unemployed morethan blacks on average, although the difference is small (less than 6 percent) and thevariance in estimates was greater for blacks than whites. Respondents’ perceptions ofthe partisan make-up of their contexts also seem to bear only a passing resemblance tothe objective measures of partisan context.20 On average, respondents underestimatedthe percentage of Democrats in their block group by 6 percentage points, and overesti-mated the percentage of Republicans by 5 percentage points. This pattern was true forboth white and black respondents. Surprisingly, having a college education and incomehad no effect on accuracy, nor did partisanship of the respondent. So, while there islittle relationship between perceptions and objective indicators of socioeconomic and

19For “objective” data about unemployment rates, we used data from the 2000 Census. For the blockgroups in our sample, unemployment ranged from 0 to 14 percent.

20For the “objective” basis of comparison of partisanship, we use the 2004 presidential vote. Becausevote returns are aggregated at the precinct level, we average the vote for any precincts that overlap with ablock group’s boundaries to create the block group’s “objective” partisanship. The block groups in oursample ranged from 62 to 96 percent Democrat.

23

partisan contexts, respondents are, at least, more accurate, on average, in their estimatesof living in predominantly Democratic block groups than in their beliefs that about oneout of every six residents of their block group is unemployed.

!!!!

!!! !!!

! !!! !! !!!

!!

! ! !! !!! !! !

!!! ! !! ! !!

!

!

! !

!!!

!

! !

0.0 0.2 0.4 0.6 0.8 1.0

0.0

0.2

0.4

0.6

0.8

1.0

Total

% Unemployed Perceived in Block Group

Cen

sus

%U

nem

ploy

ed in

Blo

ck G

roup

!!!!!! !

!!

! !!! !!!

!

!!!

! ! !!

!

!

! !

!!

! !

0.0 0.2 0.4 0.6 0.8 1.0

0.0

0.2

0.4

0.6

0.8

1.0

Whites

% Unemployed Perceived in Block Group

Cen

sus

%U

nem

ploy

ed in

Blo

ck G

roup

!

!!!

! ! !!!! !! !

! !!!!

0.0 0.2 0.4 0.6 0.8 1.0

0.0

0.2

0.4

0.6

0.8

1.0

Blacks

% Unemployed Perceived in Block Group

Cen

sus

%U

nem

ploy

ed in

Blo

ck G

roup

Figure 6: Relationship between Objective Economic Context and Perceptions at the Block GroupLevel.

Misperception of economic context of the “local community” is similar to that at theblock group level, with an average overestimate of 15 percent and very little overallrelationship between perceptions and objective measures. Objective unemploymentin the block groups ranged from 2 to 12 percent. When asked about unemploymentat the community level, blacks were much more likely to overestimate the level ofunemployment — by 10 to 15 percentage points — than were whites. Perceptions of

24

!

!! !

!

!

!

!! !

!!! !

! !!!

!!

!! !!! !!

!!

!!

!! !! !! !

! ! !

! !! !!!

0.0 0.2 0.4 0.6 0.8 1.0

0.0

0.2

0.4

0.6

0.8

1.0

All Respondents

% Democrat Perceived in Block Group

% D

emoc

rat 0

4 Pr

esid

entia

l Vot

e in

BG

!

! !!

!

!

!

! !!

! !!!

!! !!!

! !

! ! !! !!!

! !

! !

! !! !! !!

!!!

!! !!! !

0.0 0.2 0.4 0.6 0.8 1.0

0.0

0.2

0.4

0.6

0.8

1.0

All Respondents

% Republican Perceived in Block Group

% R

epub

lican

04

Pres

iden

tial V

ote

in B

G

Figure 7: Relationship between Objective Partisan Context and Perceptions at the Block GroupLevel.

partisan context at the “local community” level are also similar to that at the block grouplevel: respondents underestimated the proportion of Democrats by 7 percentage pointsand overestimated the proportion of Republicans by 4 percentage points, but hand-in-hand with this greater accuracy is very little relationship between the perceived partisanmake-up of a community and objective measures of partisanship for that same area.The objective measures for the communities ranged from 66 to 92 percent Democrat.Blacks were slightly more likely to be misinformed than whites, but all estimates werewithin 10 percentage points of the objective facts. A similar pattern of misperceptionappears at the national level as well. Misperceptions about unemployment in the U.S.are, on average, an overestimate of 15 percent. When asked about unemployment at thenational level, blacks were much more likely to overestimate the level of unemployment— by 10 to 15 percentage points — than were whites. So, in contrast to perceptionsabout racial context, people’s visions of the economic context are distorted to the sameextent, whether they are thinking of their block group, community, or nation. Therewas also a lack of variation across levels for partisan context: at the national level, therewas again only a low level of misperception on average; respondents overestimated theproportion of Democrats by less than 5 percent and underestimated that of Republicansby about 5 percent.

Table 2 presents a summary of the misperceptions of the three types of context —racial, partisan, and economic — across the three levels of context — block group,“local community”, and nation. It presents the mean percentage of whites and blacks,unemployment, and Democrats and Republicans perceived in the U.S., compared tothe 2000 Census figures and 2004 election returns. Table 2 shows that there is littlevariation in average misperceptions of unemployment and partisanship from the localto national level. It also shows that perceptions of economic and racial context arethe most distorted, while respondents have fairly accurate pictures of the partisan

25

composition of their block group, community, and nation.

National Local Community Block GroupWhites Subjective % 59 50 49Whites Objective % 75 63 53Blacks Subjective % 30 41 42Blacks Objective % 12 26 35

Unemployed Subjective % 21 21 21Unemployed Objective % 6 5 5Democrats Subjective % 52 68 71Democrats Objective % 48 75 77

Republicans Subjective % 46 27 27Republicans Objective % 51 24 21

Table 3: Average Subjective Perceptions and Objective Percentages of the Nation

3

Table 2: Misperceptions across Types of Context and Geographic Levels

5 Maps and Misperceptions

Why does our new measurement strategy reveal differing levels of accuracy about racialcontext across geographic units? People may simply have a better vision of their blockgroup, and can more easily envision 100 people, for example, than they can pictureabout 300 million Americans as a whole. For example, using the 2000 GSS data, Wongfound that respondents’ estimates of the different racial and ethnic groups in theirlocal community were much more likely to add up to 100 percent than their nationalestimates, which on average ranged from 122 to 190 percent, depending on whetherHispanic and multiracial individuals were counted (Wong, 2007). Another explanationis that people learn about their various environments in different ways. For example,in their work on the difference between sociotropic and aggregate economic studies’effects, Ansolabehere et al. speculate that one reason for the differences in evaluations isthat there are different sources of information (Ansolabehere, Meredith and Snowberg,2010). In our pilot, we explicitly address this possibility. One open-ended questionthat we asked of all respondents was the source of information for their perceptionsat the local, block group, and national level.21 As we show in Table 3, the responsesdiffer greatly between where one gathers knowledge about the nation and where onegains information about one’s local community and block group, with the responsesabout the latter two blending together. Personal experience, families, and friends play amuch larger role in knowledge about people’s community and block group; as shownin previous research, everyday observation was cited by at least half our sample asthe source of information about the local community and block group, and socialnetworks provided information to about a quarter (Huckfeldt, 1980; Gilliam, Valentino

21We asked respondents about the sources of information for their perceptions at each level, but did notask them to differentiate sources for racial, economic, and partisan context.

26

and Beckmann, 2002). In contrast, media is the primary source of information about thenation; over two-thirds of the responses given referred to media, and few mentioned anyother source as the basis of their knowledge about the United States as a whole. Whilethe media can provide information both via news shows as well as via entertainment,the misperceptions must come from the foci of news stories and images, as well asfrom the content and casts of entertainment shows (Armstrong and Neuendorf, 1992;Busselle and Crandall, 2002). Potential biases from availability heuristics seem moreprominent in national racial perceptions than local, which suggests the important roleof the media in creating perceptions of racial threat.22

Block Group Local Community United StatesMedia 4 8 67

Social Network 22 27 0Institutions 6 5 5

Everyday Observation 61 50 5Guess and DK 6 8 16

Other 2 2 7

Table 4: How did you learn about the composition of this area?

4

Table 3: Sources of Perceptions of Context across Geographic Levels

5.1 Differences Across Types of Context

These explanations are plausible at least with respect to racial context. However, whywould these same sources of information not lead to variation in misperceptions ofeconomic and partisan context across levels? Why are levels of misperceptions ofunemployment and partisan context invariant to the unit of aggregation, but in differentways? One possible explanation for perceptions of economic context is that such beliefs— regardless of level — are mainly based on media reports. Because few news sourceswould ever discuss unemployment at levels smaller than a respondent’s city, he orshe most likely is not exposed to information about unemployment at the block grouplevel. And, as we mentioned earlier, casual observation of one’s surroundings is muchless likely to give one clues about how many neighbors are unemployed than howmany are of a different racial group. Therefore, perhaps the only source of informationthat people have about unemployment rates is what they hear on the news; inferencesacross all levels are drawn from the same source of information, with an assumption ofuniformity across levels.The same hypothesis could hold for partisan context as well.However, since elections regularly flood the news with reports of who (and which party)

22We chose the open-ended format in order to get as complete a range of responses as possible. Onedownside is that we did not have follow-up questions to ask, for example, exactly which national newsshow they watched. After all, if there is only a one percent chance of an African American being a viewerof Fox News’s prime-time broadcast (Nielsen, 2010), then different news sources could explain some ofthe differences in perceptions between whites and blacks.

27

won office, there should be more localized information about partisan context. Thiswould lead one to expect greater variation across levels, if, for example, one heard of aDemocratic member of Congress being reelected in one’s district while a Republicanwon the mayoralty.

Contrary to our expectations, respondents are more accurate in their perceptions ofpartisan context relative to racial context. When it comes to understanding the accuracyof perceptions of partisan context across all three levels, one possible explanation isthe role of elections in learning. Stories about election returns may highlight partisansurroundings in a dramatic way for individuals, unmatched by any event that wouldhighlight their racial and economic contexts as clearly, regularly, and officially. Of course,one obvious wrench in the explanation that media coverage can lead to greater accuracyis the fact that the news regularly presents stories about the nation’s unemploymentrate and the 2000 Census results, yet respondents were much more misinformed aboutthese facts relative to facts about partisanship. There is another plausible explanation.In contrast to economic and racial context, the partisan composition of the nation(and likely many communities and block groups) is more evenly distributed betweenDemocrats and Republicans. Therefore, if a respondent took a guess around 50 percent,he or she is more likely to appear “accurate” about partisan context at the national level,but grossly incorrect about economic and racial contexts. Research also shows thatpeople are more likely to misperceive rare events than more normal ones. However, ifone looks more closely at the contexts in our sample, this explanation is also problematic:the county from which we drew our sample leans Democratic, and the means andmedians of the distributions in both block groups and the local communities drawnranged between 75 and 80 percent Democratic. Obviously, more study is needed togain a better understanding of how people develop their perceptions of different typesof context and across different levels.

The data gathered from our study indicate that “context” is a complicated conceptthat involves both factual as well as perceptual information. The figures and tablesin this section provide us with a better sense of just how different reality can be fromLippmann’s “pictures in people’s heads.” The content of a context that is experiencedseems to vary across levels of aggregation as well as type of context (racial, economic,or partisan). The measurement of context needs to emphasize its psychological nature,taking into account both objective and subjective context; scholars cannot simply assumethat objective context is experienced directly without filters.

28

5.2 Reflections on Past Findings

One question raised by the finding that misperception of racial context is greater atlarger units is how best to interpret previous findings that context is threatening atlarger units but not at smaller ones. For example, most of the research that finds supportfor the racial threat hypothesis measures context at the level of county, metropolitanarea, state, region, or country; in contrast, the research that has found support for thecontact theory tends to measure context at smaller units, like census tract or block group.In the past, this difference was interpreted as evidence that contact at the smaller unitswould diminish ethnocentrism, while diversity at the larger aggregate levels madesuch intergroup contact more likely. Diversity without contact, in contrast, would leadto feelings of threat (Stein, Post and Rinden, 2000; Kinder and Mendelberg, 1995).23

Another interpretation, however, is that these differences in outcomes across levelsare simply the result of the MAUP, and we should draw no substantive conclusionsabout differing perceptions of threat across levels. A third interpretation, given ourfindings, is that overestimates of the size of outgroups at larger contextual units relativeto misperceptions at smaller units is what leads to a greater sense of threat in the moreaggregated levels. In other words, if individuals had a more accurate picture of theirsurroundings — rather than exaggerating the number of outgroup members — theymight not feel a sense of racial threat at the metropolitan or county level, for example.This does not preclude the idea that personal interactions with outgroup membersdiminish ethnocentrism, but it does raise more questions as to why racial context rarelyposes a “threat” at the more localized level in previous research.

We also raise the question of how one should interpret perceptions by different racial orethnic groups. While the bulk of research on racial threat has focused on the attitudesand actions of white Americans, research in the recent past has expanded to include theeffects of context on the political judgments of racial/ethnic minorities. The theoreticalarguments advanced for all groups are similar — predominantly group conflict orcontact theory — although the specifics are left a bit vague as to why groups of suchdisparate sizes should behave in similar ways. It does seem peculiar to argue that whiterespondents who compose a majority in a county are as “threatened” by blacks asblack respondents in that same county are by whites. After all, if one considers theBlack Belt studied by Key, the threat of lynching for blacks was much more real andvivid than the threat of losing political power was for whites. Another complication

23We find in our pilot study that contact with outgroup members has no noticeable effect on people’sperceptions of their contexts; for example, whites who have contact with blacks at work, at their place ofworship, or in their circle of friends are not more or less accurate in their perceptions of how many blackslive in their neighborhood or community, compared to whites whose daily lives are more segregated.In other words, actual interactions with outgroup members do not lead to either greater accuracy ormisinformation.

29

is added by the fact that previous research has shown that Americans of all racesoverestimate the numbers of minorities in the United States and underestimate thenumber of whites in the country.24 However, what has not been mentioned is that themotivations attributed to these inaccuracies cannot be the same across groups: if whites,for example, exaggerate the numbers of blacks due to a sense of threat, then why doblacks also exaggerate the numbers of blacks? Threat cannot explain why an individualoverestimates the numbers of her outgroup and ingroup living nearby.

These misperceptions and interpretations also have policy effects. The literature onwhite flight discusses the fact that whites are unhappy and flee when blacks move intotheir neighborhoods, supporting the hypothesis of racial threat. However, Harris hasfound that blacks are also threatened by other blacks moving into their neighborhoods,a threat that cannot be motivated by fear of a racial outgroup (Harris, 2001). It is possiblethat economic threat motivates both groups’ responses, or they may have differentimpetuses; teasing out the meaning depends on understanding how different groupsperceive and interpret the same environments. When scholars study one group at atime, it is easier to find plausible theoretical explanations for the empirical findings; itis only by comparing groups that one is confronted by the more complicated picture ofhow perceptions of context can affect political judgments.

The housing literature is also relevant for another reason to the analyses presentedhere. According to Schelling’s argument about tipping points, small preferences at theindividual level can become magnified at the aggregate level, such that a person’s slightpreference for a homogeneous neighborhood leads to the dramatic levels of residentialracial segregation that exists in this country (Schelling, 2006). Furthermore, as Bruchand Mare show in their analysis of Schelling’s model, the existence of a threshold is keyto explaining how slight differences in preferences could lead, in a theoretical model,to an “apartheid America” (Bruch and Mare, 2006). While the strikingly low thresholdobserved in real life may reflect the “objective” status of the neighborhood’s make-up,it is also possible that the neighborhood is perceived as more diverse than it is in reality,and that the threshold for white flight in people’s minds is higher than the one observed.In other words, someone might be willing to live in a neighborhood that is one thirdblack, but because he overestimates the percentage of blacks living in his block group,he ends up moving away well before his real threshold has been reached. While thepractical implications may be the same in terms of policy interventions to prevent flightand hypersegregation, there may be a role for public education about the demographicsof one’s surroundings that could at least slow down flight.

24This similarity in itself is curious, given that research has shown that whites and blacks, for example,make divergent media choices, and media is a major source of information.

30

6 Conclusion: Attitudes are not Asthma

Medical researchers who study the effects of pollutants on the development of asthmameasure airborne particulates in a geographic region and the presence of asthma andother respiratory symptoms for the population living in that area. When politicalscientists study the effects of racial context on ethnocentrism and policy opinions, theirstudies mimic the pollution studies: outgroup members are the threatening allergensand attitudes are the asthma. However, in general, perceptions of context do notsurreptitiously affect one’s attitudes and actions like smog can invisibly penetrate one’slungs, even if perceptions can lead to fear (which in turn can lead to a panoply ofunconscious reactions). Parsons and Shils include this perceptual stage — Lippmann’spseudoenvironments — in their conceptualization of context, and our map-basedmeasure of context is valid for this conceptualization. Our measure of context allows usto ask questions about respondents’ self-drawn communities as compared to places thatgroup people by administrative fiat, thus sidestepping the MAUP. And, while we arebuilding on an interdisciplinary tradition of research in mental mapping, our specificuses of these maps to tap the inner states of a sample of ordinary people is relativelynew – and it is especially novel in political science.

We are not trying to generalize from the findings we present here to the Americanpublic; not only do we have a very small sample size, we recognize that the county wesurveyed is not representative of the nation. Nevertheless, the county is very near oneof the most segregated cities in the nation (Farley 1999), which makes it more likelythat we should find support for the power threat hypothesis.25 Therefore, while ouranalyses are based on only a pilot study, we believe the findings point to a strong needto “bring the person back in” to how we conceptualize and measure context.26

Our measure of context “brings the person back in” through two ways: we allow the25While perceptions of one’s context may vary depending on where one lives (i.e., many more residents

of a small island may share a perception of their community’s boundaries, compared to residents of ablock group in New York City), we expect that misperceptions and varying interpretations of context occureverywhere. Furthermore, there is no reason to assume the threat hypothesis only holds in particularlocales, even if it originated in the South; its application has been much more widespread since Key’sresearch (Taylor, 1998). More broadly, group conflict theories have been applied internationally, in smalland large settings (LeVine and Campbell, 1971).

26Again, we recognize that place may matter in ways other than perceptions. The environment inwhich people live can provide collective resources, like lower neighborhood rates of delinquency, or causeproblems, like lower rates of residential stability, independent of individual assets or limitations (Sampson,Raudenbush and Earls, 1997; Cantillon, 2006; Shaw and McKay, 1942). And, people’s perceptions ofcommunity boundaries are not always relevant; epidemiologists may determine that a particular pollutantis best studied at the neighborhood level because exposure around one’s home is what has the greatesteffect on health outcomes, regardless of how its residents conceive of their communities. While a certainparts per million of a pollutant may be dangerous to all residents of a county, there may be too muchvariation between neighborhoods for the larger aggregate measure to be useful. However, we argue thatwhen one is interested in psychological outcomes, the relevant environment will often be that defined byindividuals.

31

respondent to define the boundaries of the context and also describe the content ofthe context. However, these two aspects can also be separated and studied more indepth. For example, Kwan and her colleagues in geography have used georeferencedactivity-travel diaries to map where people travel within a city over the course of daysor weeks. If one defines a person’s context by where she spends her time and interactswith others and the environment, then again the MAUP can be avoided; the geographicunit is personalized, not arbitrary or modifiable (Kwan, 2000). This is especially relevantbecause she finds that different racial groups have different travel trajectories withinthe same diverse area; African Americans, for example, tend to move around a limitedarea within the larger city. These time-space measures are similar to those of publichealth and sociology scholars who ask respondents—often adolescents and seniorcitizens—about their abilities (and habits) for walking and driving short distances ona regular basis (Colabianchi et al., 2007). These measures aid in studying the extentof a place’s influence on behaviors like buying drugs as well as outcomes like obesity,but they could also be used to measure access to town hall meetings, for example.Other scholars have used photography to measure communities, either via photovoice(where subjects take pictures of their own surroundings) or by looking at aerial photos(Nicotera, 2007; Wang and Burris, 1997), and political scientists could also borrow fromthese alternative measures.

If pictures fixed in people’s heads do not match Census pictures, the practical implica-tions extend beyond the confines of citizens’ minds or the voting booth. The “fear ofcrime” literature in sociology has explained that personal and altruistic fear—regardlessof accuracy—leads to purchases (e.g., guns and personal safety devices), behavioralchanges (e.g. not going out at night, leaving lights on), and abandonment of locations(e.g., parks and industrial areas), particularly in metropolitan areas that are seen asdangerous (Warr and Ellison, 2000). Political scientists need to understand whether per-ceptions of community heterogeneity and interracial competition have equally seriousconsequences for political actions and outcomes.

32

ReferencesAchen, C.H. and W.P. Shively. 1995. Cross-Level Inference. Chicago: University of Chicago Press.

Adcock, R. and D. Collier. 2001. “Measurement validity: A shared standard for qualitative and quantitativeresearch.” American Political Science Review 95(03):529–546.

Agnew, J. 1996. “Mapping politics: how context counts in electoral geography.” Political Geography15(2):129–146.

Aitken, S.C. and R. Prosser. 1990. “Residents’spatial knowledge of neighborhood continuity and form.”Geographical Analysis 22(4):301–325.

Alesina, A. and E. La Ferrara. 2000. “Participation in Heterogeneous Communities.” Quarterly Journal ofEconomics 115(3):847–904.

Alesina, A., R. Baqir and W. Easterly. 1999. “Public Goods and Ethnic Divisions.” Quarterly Journal ofEconomics 114(4):1243–1284.

Anderson, E. 1976. A Place on the Corner. University of Chicago Press.

Anselin, L. 1988. Spatial econometrics: methods and models. Springer.

Ansolabehere, S., M. Meredith and E. Snowberg. 2010. “Mecro-Economic Voting: Local Information andMicro-Perceptions of the Macro-Economy.”.

Armstrong, G.B. and K.A. Neuendorf. 1992. “TV entertainment, news, and racial perceptions of collegestudents.” Journal of Communication 42(3):153–176.

Ayres, I. and P. Siegelman. 1995. “Race and gender discrimination in bargaining for a new car.” TheAmerican Economic Review 85(3):304–321.

Bartels, L.M. 1996. “Uninformed votes: Information effects in presidential elections.” American Journal ofPolitical Science 40(1):194–230.

Bartels, L.M. 2002. “Beyond the running tally: Partisan bias in political perceptions.” Political Behavior24(2):117–150.

Baybeck, B. 2006. “Sorting Out the Competing Effects of Racial Context.” Journal of Politics 68(2):386–396.

Behrens, A., C. Uggen and J. Manza. 2003. “Ballot manipulation and the ”menace of Negro domination”:Racial threat and felon disenfranchisement in the United States, 1850-2002.” American Journal of Sociology109(3):559–605.

Bhat, C.R. and J. Guo. 2004. “A mixed spatially correlated logit model: formulation and application toresidential choice modeling.” Transportation Research Part B 38(2):147–168.

Blalock, H.M. 1967. Toward a theory of minority-group relations. Wiley.

Blalock, Hubert M and Ann B. Blalock. 1968. Methology in Social Research. NY: McGraw-Hill.

Bobo, L. and V. L. Hutchings. 1994. “Race, the Sense of Group Position, and Immigration Politics.” Paperpresented at the American Political Science Association Annual Meeting, New York, August, 1994.

Bobo, L. and V.L. Hutchings. 1996. “Perceptions of racial group competition: Extending Blumer’s theoryof group position to a multiracial social context.” American Sociological Review 61:951–972.

33

Bowers, J. 1997. “Threat, Participation, and Mobilization: Crossburnings and Black Leadership in NorthCarolina, 1987-1997.”.” Paper prepared for delivery at the Annual Meeting of the American PoliticalScience Association.

Branton, R.P. and B.S. Jones. 2005. “Reexamining racial attitudes: The conditional relationship betweendiversity and socioeconomic environment.” American Journal of Political Science 49(2):359–372.

Braun-Fahrlander, Charlotte, Et. Al. 2002. “Environmental Exposure to Endotoxin and Its Relation toAsthma in School-Age Children.” New England Journal of Medicine 347(12):869–877.

Bruch, E.E. and R.D. Mare. 2006. “Neighborhood choice and neighborhood change.” American Journal ofSociology 112(3):667.

Buck, N. 2001. “Identifying neighbourhood effects on social exclusion.” Urban Studies 38(12):2251–2275.

Burbank, M.J. 1997. “Explaining contextual effects on vote choice.” Political Behavior 19(2):113–132.

Busselle, R. and H. Crandall. 2002. “Television viewing and perceptions about race differences in socioeco-nomic success.” Journal of Broadcasting & Electronic Media 46(2):265–282.

Cain, B.E., J. Citrin and C. Wong. 2000. Ethnic context, race relations, and California politics. Public PolicyInstitute of California.

Campbell, A.L., C. Wong and J. Citrin. 2006. ““Racial Threat”, Partisan Climate, and Direct Democracy:Contextual Effects in Three California Initiatives.” Political Behavior 28(2):129–150.

Campbell, D.E. 2008. “Religious “threat” in contemporary presidential elections.” The Journal of Politics68(01):104–115.

Cantillon, D. 2006. “Community social organization, parents, and peers as mediators of perceivedneighborhood block characteristics on delinquent and prosocial activities.” American journal of communitypsychology 37(1):111–127.

Carpini, M.X.D. and S. Keeter. 1997. What Americans know about politics and why it matters. Yale UniversityPress.

Chiricos, T., M. Hogan and M. Gertz. 1997. “Racial composition of neighborhood and fear of crime.”Criminology 35:107.

Chiricos, T., R. McEntire and M. Gertz. 2001. “Perceived racial and ethnic composition of neighborhoodand perceived risk of crime.” Social Problems 48(3):322–340.

Cho, C.M. and Y.S. Choi. 2005. “The Effect of Resident-Perceived Neighborhood Boundary on the Equityof Public Parks Distribution: Using GIS.” Web and Wireless Geographical Information Systems 3833:296–307.

Cho, Wendy Tam and Neil Baer. 2010. “Environmental Determinants of Racial Attitudes Redux.” Unpub-lished manuscript.

Cho, W.K.T. and C.F. Manski. 2008. Cross Level/Ecological Inference. In Oxford Handbook of PoliticalMethodology, ed. H. Brady, D. Collier and J. Box-Steffensmeier. Oxford Univ Press chapter 22, pp. 530–569.

Cho, W.K.T., D. Darmofal and N. Baer. 2007. “Environmental Determinants of Racial Attitudes Redux.”Unpublished manuscript.

Clark, W.A.V. 1991. “Residential preferences and neighborhood racial segregation: a test of the Schellingsegregation model.” Demography 28(1):1–19.

34

Colabianchi, N., M. Dowda, KA Pfeiffer, DE Porter, MJ Almeida and RR Pate. 2007. “Towards an un-derstanding of salient neighborhood boundaries: adolescent reports of an easy walking distance andconvenient driving distance.” Int J Behav Nutr Phys Act 4:66.

Conover, P.J., S. Feldman and K. Knight. 1987. “The personal and political underpinnings of economicforecasts.” American Journal of Political Science 31(3):559–583.

Converse, P.E. 1964. The nature of belief systems in mass politics. In Ideology and discontent, ed. D.E. Apter.Free Press pp. 206–61.

Coulton, C.J., J. Korbin, T. Chan and M. Su. 2001. “Mapping residents’ perceptions of neighborhoodboundaries: a methodological note.” American journal of community psychology 29(2):371–383.

Coulton, C.J. and J.E. Korbin. 2007. “Indicators of children’s wellbeing through a neighborhood lens.”Social Indicators Research 84:349–361.

Covington, J. and R.B. Taylor. 1991. “Fear of crime in urban residential neighborhoods: Implications ofbetween-and within-neighborhood sources for current models.” Sociological Quarterly 32(2):231–249.

Custovic, Adnan, Angela Simpson, Martin D. Chapman and Ashley Woodcock. 1998. “Allergen Avoidancein the Treatment of Asthma and Atopic Disorders.” Thorax 53:63–72.

Duch, R.M., H.D. Palmer and C.J. Anderson. 2000. “Heterogeneity in perceptions of national economicconditions.” American Journal of Political Science 44(4):635–652.

Feldman, S. and K. Stenner. 1994. “Threat and Political Intolerance.” Paper prepared for delivery at theAnnual Meeting of the American Political Science Association.

Fossett, M.A. and K.J. Kiecolt. 1989. “The relative size of minority populations and white racial attitudes.”Social Science Quarterly 70(4):820–35.

Gaines, B.J., J.H. Kuklinski, P.J. Quirk, B. Peyton and J. Verkuilen. 2008. “Same facts, different interpretations:Partisan motivation and opinion on Iraq.” The Journal of Politics 69(04):957–974.

Galster, G. 2001. “On the nature of neighbourhood.” Urban studies 38(12):2111–2124.

Gans, H. 1962. The urban village. Free Press.

Garling, T. and R.G. Golledge. 2000. Cognitive mapping and spatial decision-making. In Cognitive mapping:Past, Present and Future, ed. R. Kitchin and S. Freundschuh. Routledge p. 44.

Gauvin, L., E. Robitaille, M. Riva, L. McLaren, C. Dassa and L. Potvin. 2007. “Conceptualizing andOperationalizing Neighbourhoods.” Revue Canadienne De Sante Publique 98 Supplement 1:S18–S26.

Gay, Claudine. 2004. “Putting Race in Context: Identifying the Environmental Determinants of BlackRacial Atittudes.” American Political Science Review 98(4):547–562.

Gehlke, C.E. and H. Biehl. 1934. “Certain Effects of Grouping Upon the Size of the Correlation Coefficientin Census Tract Material.” Journal of the American Statistical Association, Supplement 29:169–170.

Gerber, A.S. and G.A. Huber. 2008. “Partisanship, Political Control, and Economic Assessments: Resultsfrom a Natural Experiment.” Working Paper, ISPS Yale University.

Gilens, M. 2000. Why Americans hate welfare: Race, media, and the politics of antipoverty policy. University ofChicago Press.

Gilens, M. 2005. “Political ignorance and collective policy preferences.” American Political Science Review95(02):379–396.

35

Giles, M.W. 1977. “Percent Black and Racial Hostility: An Old Assumption Reexamined.” Social ScienceQuarterly 58(3):412–7.

Giles, M.W. and A.S. Evans. 1985. “External threat, perceived threat, and group identity.” Social ScienceQuarterly 66(1):50–66.

Giles, M.W. and K. Hertz. 1994. “Racial threat and partisan identification.” American Political Science Review88(2):317–326.

Giles, M.W. and M.A. Buckner. 1993. “David Duke and black threat: An old hypothesis revisited.” TheJournal of Politics 55(3):702–713.

Gilliam, F.D., N.A. Valentino and M.N. Beckmann. 2002. “Where you live and what you watch: Theimpact of racial proximity and local television news on attitudes about race and crime.” Political ResearchQuarterly 55(4):755.

Gladwell, M. 1995. “Fundamental Ignorance About Numbers.” The Washington Post National Weekly Editionp. 7.

Glaser, J.M. 1994. “Back to the black belt: Racial environment and white racial attitudes in the South.” TheJournal of Politics 56:21–41.

Glaser, J.M. 2002. “White voters, black schools: structuring racial choices with a checklist ballot.” AmericanJournal of Political Science 46:35–46.

Golledge, R.G. and A.N. Spector. 1978. “Comprehending the urban environment: Theory and practice.”Geographical Analysis 10(4):403–426.

Gotway, Carol A. and Linda J. Young. 2002. “Combining Incompatible Spatial Data.” Journal of the AmericanStatistical Association 97(458):632–648.

Grannis, R. 1998. “The Importance of Trivial Streets: Residential Streets and Residential Segregation.”American Journal of Sociology 103(6):1530–1564.

Green, D.P. and L.M. Waxman. 1987. “Direct Threat and Political Tolerance an Experimental Analysis ofthe Tolerance of Blacks Toward Racists.” Public Opinion Quarterly 51(2):149–165.

Harris, D.R. 2001. “Why are whites and blacks averse to black neighbors?” Social Science Research 30(1):100–116.

Hatt, P. 1946. “The concept of natural area.” American Sociological Review 11(4):423–427.

Heer, D.M. 1959. “The Sentiment of White Supremacy an Ecological Study.” American Journal of Sociology64(6):592–598.

Hetherington, M.J. 1996. “The media’s role in forming voters’ national economic evaluations in 1992.”American Journal of Political Science 40:372–395.

Highton, B. and R.E. Wolfinger. 1991. Estimating the Size of Minority Groups. Technical report Memo toBoard of Overseers, National Election Studies.

Hochschild, J.L. 2001. Where you stand depends on what you see: Connections among values, perceptionsof fact, and political prescriptions. In Citizens and Politics: Perspectives from Political Psychology, ed. J.H.Kuklinski. Cambridge University Press pp. 313–340.

36

Hopkins, D.J. 2007. “Threatening Changes: Explaining Where and When Immigrants Provoke LocalOpposition.” Presented at the 2007 annual conference of the Midwest Political Science Association andthe American Political Science Association.

Hopkins, D.J. 2010. “Politicized places: Explaining where and when immigrants provoke local opposition.”American Political Science Review 104(01):40–60.

Huckfeldt, R. and J. Sprague. 1993. Citizens, contexts, and politics. In Political science: The state of thediscipline II, ed. A. Finifter. American Political Science Association pp. 281–303.

Huckfeldt, R.R. 1979. “Political participation and the neighborhood social context.” American Journal ofPolitical Science pp. 579–592.

Huckfeldt, R.R. 1980. “Variable responses to neighborhood social contexts: Assimilation, conflict, andtipping points.” Political Behavior 2(3):231–257.

Jacobs, D. and D. Tope. 2007. “The Politics of Resentment in the Post-Civil Rights Era: Minority Threat,Homicide, and Ideological Voting in Congress.” American Journal of Sociology 112(5):1458–1494.

Jacobs, J. 1961. The death and life of great American cities. Vintage.

Key, V. 1949. Southern politics in State and Nation. University of Tennessee Press.

Kinder, D.R. and D.R. Kiewiet. 1981. “Sociotropic politics: the American case.” British Journal of PoliticalScience 11(2):129–161.

Kinder, D.R., G.S. Adams and P.W. Gronke. 1989. “Economics and politics in the 1984 American presidentialelection.” American Journal of Political Science 33(2):491–515.

Kinder, D.R. and L.M. Sanders. 1996. Divided by color: Racial politics and democratic ideals. University ofChicago Press.

Kinder, D.R. and T. Mendelberg. 1995. “Cracks in American apartheid: The political impact of prejudiceamong desegregated whites.” Journal of Politics 57:402–424.

King, G. 1996. “Why context should not count.” Political Geography 15:159–164.

King, Gary. 1997. A Solution to the Ecological Inference Problem. Princeton, NJ: Princeton University Press.

Kramer, G.H. 1983. “The ecological fallacy revisited: aggregate-versus individual-level findings oneconomics and elections, and sociotropic voting.” The American Political Science Review 77:92–111.

Kuklinski, J.H., P.J. Quirk, J. Jerit, D. Schwieder and F. Robert. 2000. “Misinformation and the Currency ofDemocratic Citizenship.” Journal of Politics 62(3).

Kwan, M.P. 2000. “Interactive geovisualization of activity-travel patterns using three-dimensional ge-ographical information systems: a methodological exploration with a large data set.” TransportationResearch Part C: Emerging Technologies 8(1-6):185–203.

Lazarsfeld, Paul F., Bernard Berelson and Hazel Gaudet. 1968. The People’s Choice. Columbia UniversityPress.

Lebel, A., R. Pampalon and P.Y. Villeneuve. 2007. “A multi-perspective approach for defining neighbour-hood units in the context of a study on health inequalities in the Quebec City region.” InternationalJournal of Health Geographics 6(1):27.

Lee, T. 1968. “Urban neighborhood as a socio-spatial schema.” Human Relations 21(3):241–267.

37

Lefebvre, Henri. 1991. The Production of Space. Oxford: Blackwell Publishing.

LeVine, R.A. and D.T. Campbell. 1971. Ethnocentrism. J. Wiley.

Lewin, Kurt. 1936. Principles of Topological Psychology. New York: McGraw-Hill.

Lippmann, W. 1991. Public opinion. Transaction Publishers.

Liu, B. 2001. “Racial contexts and white interests: beyond black threat and racial tolerance.” PoliticalBehavior 23(2):157–180.

Lynch, K. 1973. The image of the city. MIT press.

Mackuen, Michael and Courtney Brown. 1987. “Poltiical Context and Attitude Change.” American PoliticalScience Review 81(2):471–490.

Matei, S., S.J. Ball-Rokeach and J.L. Qiu. 2001. “Fear and misperception of Los Angeles urban space.”Communication Research 28(4):429–463.

Merry, S.E. 1981. Urban danger: Life in a neighborhood of strangers. Temple University Press.

Nadeau, R. and R.G. Niemi. 1995. “Educated guesses: The process of answering factual knowledgequestions in surveys.” Public Opinion Quarterly 59(3):323–346.

Nadeau, R., R.G. Niemi and J. Levine. 1993. “Innumeracy about minority populations.” Public OpinionQuarterly 57(3):332–347.

Nicotera, N. 2007. “Measuring neighborhood: A conundrum for human services researchers and practi-tioners.” American Journal of Community Psychology 40(1):26–51.

Nielsen. 2010. “Harper’s Index.” Harper’s Magazine p. 11.

Nyhan, Brendan and Jason Reifler. 2010. “When Corrections Fail: The Persistence of Political Mispercep-tions.” Political Behavior 32:303–330.

Oliver, J.E. 2010. The paradoxes of integration: race, neighborhood, and civic life in multiethnic America. UniversityOf Chicago Press.

Oliver, J.E. and T. Mendelberg. 2000. “Reconsidering the environmental determinants of white racialattitudes.” American Journal of Political Science 44:574–589.

Olzak, S. 1994. The dynamics of ethnic competition and conflict. Stanford University Press.

Olzak, S., S. Shanahan and E.H. McEneaney. 1996. “Poverty, segregation, and race riots: 1960 to 1993.”American Sociological Review 61:590–613.

Openshaw, S. 1984. The Modifiable Areal Unit Problem. Concepts and Techniques in Modern Geography No. 38.Norwich, England: Geo Books.

Openshaw, S. and P.J. Taylor. 1979. A Million or So Correlation Coefficients: Three Experiments on theModifiable Areal unit Problem. In Statistical Applications in the Spatial Sciences, ed. N. Wrigley. PionLimited pp. 127–144.

Orbell, J.M. and K.S. Sherrill. 1969. “Racial attitudes and the metropolitan context: A structural analysis.”Public Opinion Quarterly 33(1):46.

Parsons, Talcott and Edward A. Shils. 1951. Toward a General Theory of Action. Cambridge, MA: HarvardUniversity Press.

38

Prysby, C.L. 1989. “Attitudes of southern Democratic party activists toward Jesse Jackson: The effects ofthe local context.” The Journal of Politics 51:305–318.

Putnam, R.D. 1966. “Political attitudes and the local community.” The American Political Science Review60(3):640–654.

Quillian, L. 1995. “Prejudice as a response to perceived group threat: Population composition andanti-immigrant and racial prejudice in Europe.” American Sociological Review 60:586–611.

Quillian, L. 1996. “Group threat and regional change in attitudes toward African-Americans.” AmericanJournal of Sociology 102:816–860.

Quillian, L. and D. Pager. 2001. “Black Neighbors, Higher Crime? The Role of Racial Stereotypes inEvaluations of Neighborhood Crime 1.” American Journal of Sociology 107(3):717–767.

Ridley, S., J.A. Bayton and J.H. Outtz. 1989. “Taxi service in the District of Columbia: is it influenced bypatrons’ race and destination.” Mimeograph, Washington Lawyers’ Committee for Civil Rights under the Law,Washington, DC .

Ross, L. 1978. “The intuitive psychologist and his shortcomings: Distortions in the attribution process.”Cognitive Theories in Social Psychology: Papers from Advances in Experimental Social Psychology p. 337.

Sampson, R.J., J.D. Morenoff and T. Gannon-Rowley. 2002. “Assessing neighborhood effects: SocialProcesses and New Directions in Research.” Annual review of sociology 28(1):443–478.

Sampson, R.J., S.W. Raudenbush and F. Earls. 1997. “Neighborhoods and violent crime: A multilevel studyof collective efficacy.” Science 277(5328):918.

Sastry, N., A.R. Pebley and M. Zonta. 2002. Neighborhood definitions and the spatial dimension of dailylife in Los Angeles. In 2002 annual meeting of the Population of Daily Life in Los Angeles. Vol. 8 RAND.

Schelling, T.C. 2006. Micromotives and macrobehavior. WW Norton & Company.

Schulman, K.A., J.A. Berlin, W. Harless, J.F. Kerner, S. Sistrunk, B.J. Gersh, R. Dube, C.K. Taleghani, J.E.Burke, S. Williams et al. 1999. “The effect of race and sex on physicians’ recommendations for cardiaccatheterization.” New England Journal of Medicine 340(8):618–626.

Semyonov, M., R. Raijman and A. Gorodzeisky. 2008. “Foreigners’ Impact on European Societies: PublicViews and Perceptions in a Cross-National Comparative Perspective.” International Journal of ComparativeSociology 49(1):5.

Shamir, M. and J. Sullivan. 1983. “The political context of tolerance: The United States and Israel.” TheAmerican Political Science Review 77:911–928.

Shaw, C. and H.D. McKay. 1942. Delinquency and urban areas. Chicago: University of Chicago Press.

Shinn, M. and S.M. Toohey. 2003. “Community contexts of human welfare.” Annual Review of Psychology54(1):427–459.

Sigelman, L. and R.G. Niemi. 2001. “Innumeracy about Minority Populations: African Americans andWhites Compared.” Public opinion quarterly 65(1):86–94.

Smith, T.W. 2001. “Intergroup relations in a diverse America: data from the 2000 General Social Survey.”New York: American Jewish Committee .

Stein, R.M., S.S. Post and A.L. Rinden. 2000. “Reconciling context and contact effects on racial attitudes.”Political Research Quarterly 53(2):285.

39

Stipak, B. and C. Hensler. 1983. “Effect of neighborhood racial and socioeconomic composition on urbanresidents’ evaluations of their neighborhoods.” Social indicators research 12(3):311–320.

Svendsen, E., L. Campbell and D. Fisher. 2008. Understanding urban environmental stewardship in NewYork City. In Proceedings of the international symposium on society and resource management (ISSRM): Pastand future. pp. 10–14.

Svendsen, E.S. and L.K. Campbell. 2008. “Urban ecological stewardship: understanding the structure,function and network of community-based urban land management.” Cities and the Environment 1(1):1–31.

Taylor, M.C. 1998. “How white attitudes vary with the racial composition of local populations: Numberscount.” American Sociological Review 63:512–535.

Taylor, M.C. 2000. The Significance of Racial Context. In Racialized Politics: The Debate About Racism inAmerica, ed. David O. Sears, Jim Sidanius and Lawrence Bobo. University of Chicago Press pp. 118–36.

Tingsten, Herbert. 1937. Political Behavior. London: King.

Tolnay, S.E. and EM Beck. 1995. A Festival of Violence: An Analysis of Southern Lynchings, 1882-1930. Universityof Illinois Press.

Tolnay, S.E., EM Beck and J.L. Massey. 1988. “Black lynchings: The power threat hypothesis revisited.”Social forces 67:605.

Turner, M.A. and F. Skidmore. 1999. “Mortgage lending discrimination: a review of existing evidence.”Washington: Urban Institute .

Tversky, A. and D. Kahneman. 1974. “Judgment under uncertainty: Heuristics and biases.” Science185(4157):1124–1131.

Tversky, B. 2000. Remembering spaces. In The Oxford handbook of memory, ed. E. Tulving and F. Craik.Oxford University Press pp. 363–378.

Voss, D.S. 1996. “Beyond racial threat: Failure of an old hypothesis in the new South.” Journal of Politics58:1156–1170.

Wang, C. and M.A. Burris. 1997. “Photovoice: Concept, methodology, and use for participatory needsassessment.” Health Education & Behavior 24(3):369.

Warr, M. and C.G. Ellison. 2000. “Rethinking Social Reactions to Crime: Personal and Altruistic Fear inFamily Households.” American Journal of Sociology 106(3):551–578.

Weinmayr, Gudrun Et Al. 2007. “Atopic Sensitization and the International Variation of Asthma SymptomPrevalence in Children.” American Journal of Respiratory and Critical Care Medicine 176:565–574.

Weston, L. and S. Handy. 2004. “Mental maps.” Handbook of transport geography and spatial systems 5:533–545.

Wilson, W.J. 1987. The truly disadvantaged. University of Chicago Press.

Wissoker, D.A. 1997. Testing for discrimination in home insurance. The Urban Institute.

Wong, C.J. 2007. “” Little” and” Big” Pictures in Our Heads: Race, Local Context, and Innumeracy aboutRacial Groups in the United States.” Public Opinion Quarterly 71(3):392.

Wong, C.J. 2010. Boundaries of Obligation in American Politics: Geographic, National, and Racial Communities.Cambridge University Press.

40

Wright Jr, G.C. 1977. “Contextual models of electoral behavior: The southern Wallace vote.” The AmericanPolitical Science Review 71:497–508.

Yinger, J. 1997. Closed doors, opportunities lost: The continuing costs of housing discrimination. Russell SageFoundation Publications.

Yule, G.U. and MG Kendall. 1950. “An Introduction to the Theory of.” Statistics .

41