ob, pp, iro-20'i copv-righi © 2008. bl- rural sodulogjcal...

OB, pp, IRO-20'ICopv-righi © 2008. bl- ihc Rural Sodulogjcal Sotici^

Political Socialization and Reactions to Immigration-Related Diversity in Rural America*

James G. GimpelDepartment of GovemmentUniversiiy of Maryland

J. Celeste LayDepartment of Political ScienceTulane University

ABSTRACT We explore the roots of lolerance for immigration-relateddiversity from a political .socialization perspective. Among rural adolescentrespondents, we find that attitudes toward immigrants are stirprisinglyvariahle along a number of important dimensions: anticipated socioeco-nomic status, family longevity in the community, and employment inagriculture. The extent to which an adolescent's family is anchored in thecommunity proves to be an important determinant of diversity attitudes.Tolerance for diversity is also contextually conditioned by the percentage ofimmigrants settled in a tieighborhood. and the percentage of the localpopulation employed in farming. Interestingly, lower income youth aremore welcoming of immigration than the affluent, particularly when theylive near thetn. Without quite labeling these laiial adolescent populationsracially "progressive," the youth we encountered mostly expressed thenorms of tolerance and civility essential for avoiding unpleasant intergroupconflict.

\\Tiere we do find the roots of diversity attitudes in rural America? Doesthe recent arrival of immigrants in rural commtinilies generate lessethnic hostility and more tolerance for diversity than one would fmd inotheiTvise similar communities not touched by immigration?

Immigration-related diversity is new to most rural area.s outside thesouthwestern United States. The arrival of Latino and Asian immigrantsin the Midwest and South has heen closely linked to the long termeconomic restrticturing of rural industry and agriculture (Baker andHotek 2003; Kandel and Cromartie 2004; Stull. Broadway and Griffith1995;). The current generation of youth are front and center of thisnew societal experiment. In adult life, occupational and housingsegregation ensure limited contact with immigrants, though thiscontact is arguahly more frequent in small towns with immigrantpopulations than it is in large cities. In schools, however, the immigrantpresence is a constant and cannot be avoided by natives through

* Direct correspondence to: Jim Gimpel. University of Marvland. Depanment ofGovemnieni, 3140 Tydiiigs Hall, College Park. Man'land 20742, jgimpel@gv{)t.umd.edu.

Reactions to fmmigration-ReUited Diversity — Gimpel and Lay 181

discriminatory practices. This makes adolescent populations anespecially informative test of how rural citizens are adjtisting to thenewcomers.

For the increasing ntimber of locales experiencing immigrant influx,the social and economic conditions in which younger generations arebeing raised are veiy different from the circumstances thai prevailed inthose same locations for previous generations. That conditions havechanged so dramatically provides us with an excellent reason for posingqttestions about tolerance and openness to immigration to youngpeople who; 1) find thetnselves in regular contact with immigrantyouth; and 2) are also being socialized by elders who influence theiropinions and shape their biases. Although cultural and politicallearning does not end when a person reaches 21, there is abundantevidence to support the conclusion that what is learned duringadolescence is highly predictive of adult behavior and opinions (Alwinand Krosnick 1991; Beck and Jennings 1982; Green and Palmquist1994; Jennings and Markus 1984; Plutzer 2002). Studying the views andopinions of the curtent generation of youlh may therefore providesome glimpse of what lies Just around the corner for rural communitiesfacing the pres.sures of globalization and ethnic change.

Group Threat and Diversity Attitudes

The research we report foctises on the political socialization ofadolescents (between 14 and 18 years of age) in understudied ruralareas in two states: Iowa and Maryland. We are most interested in thediversity attitudes children leam from the socialization agentsproximate to them, embedded in their local communities (FalkSchulman and Tickamyer 2003; Gimpel, Lay and Schuknecht 2003).The socialization perspective on adolescent development suggests thatchildren are raised within a specific strtictural context, a local socialenvironment, tliat influences what they leam and the identities tlieyadopt. Geographically constrained interactions structure the informa-tion flow on the basis of which individuals develop their posture towardethnically distinct populations, including their own, as well as relatedbeliefs about equality and morality- (Greene, Way, and Pahl 2006;Hughes et al. 2006; Phinney and Chivara 1995; Stiarez-Orozco andSuarez-Orozco 2001). Moreover, socialization experiences are nottmiform within a society but are shaped by local socioeconomic atidpolitical circiunstances—conditions which can vary at a surprisinglygranular geographic scale—and shape the viewpoints of both adultsand children alike. Most children are not highly mobile, and their

1S2 Rural Sodology, Vol. 73, No. 2, June 2008

socializing experiences occur within a limited number of .settings(Gimpel, Lay. and Schuknecht 2003).

Given that the respondents to our survey are adolescents, we wouldnot expect youth attitudes to perfectly mirror those of the local adultpopulation. For instance, in the formation of adtilt attitudes towardimmigrants, economic and status insecurities have been found to beconscious considerations. Prejudice and intolerance are commonlyunderstood to be the product of competition for economic benefits(Bonacich 1976; Cummings 1980; 1977; CXimmings and Lambert 1997;Olzak 1992; Scheve and Slaughter 2001), positional benefits (Beggs,Villemez, and Arnold 1997; Horowitz 1985; Lipset and Raab 1973), or acombination of both (Olzak 1992). These works connecting groupthreat to prejudice portend a difficult adjustment process in ruralAmerica, even without the aggravating circumstances of industrialrestructuring and rootedness that are present in many small towns.

Even those wlio have suggested that contact amung rival grotips isproductive of friendship and peaceful intergroup relations say that thisis the most likely outcome only when those groups are rotighly of equalstatus (Allport 1954; Dovidio, Gaertner, and Validzic 1998; Jackman andCrane 1986; Sigelman et al. 1996; Sigelman aiid Welch 1993; Welchetal. 2001). True, small towns have characteristically narrow incomedistribtitions, indicative of greater equalit)- than exists in metropolitanareas (Lay 2006). But in spite of this greater leveling, there are very fewrural areas where the immigrant and native bom populations are trulyon an equal footing. There are sound reasons to believe, then, thatrural areas could contain the seeds for interethnic conflict.

As far a.s yotith are concerned, however, theories of grotip threat mayapply only to attitude formation among a small subset of the teenpoptilation. Most adolescents are not in the labor market, and if theyare, it is tisually only for part-time employment. Immigrants are unlikelyto pose the same direct economic threat lo a high schooler's economicwell-being that they might for her parents or grandparents. But theoriesof group threat are stirely relevant to predicting adolescent diversityattitudes. Certainly children may be concenied on behalf of parents orkin who are losing ground in the competition for economic goods.From this standpoint, one might theorize that anti-immigrant attitudesoriginate from broader family- or groupnjriented calculations ofinterest, rather than from a personal perception of threat

Threat may also he an anticipated rather than experienced directly.Prejudice anchored in economic insecurity is not likely to be theforemost sotuce of anti-immigrant bias among adolescents who areupwardly mobile. But it is reasonable to hypothesize thai there will be a

Inactions to !mmigration-Related Dix)ersily — Gimpel and iMy IH3

negative evaluation of immigrants among those students with lesseducational amlntion—who have no plans to continue their formaleducation in college. Low achieveis may perceive that their most likelyemployment opportunities are going to be in labor market niches thatlow skilled immigrants could also fill. Academic aspirations after highschool could be a critical explanatory variable in predicting the extentof current anxiety about diversity.

Wliile we evaluate whether economic conditions and immigrationattitudes are linked, we are mindful of botb economic and noneco-nomic influences on intergroup relations. Tolerance may be a functionof each grottp's position and size in the commtinity, even wlien Jobs andother economic goods are plentiful (Bltuner 1958; Bobo and Hutchings1996; 1965; Smith 1981; Taylor 1998). Intolerant attitudes have beenfound among white populations in sparsely populated mral areas whereblacks, immigrants, or other ethnic "rivals" do not constitute a credibletbreat to persona! economic well-being (Glaser 1994; Oliver andMendelberg 2000; Voss 1996). Personal economic circumstances, then,may be irrelevant in accounting for the levels of prejtidice or for raciallyrelevant political choices in some places (Kinder and Sears 1981:427).Given the existence of intolerance in communities isolated fromdiversity, ecological conditions qtiite apart from proximity to potentiallyrival groups must play some role in generating diversity attitudes. Inrelated research conducted in Iowa, even white ethnic diversity wasassociated with greater suspicion, regardless of socioeconomic status(Rice and Steele 2001).

Prejudice and Rootedness in the Past

A number of studies have shown that prejudicial attittides \'ary by agecohort; with far lower levels of prejtidice expressed among the veryyoungest age cohorts, compared witli tbe very oldest (Schuman, Steeh,and Bobo 1985). Inasmuch as older generations harbor out-dated andunyielding valties on matters such as race and ethnicity, one certainsource of diversity attitudes is the relative rootedness of communitiesand the volume of older voices in the opinion dynamics of thoselocales. If the mral population were more highly mobile, rtiral life andattitudes would take on a more urban or subiuban quality, namely theywould be more fluid and less shaded by the Judgments of previousgenerations. Rural poptilations are resistant to change, one mightargue, because they have stronger anchorage to the unbending valuesand beliefs of past generations. And grandparents take a far more activerole in parenting grandchildren in mral America than in urban areas

184 Rural Soaohgy, Vol. 73, No. 2, June 2008

(Elder and Conger 2000; Ramirez-Barranti 1985). For youth, thisintergenerational connection is a cost-cutting source of information, animportant font of learning, and a source of attitude consLraint. Withoutquestion, the values passed down across two or more generations areoften an important cultural resource and a positive force in the lives ofchildren (Elder and Conger 2000). But there may also be someunhelpful or retrogressive elements in this package—perhaps a darkside to the rich social capital one finds in small towns.

An important emerging literature has found that elders do pass alongmessages in the form of cautions and admonitions abotit interactionwith other racial and ethnic groups. These socializing communicationspromote mistrust, social distance, and suspicion in relationships withthose wlio are ethnically or racially dissimilar. Apparently, small butnontrivial subsets of immigrant and African American youth are alsowarned by their elders about the undesirable characteristics of specificgroups (Hughes etal. 2006). The presence within a community of oldergenerations as vocal socializing agents stistains traditional forms ofdiscriminator) thinking, dating to a time when open expressions ofconsciotis, intentional, prejtidice were more widely accepted. Frequentenough exposure to these expressions introduce youth to olderstereotypes ihat sustain intolerance and convey the impression thatavoidance of Latinos, Asians or other immigrants is preferable oracceptable.

A Diversity of Views

We do not expect lo find the iiiral response to immigration to bemonolithic, even within the limited setting of our study. Consistent withthe notion that rural areas are not all equal, the rootedness of ruralpopulations is also subject to variation. WTiile there is no qtiestion thatmral poptilations show more residential stability tlian suburban ones,not everyone in rural America is a long-time resident with kinship tiesextending back two or more generations. Because housing is cheap andmral areas are considered havens from crime, pollution and trafficcongestion, small towns do attract new residents, and not all of themare immigrants seeking low-skill jobs. For many metropolitan fringelocations, such as three locations we studied in Maryland, and two inIowa, urban areas are accessible, particularly with modern transporta-tion and commtmications networks. We should expect some variationwithin the rural population in its extent of rootedness.

Aside from rootedne.ss in the community, one might speculate thatthe unique nature of loiral economies contributes to conservative

Reactions to Immigration-Related Diversity — Gimpel and Lay 185

political views on many matters, including ethnic diversity. To theextent that agriculture does remain a critical aspect of rural economies,the seasonal nature of this business, and the associated irregularity ofincome, contributes to a frugality and economic conservatism thattranslates into political and economic attitudes.

Finally, small businesses and entrepreneurship aie strong character-istics of the rural economy, agricultural or otherwise, that maycontribute to political conservatism. Entreprenetirs are in control oftheir enterprises, so they are more likely to be threatened thanbenefitted by a more activist regulatory regime, or one that comes withthe expectation of higher taxes. Wage workers in urban industries, onthe otlier hand, have no quarrel with more regulation, because it isexisting conditions that they find dissatisfying. We hypothesize, then,that rural conservatism may rest in self-employnwnl znd m JuU-time fanningmore than in other types of occupations found in the mral economy.

In summary, these foregoitig considerations from establishedresearch lead us to test several h)'potheses. First, consistent with thepower-threat theory (Blalock 1967). we anticipate that hostility toimmigration-related diversity is likely to be greatest among adolescentsfrom families utith lower incomes^ as well as among youth with lowereducational aspirations and consequent lower futtire income-earningpotential.

Second, rootedness in the cotnmunity, as measured by having multiplegenerations residing in the same location, is likely to be associated withconservative political values more generally, including resistance todiversity. In connection with this hypothesis, we further suggest thatthose whose derisily of acquaintanceship in the community is high will bemore resistant to diversity than tho.se whose local networks are not wellformed and tho.se who report knowing fewer community residents.Third, we are likely to find greater opposition to diversity amongchildren of farmers viho are likely to cling to more conservative viewpoints.Finally, we hypothesize that gender matters to diversity attitudes. Thismay have to do with the fact that mral hoys are more likely to anticipateassuming the breadwinner role as adults than girls and may fear greatereconomic competition in the workforce than girls do. In some ruralworking-class households, boys grow-up to enter labor market niches,such as home construction, or mral industry, in which immigrants havea presence. Girls will often pursue two-year degrees at local communitycolleges, then take positions as office assistants or bookkeepers thatrequire .skills or education that many immigrants will not have. Somerural male children have also had greater exposure to the workplace,commonly in family businesses, where they may have encountered an

186 Rural Sodology, Vol. 73, No. 2, June 2008

immigrant presence or heard parents or other adults complain about itin a work setting.

Reflecting the broader theoretical aims of our work, we test for theinfluence of community or "ecological" variables on diversity attitudes,believing that individual opinions are the product of social interactionand neighborhood characterisdcs that cannot be captured by recordingindividual traits alone. In brief, some locations are more likely topromote egalitarian views of interethnic relations than others. Wespecifically examine the impact of immigrant presence in theneighborhood, the percentage employed in farming, and the incomeof neighborhoods. By controlling for both the contextual andindividual level influences on local attitudes, we are likely to obtainmore acctirate estimates of each force acdng to shape attitudes towarddiversity.

Research Setting

The design for otir research was shaped principally by a desire to testthe straightforward notion that rural areas are more diverse than manysocial scientists and non-experts believe. We focus on schools andcommunities at the lower end of the population distrihution. If there isno variation in the way rural American youth respond to questionsabout diversity, then we are likely to find very little statisticalsignificance in any of our hypothesis tests.

We are far from the first to suggest that political socialization wil! notbe uniform within a society, but viill vai y acnxss communities (Conoverand Searing 2000; Gimpel, l.ay, and Schuknecht 2003; Sanchez-Jankowski 1986; Litt 1963; Levin 1961). But the idea that we can takea sample exclusively from small towns in a limited geographic area andfind significant differences along a number of key continua, defiesmany stereotypes about loiral populations and the complexity ofdiversity opinions in rural areas.

In two cross-sections, in which a lai'ge share of the respondents weresuiTeyed twice, we surveyed nearly 2,450 high school students in fiveIowa and four Maiyland communities. Locations were selectedpurposively because they repre.sent diverse settings within these states.Specifically, the Iowa and Maryland locations also present a very helpfulmix of elhnic heterogeneity and homogeneity. Two ofthe Iowa towns,

^ The first cross-section of students was surveyed in Ociober-November of 2001, and thesecond cross-section was sur\('yed in Apiil-Mxy 2002. for a total of 2,439 respondents.There was considerable overlap between the two cross-sections, as approximately l.lfiOstudents were capttired in both surveys, producing a two-wave panel.


Perry and Storm Lake, have a large immigrant presence. The WesternIowa towns of Carroll and Harlan are predominantly white, althotigbresidents in each town are clearly conscious of the immigrantpopulations in nearby towns. The town of Boone, Iowa, the largest ofthe set, has a small immigrant population.

In Mai-yland, the Garrett (bounty location is the most insulated fromdiversity and represents a very remote rural setting in the state's farwestern Appalachian region. The towns of Elkton and Havre de Gracehave small immigrant populations, but long-standing and establishedAfrican American populations, approaching as high as 20-25 percent inthe local .school system K-12 (see Table 2). The North Harford locationis predominantly white, akhough close enough to Bakimore that outer-ring suburbs are beginning to encroach upon it, bringing near someAsian American and African American diversity. Elkton and Havre deGrace, arguably, are also in the midst of rural-to-suburban transition,though both would he considered small towns rather than suburbs.

The participating schools ranged in size from 500 to about 1,400students, but more importantly, none ofthe towns in which the schoolswere located were larger than 10,000 people. To the extent that theserural populations did vary, they did so in two primary respects: 1) tbeextent to which they were presented with population diversity in theirmidst, and 2) the extent to whicb their local economies andpopulations were either expanding and growing, remaining stable, orcontracting. Some towns exhibited new diversity', as represented bynewly arriving immigrant populations that had .settled in just theprevious ten years. At least two towns exhibited old diversity—asrepresented by longstanding sizable minority populations, primarily ofAfrican American ancestry. Finally, three towns were havens of low tono diversity—the towns exhibiting a longstanding native-born popula-tion that remains tnore than 97 percent white.

As for the schools, the immigrant-receiving school systems provideone-year of intensive compensatory English language education forimmigrant youth if they are evaluated as deficient in English, but everyeffort is made to keep them integrated with the native-born population.Many parents do not think much of taking kids out of .school for weeksor months to travel back to Mexico or Centra! America. Transiency is anactive choice of some parents, althotigh the Latino population is aisosettling down as more immigrant families have moved into the area toreplace the single men who once dominated the industrial workforce.

School officials admitted that in the eiu"lier years of immigration, thenative and foreign-born populations did not readily mingle, but tbis haschanged with time, as the immigrant youth decided to participate in


sports and other extracurricular activities that provided a foundationfor social interaction and acclimation. Soccer, for example, is a popularand traditional sport in Spanish-speaking countries. In Perry, 80percent of the players on the soccer team are Latino. But the Latinoyouth also eventually tried-out for football and wrestling and after a fewyears were regular participants on those teams as well. Interethnicdating eventually developed. In Storm Lake, a teacher reported that amajor breakthrough in ethnic relations occurred when the immigrantyouth came dressed in the school colors the day of the homecomingfootball game. This was an important signal to the native-bom youththat the newcomers wanted to be a part ofthe school, to be supportive,and partake of school spirit.

Even so, tlie majority of the immigrant high school students are noton an academic, college-prep, track. They are more likely to minglewith the native-born "have-nots" than the "haves." According to theestimates of one school counselor, fewer than five percent are headedfor college, compared with about 50 percent of the native born. Theirlack of college aspiration influences the courses they take and theamount of eftbrt they put into academic studies.

Sampling Methods and Representativeness

While probabilistic or random sampling methods are preferable tonon probabilistic ones, there are many circumstances when it is notfeasible to randomly sample. Our sample of student respondents couldnot be randomly drawn from school enrollment lists becausecooperating school administrators considered such data collectionprocedures to be entirely too disruptive of the school day. By choosingto survey entire classrooms of varying achievement levels along with verylai ge proportions of the student population at each school, we wereconfident that we could represent the total school population, as well asthe local rural population along key dimensions such as race,socioeconomic status, parents' partisanship, and parental occupation(farm vs. non-farm) (see Table I). By making a Spanish language formavailable, we were also able to include immigrant youth with limitedEnglish proficiency. The vast majority of these youth were born outsidethe United States, chiefly in Mexico.

Nonprobability or purposive methods do not nece.ssarily fail torepresent sampled populations. Comparisons of the sample, school,and broader community populations by race/ethnic characteristics andfarm population are presented in Table 1. The community populationfigures are calculated by aggregating 2000 U.S. census data for the


Table I. Representativeness of Sample and Local Populations fromRural Civic Values Survey, 2001-2002, Unweighted Data

Iowa Schools

Boone Sample N = 162Boone PopulationBoone High School =868Carroll Sample N = 235Carroll PopulationCarroll Hifrh S<:hool = 579Harlan Sample N=205Harlan PopulationHarlan High School = 6l2Perry Sample N=222Perry PopulationPerry High School=564Storm Lake Sample N = 266Storm l^kf PopulationStomi Lake High

School = H39

Maryland Schools

Elkton Sample N=100Elkton PopulationElkton High SC:1K>O1-93HGarrett Sample N=8«Carrett PopulationCarrett High School = 820Havre de Grace Sample

N=109Havre de Grace PopulationHavre de Grace High

School=7t7North Harford Samplt'

N=128Nortli Harford PopulationNorth Harford High

Sdiool= 1.430

Farm*

4.54.7-8.08.1-8.4

10.4-

13.64.1-

10.68.2

-

Farm*

7.3.6-

3.52.2-

.'i.8.0

-

2.41.1

-

White

96.997.998.598.798.398.497.698.098.466.775.276.672,078.2

66.0

White

76.889.779.097.698.499.fi

68.982.9

73.0

95.295.4

96.8

Black

.6

.3

.1

.4

.2

.50.0

.102.31.02.10.0

.3

1.0

Black

11.15.0

15,01.2

.5

.3

17.911.4

21.0

.82.3

1.6

Asian

0.02.8.4.3.2.5.3.2

2.3.5

2.013.03.9

12.8

Asian

2.01.0l.O1.2.2.1

0.01.5

1.4

1.6.5

.5

Latino

.6

.8

.30.0

.5

.90.01.01.3

22.414.519.312.611.1

20.0

Latino

4.02.04.00.0

.4

.1

3.81.5

4.0

.8

.8

1.0

NativeAmer

.6

.2

.2

.4-1

0.0.5.3.2

2.32

0.0.•I1

.2

NativeAmer.

1.0.3

1.00.0

.0

.0

0.0.3

.6

0.0.2

.1

Biracial

1.3.5-

0.0.5-1.5.5-

3.21.1-

1.91.3

-

Biracial

• \ \

1.4-

0.0.3-

9.41.8

-

1.6.7

-

* Sample reflects percentage reporting thai both parents arc employed full time infarming.

Populadon Figures are calculated from census tracts and block groups corresponding tothe high schools' catchment areas. For school pupitlation figures, cells are left blankwhere schools do not collect or report this information.

entire population (adults included) from census tracts and blockgroups approximately within the schools' catchment areas. Thesefigures show that the proportions in our samples are not far off. Wehave slightly oversampled the Asian and Latino populations of Perryand Storm Lake, Iowa, and oversampled African American and biracial


populations in Havre de Grace and Elkton, Maryland. It would alsoappear that we probably included a greater proportion of students inour sample whose parents are employed in farming than ihe underlyingcommunity population proportions would dictate. Of course, theproportions represented in the school-ba.sed samples may better reflectthe underlying farm population with school-age children than the totalemployed in farming, which provides the basis for the populationfigures in Table L Comparing the samples to the high schools'poptilation revealed close matches between school and samplepopulations for four criterion variables: percent white, percentHispanic, percent black, and percent Asian and Native American, alsosummarized in Table 1. The slight oversampling of some subgroupsand undersampling of others is nol a great concern in survey researchbecause most samples are adjusted by weighting the data to obtain theappropriate sizes of sample subclasses (Kish 1987), and oversamples ofminority populations prove conducive to generating stable parameterestimates in statistical analysis.

As for the external validity of our study, or the generalizability of thefmdings, we are appropriately cautious. While our sampling strategysucceeds in representing the communilies whose youth participated inthe research, we do not imagine that these results apply to otherlocations, or to die same locations at other points in time, much less dieentire nation. Campbell and Stanley (1963:17) point out tbat theextemal validity of research fmdings is never finally setded since itinvolves extrapolation to subjects not in one's sample. NevertJieless, wedo believe ihat nature is "sticky," in that similar studies should retumcomparable results. Only the conduct of additional studies in similarsettings, however, will reveal the extent to which our findings holdacross unexamined locations.

A Multilevel Approach

Because we have individual level survey respondents nested withinneighborhoods which we can measure at the census tract level of^gregation, a hierarchical data analysis procedure is ideal forestimating the effect of both neighborhood and individual character-istics on diversity attitudes. Social scientists and statisticians havedeveloped a number of regression-related methods specifically tailoredto the analysis of multi-level data. These innovalions have been pursuedbecause of the special statistical challenges of data analysis whenindividual observations are clustered within geographic jurisdictionsrather than independently distributed. Standard regression approaches

Reactions to hnmigration-Related Diversity — Gimpel and Lay 191

assume that observations are independent of each other in space,clearly a faulty assmnption in our case, becau.se students hi particularneighborhoods interact with each other and share the same condition-ing environment Given our theory of the spatially contingent nature ofthe socialization process, a student living among immigrant andminority populations will look more like another student living in thatsame area, than they will look like a student from a census tract on theother .side of the county housing no minority population. The spatialclustering of students in particular communilies necessitated methodsthat would not require the strong assumption that the surveyobsei'vations be independently distributed in space.

Hierarchical Linear Modeling (HLM) is employed to examine ourhypotheses (Raudenbush and Biyk 2002; Lee and Biyk 1989;Steenbergen and Jones 2002). Given the substantive aims of ourresearcb, we will only lay a foundation for an intuitive understanding ofthe method in ihe body of this article. Briefly, we use HLM in order toevaluate hypotheses about how our neighborhood level characteristicsinfluence relationships at the individual level. So, for example, if wehypothesize that poor respondents will come to think differently aboutthe value of immigration-related diversity' if they live near immigrantsrather ihan in homogeneously white areas. HLM is a tool designedexplicitly to lesl such hypotheses.

Ifwe were studying adolescents all from a single neighborhood, HLMwould not be necessary because the neighborhood or "environment"would not vaiy. Ordinary regression techniques could be employed inassessing the outcomes of interest. Ifwe were working with data drawnfrom multiple sites, but we had no theoretical expectation thatneighborhood characteristics would matter to the individual levelrelationships, then HLM would not be necessary either. Because ouriheon- suggests thai neighborhood characteristics may matter, andbecause we are working witti dala drawn from many distinct ruralneighborhoods, we reached the conclusion that HLM was a method wecould not do without.^

^ HLM does demand a lot of the data in that it requires that each neighborhood have asiiHiciem number of individual oh.sei\atians to permit the estimalion ot separateneighborhood slopes and intercept.*;. Unlike ordinar\' regre.ssiun, it will exclude thosencighborhood.s that do nol contain a siiflkient niiml)er oi indiwliial cases. Given thegt'ographir dispersifin of the niral puptilaiion, this limitation could hinder a completeexamination of the data from our most rtiral locations. Wlien we considered ihis costalongside alternative methods of estimation reported in previous versions of this paper,we found that il made littk' stibstantive ditlercnce to the rcsulLs. Alternative results fromstructural equation estimation (SKM) are avaiiabie upon rt-qiiesi of the authors.


Results

Table 2 presents results from an estimation of a hierarchical linearmodel that evaluates the interaction of individual with neighborhoodcharacteristics. Responses are pooled from bolh the Spring 2002 andFall 2001 surveys, and a conuol variahle is added lo gauge the effect ofsurvey timing. The model in Table 2 includes a measure of incomederived from students' descriptions of their parents' occupations,resulting in a continuous measure of income (see appendices forquestion wording and responses). The correlation between this incomemeasure and siudeni self-reports of their parents' income wasmoderately positive, but statistically significant (r=.33; p^.OOl).^ Theoccupation-based measure was judged to be superior given thatstudenLs are characteristically ignorant of ihe incomes of their parents,and wind up guessing much of the time when asked.

Of particular interest is the way in which the relationship betweenfamily rootedness and opposition to diversity changes from place-to-place in intcraclion with the local size oi" the foreign-born population.In locations with virtually no foreign-boni presence, we see substantialconsensus in diversity attitudes, regardless of how well rooted therespondent's family is in the community. But in communities with largeforeign born populations, the coefficient (7=.012) indicates that asubstantial gulf emerges between the well-rooted and more recentarrivals as the foreign born population increases (see Figure 1).Specifically, the well-rooted express far greater reservations aboutdiversity since the slope for family longevity is steeper as the foreignborn population increases, conti'oUing, of course, for the immigrantparentage of the individual respondents. Specifically, there is nearly a20 percentage point difference in diversity attittides between those atthe 25' and those at the 75" ' percentiles of rootedness in locations withthe most significant immigrant concentration (see Figure 1). Butlocations with few to no foreign born residenLs exbibit a far smaller twopoint difference of opinion across this same range of rootedness(Figure 1). The model coefficients show that anti-immigrant currentsare highly related to rootedness in the community at those locations toherethe immigrant fnesence is most x'isihle.

Opposition is also much greater among the well-rooted in the ruralfarming communities {7=.006) than it is in small towns where fanning

The differences in ihe resulting estimates from twitig both income measures were notvast, but were sufTiciently different to warrant menlion in this footnote. We utilize thecontinuous measure of income in these resulLs. The resulLs from estimaiion with thestudent self-reports of parental income are availahle upon request of the authors.


Table 2- Two Level Model of The Influence of Foreign Bom Contexton Opposition to Diversity among Rural Adolescents

Fixed Effects

Locality Means (70)

Family Longevity (71)

Inverse of Income (Poverty) (7a)t

No College Plans (73)

Male Respondent (y,,)Parents lmmi(j;rants (71 )Pareul.s Farmers (75)Recognize Me: Ever>one (y?)Recognize Me: Hardly Anyone (•/«)Frequency of Discussion (7i>)Plan to Attend 4 Year College (710)April 02 Survx'v (711)N

Random Effects

Locality MeansFamily LongevityNn O)IIege PlansInverse Income (Poveny)Level 1 RProportion Reduction in Variance

Explanatory Variable

InterceptState. (O=MD. 1=IA)Percent FarmPcrrc-nt Foreign BomFamily Income (Z)InterceptPercent FarmPercent Foreign BomFamily Income (Z)InterceptPercent FarmPercent Foreign BomFamily Income (Z)InterceptPercent FarmPerccni Fnreign BomFamily Income (Z)InterceptInterceptInterceptInterceptInterceptInterceptInterceptIntercept

Pooled 2001-02 Survey

y

54.570**-.224-.523**

.187- .078

.OCR

.006**

.012**

.003

.074

.004- . 0 4 7 * *- . 8 4 6

.374

.603**

.209*-.84t>2.281**

- 5 . 4 1 9 * *.822*

2.050**

-.471- .309-.709

-2.214**

SF-Y

3.0371.627

.2.'^3

.3861.861.016,002.002.014.094.012.016

1.0211.847.156.117

1.021.466

1.196.499

1.063I.0S4.200

1.405.443

Variance Component/Z2 Sig.

39.90.001

19.18.0. 9

286.115.5%

.001

.00.3

.001

.007

Hieiarchical Linear Model; Slopes and Intercepts Estimation.cell entries in lop portion of the tahle are regression coeflicients in the first coltimn,

with standard error's in the second column.*p<.10; **p<.05.

is not a major source of employment (Figure 2), though in the mostagricultural locations, the well-rooted rank only 11 percent higher inopposition to diversity, across ihe interquartile range, than those withless tenure. There is some foundation, then, for helieving that well-rooted fartn families pixiduce offspring with less inviting instincts.

194 Rural Sociology, Vol. 73, No. 2, June 2008

65

50

O.I 10.5 15 7

% Foreign Bom

Figure 1. Opposition to Diversity Increases atnong Well-Rooted Respondents inl i n t Receiving Areas

Notably, we do not find that rootedness yields greater or less oppositioncontingent on neighlxtrhood income levels, as thai coefficienl (y —.003) isstatistically indistinct fnnn zero.

Do we fmd that the family's poverty has an effect on diversity attitudescontingent upon neighborhood context? Yes, hut what the data show isthat lower income youth are actually more positive about immigrantswhen they live in closer proximity to them {y=-.O47). Undoubtedlythis points to the effect of contact with immigrants on eqtial terms thattakes place in high school classrooms and neighborhoods where lowerincome native and immigrant youth interact.^ The familiar finding thatgreater lolerance emerges on the basis of equal-status contact receivessome support from these findings.

But there may also be evidence that group economic threat is alsopresent. The other results in Table 2, including the control variables,suggest that those with low educational aspiration.s after high scliool areslightly more hostile to immigration if they live in immigrant-receivingareas than when they do not, although the effect is not substantively

Each community had only a single high school out of wliich respondents werestirveyed, so we are unahle to separately isolate the influence oi" the school setung fromIhe community in which it is located. School ctirricula did not rary significantly byconuntmity, afthough there were differences between Ioi*'a and Mai-yland schools.

Reactions to Immigration-Related Diversity — Gimpel arid Lay 195

5.8 11.5 173 23.0 20.8

% Farm Population

|—Well-Rooted —More Recent [

Figtire 2. Opposidon to Diversity Increases among Welt-Rooted Respondents in.•\gHcuttural Area.s

large, and is of marginal statistical significance. Our data also indicatethat those yotith whose parents are full time farmers are more reservedabout diversity than those whose parents ate in non-farm occupations.Similar to our results for the extent of rootedness of families in acommunity, a child's density of acquaintanceship is a source of negativeevaluations of diversity. Students who reported that nearly everyonewould recognize them on the street were about 2 percentage pointsmore wary of diversity than those who were less .sure about their abilityto be recognized. This is a similar effect to being male versus female,and a greater effect than is associated with having parents in fanning.Finally, we found that the timing of the survey mattered. Specifically,the Spring respondents were more accepting of diversity than those inthe Fall—perhaps the effect of learning.

In a final effort to gauge the effects of context on attitudes towarddiversity, we present the results in Table 3, showing the effect of familylongevity in the community, controlling for previous opposition todiversity, on attitudes toward diversity exhibited in the Spring 2002survey. Included in this analysis are only the 889 students who wereincluded in both surveys—their responses constitute a two-wave panel.

Unsurprisingly, what we see is that most of the variation in attitudestoward diversity is absorbed by the previous values of the same variable

196 Rural Sodobgy, Vol. 73, No. 2, June 2008

Table 3. Two Level Model of The Influence of Foreign Bom Contexton Opposition to Diversity Among Rural Adolescents, Controlling forPrior Diversity Attitudes

Fixed Effects

Locality Means (-fa)

Previous Opposition To Diversity

PareriLs are in FarmingParents are ImmigrantsN

Random Effects

Locality MeansPrevious Diversity AttitudesLevel 1 RProportion Reduction in Variance

Explanatory Variable

InterceptState, (« = MD. 1=IA)Percent FarmPercent Foreign BornFamily Income {Z)InterceptPercent FarmPercent Foreign BornFamily Income (Z)InterceptIntercept

Spring

y

29.998**-1.146-1.275**-1.736**

2.872.447**.024**.03S**

- 0 6 81.339**

-1.929

"02 Survey

SE7

3.2741.419.346.433

2.796.059.006.010.048.595

1.216889

Variance Component/ ^2 Sig.

213.48.077

137.6315.6%

.001

.001

Hierarchical Linear Model; Slopes and Intercepts Estimation.cell entries in the top portion of the table are regression coeffitients in the first column,

and standard errors in the second column.

from the Fall survey, and we would fully expect to see this continuityover such a short period of time. But what is still remarkable is that theintlvience of previous attitudes on curreni attitudes is much strongeramong those youth who live in contexts where there are farmpopulations and where there are immigrants.

For instance, the results in Table 3 indicate that if we take twoadolescents who were equally strong in their opposition to diversity inthe Fall survey, but one was from a farming community, and the olherwas not, the child in the farming community was about 10 percentagepoints more hostile to diversity by the Spring. Farm youth remain moreconsistently opposed to diversity over the six-month period than thenon-farm youth (see Table 3).

Similarly, for those living in immigrant-receiving areas, the relation-ship between the first and second measurement of attitudes towarddiversity is stronger than in non-immigrant areas. Again, taking twostudents who were equally strong in their opposition to diversity in theFall survey, but one is living near immigrants, and the other one is not.


the difference in attitudes toward immigration by Spring is about 13percentage points— a substantively large effect. Control variables inTable 3 show that those whose parents are employed in farming aremore opposed to diversity, while those whose parents are immigrantsare more welcoming of it, though the latter coefficient does not reachcustomary levels of statistical significance.

Without question, residential context does diversify rural attitudesabout immigration - even within predominantly rural settings! Ruralresidents are not monolithic on matters of diversity, as they are oftenconsidered to be by naive journalists and city dwellers. Two dimensionsof residential context matter most to views of immigrants: the presenceof immigrants in the respondent's neighborhood and the presence offarm populations. Botb of these contextual factors appear to greatlyaugment the negative attitudes of the most well-rooted natives, andtbose witb no college plans after high school, towaid immigrants.Importantly, however, the rural poor are more positive towardimmigrants when they live in closer proximity to them (see Table 2).

Discussion

Reflecting its unrecognized diversity, rural America has given immigrantsa variable welcome, neither entirely hostile nor wholeheartedly warm.Clas.sic fomiulations of the contact-friendship hypothesis suggest tbatequal status contact between (among) ethnic groups yields !e.ssinterethnic tension than might exist if that contact were between(among) unequals (Welch et al. 2001). We find support for tbatgeneralization in the.se results. Hostility to immigrants among the ntraladolescents we surveyed was concentrated among youth from weil-rootedand more affluent families, not the poor. Lower income natives in ritralscbools wind up sharing classes and school activities with the immigrantyouth, many of whom are tracked into non college-prep curricula. Theymay also live closer to them in small town and rural neighborhoods,perhaps having grown up together in the same housing development orpart of town. They attended the same elementary schools, and highschool does not constitute their first exposure to diversity.

By contrast, tbe affluent youth, the children of farmers, and thosewhose families are well-rooted in the community have far fewer of theseclassroom and neighborhood interactions. They do not live amongimmigrant families (in the same neigbborboods), and because of theirthick kinship ties, tliey interact far more with tlieir ovm family membersthan with others. Moreover, these embedded youth exhibit greatersuspicion about immigradon, particularly when there are immigrants


present locally. It is not rootedness in the community, by itself, thatproduces resistance to diversity, as much as it i.s rootedness interactingwith contextual factors, such as the presence of local immigrantpopulations and the dominance of farming. Conversely, it is the lowerincome youth who live in these very same areas who wind up exhibitingthe least opposition to diversit)'.

So where do our findings leave threat-motivated explanations forprejudice? At least among adolescents, personal economic tbreat doesnot appear to undergird hostility to diversity. Few native born youth,regardless of economic standing, intend to seek work in themeatpacking and food processing planLs of Peny and Storm Lake afterhigh school. Nor did we obsen'e acute shortages of housing, orcompetition over similar private goods, that might create tensionbetween immigrants and natives. These youth do not view immigrants,or their offspring, as a highly personal threat to their futures. The moststraightforward formulations of the power-threat hypothesis do nothold-up well in this analysis.

Anti-immigrant prejudice is most likely to be an expression of whitenative born youth from families with long tenure—a fascinating findingthat makes sense from a political socialization standpoint. Apparently,opposition to diversity resLs among those youth whose families areensconced in the community and commonly voice unease about changeon a variety of fronts, economic and otherwise. These are youth whosefamilies appear to be major stakeholders in the commimity, those whohave the most to lose from shifting demographics. Because the parentsand grandparents of these youth are so well-rooted in the locale, they areless likely to consider leaving when uncomfortable changes occur. Thisleaves them three primary options: 1) keep quiet, or 2) complain, or 3)do a bit of both—complaining quietly to those around them, includingtheir offspring—and this is the socializing force.

Anti-diversity attitudes are also the expression of a more generalsense of anxiety based on collective memories of how things used to be.The children whose parents and grandparents went to the same localschools hear stories about a previous era when these towns were chock-full of good-paying jobs, main street stores thrived, and no one lockedtheir doors. Undoubtedly many of these tales are romanticized andexaggerated, but their effect on adolescents is still pronounced,especially when no alternative socializing messages are heard. Steadyexposure to stories of an earlier golden era leads children to theconclusion that tbe foreign born presence has eroded the generalquality of small town and rural life, adversely affecting their family, andthe position of other well-established families. Teenagers with thick


kinship ties and frequent contact with senior family members areexposed to outmoded or archaic socializing messages on issues such asrace and ethnicity. Children from these towns whose extended familiesare not proximate never experience this additional source of attitudeconstraint.

Farming families from these towns also exhibit evidence of socializingoffspring into views that are more disparaging of diversity andimmigration than those held by townsfolk. WTiile one might think thatagricultme producers would better understand Lhe pressures ofglobalization faced by meatpackers, including the argument forcheaper labor, there has long been resentment of corporate foodprocessing giant-s among farmers who resent how IBP (Iowa BeefProcessors) and other major food processing industries conspire tokeep livestock and commodity prices low. To farmers, the wage workerin the packing plant does not cut an especialiy sympathetic figureeither. The immigrant laborer who will break U.S. immigration lawsand work for next to nothing is not necessarily viewed as a personaleconomic threat, bvit is associated with the difliculty of reevahiatingentrenched patterns of thought and adjusting to new realities. Therewas a time in the not so distant past when native born workers—aunts,uncles, and cousins—occupied many of these positions on the factoryfloor (StuU et al. 1995). But this memory is not shared by the shorter-term residents in these towns who do not have such points of reference.

While we cannot generalize ver\' confidently beyond the nine mraltowns we visited, and the adolescents we interviewed, we do believe thatsentiments sympathetic to active discrimination against immigrants andethnic minorities are fading with younger generations, in spite of thefact thai older residents seem to guide public opinion in these placesmore than they might in a fast-growing snburb or a large city. Thereplacement of the pre-World War II cohort with younger generationsover the next 20 years will smooth the way for the geographicdispersion of immigrants away from major port-of-entry locations at atime when tJie worldwide pressure to emigrate to the United States isonly rising. This docs not mean that tlie new arrivals will receive a warmwelcome in every town, as suspicions are still present as a consequenceof cultural differences with natives. Some locales undoubtedly havemore tioubled histories of intergroup tension than the ones we haveexamined. But we did find a prevalent graciousness and civility amongthe majority of the rural denizens of the towns we visited. This civilityand ordinary decency constrains the degree of confrontation that canoccur along ethnic or racial lines. Without quite labeling thesepopulations racially and ethnically "progressive," the youth we

200 Rural Sodohgy, Vol. 73, No. 2, June 2008

encountered mostly express the norms of tolerance and culturaldiversity necessary to avoid open racial and ethnic conflict. Diversity isbeing accepted, at least at a distance, and lhat is a good start towardharmony in places that have no previous experience with suchtransformation. To borrow a phrase from Seneca, time may indeedheal what reason cannot.

ReferencesAllport, G,W', 1954, Tbe Nature of Prejudice. Reading, MA: Addison-Wesley.Alvvin, D.F, and J. Krosnick. 1991. ".\ging. Cohons and the Siability of Sociopolitical

Orieniatioiis Over ihc Lift- Span," Ainerictm fmimal of Sociology 97:169-96.Baker, P.I.. and D,R. Hotek, 2003. "Perhaps a Bkssing: Skills aiid Contributions of Recent

Mexican Immigrants in the Rural Midwest." Hispanii Joumal of Behavioral Sciences25:448-68,

Beck, P.A. and M.K. Jennings. 1982, "Pathways lo Parncipation." Ammcan Political Science/^^irtd 76:94-108.

Be( g.s, ),J,. W.J, Villeiiif^, and R, Arnold, 1997. "Black Population Concentration andBlack-Wiiite Inequalit)'; Expanding tlic Consideraiion of Place and Space Effects."Sorifil Forces 76:65-91.

Blalock, M,M. 1967. Tmvarti a Theory of Minority Cjrottp Relations. New Yorlc: CapricornBooks.

Blumer, H, 1958, "Race Prejudice as a Sense uf Group PosiUdn," Parific Sociological Review1:3-7.

. 1965. "Indu.stria!ization and Race Relations." Pp. 220-53 in Industrialization andRace Relations, edited by G, Hunter. London: Oxford Universit)' Press,

liobo, L, and V.L, Mulchings. 1996. "Perceptions of Racial Group Competition:Extending Blunier's Theory of Group Position to a Multiraaal Social Context."Americtin Sociological Heineru 61.95]-72.

lionacich, E. 1976, "Advanced Capitalism and Black/White Race Relations in the UnitedStates; A Split Labor Market Interpretation." .American Sociological Review 41:34-51,

Campbell, D,T, and J,C. Stanley. 1963, Expeiintental anii (hiasiExpnimtmtat Oe.ugns forResearch. Boston, M.' : Houghton Mifflin Company.

Cummings. S. 1980. "Wbite Ethnics, Racial Prejudice and Labor Market Segmentation,"Amencan Joumal of Sociology 90:938-50.

Cummings, S. and T, I.ambert. 1997. "Anti-Hi.spanic and Anti-Asian Sentiments amongAfrican-Ameritans." Soiial Science Qtuirterly 78:338-53,

Conover, P,J. and D, Searing. 2000. "A Political Socialization Perspective." Pp. 91-126 InRediscovering the Democratic Purposes of Education, edited by L.M. McDonnell.P.M. Timpane. and R, Benjamin. Lawrence, KS: University Press of Kansas,

Dovidio,J.F., S.L, Gaertner, and A, Valid/it, 1998, "Intergroup Bias: Status Differentiationand a Common ln-(iroup Identity," foumal of Personality and Social Psychology7.'i:]09-120,

Elder, G.H., Jr. and R-D. Conger. 2000. Children ofthe Land: Adversity and Success in RuralAmerica. Chicago: University of Chicago Press.

Falk. W,W.. M.D. Schulman, and A.R, Tirkaniyer. eds, 2003. Communities of Work:Restructuring in lacal and (HOIMII Contexts, Athens, OH: Ohio University Press.

Giles, M. and K. Hertz, 1994, "Racial Threat and Partisan Identification." AmericanPolitical Science Review 88:317-26.

Gimpel. J.G.,J,C, hiy, and J,E. SchuknechL 2003, Cultivating Democracy: Civic Environmentsand Political Sodalizalion in Airurica. Washington, DC: Brookings Institute Press.

GIaser,J.M. 1994, "Back to the Black Belt: Racial Environment and White Racial Attitudesin lhe Howih." Joumal of Politics 56:21-41.


Green, D,P, and B. Palmquist. 1994. "How Stable is Party Identification?" Political Behainor43:437-66.

Greene, M,L., N. Way, and K. Pahl, 2006. "Trajectories of Perceived Adult and PeerDiscrimination among litatk. Latino and Asian .\merican Adolescents: Patlerns andPs-ycholngiral Correlates," Deiiehf»nental Psychology 42:218-38.

Horowitz, D.L. 1985. Ethnic C/roups in CA>nflict. Berkeley, CA: LIniversity of Califomia Press,Hughes, D,, E.P. Smith, H.C. Stevenson, J, Rodriguez, D,|, Johnson, and P, Spicer. 2006,

"Parents' Ethnic-Racial .Socialization Practices: A Renew of ReseaJch and Directionsfor Fuiure Study," Developmental Psychology A2:1A1-10.

[acknian, M.R. and M. tlrane. 1986. "'Some of My Best Friends are Black..,' Inten-acialFriendship and Whites' Racial Attitudes." Public Ofmiion Qtiarterly 50:459-86,

Jennings, M.K. and G.B. Markus. 1984. "Partisan Orientations over the Long Haul: Resuitsfrom a Three-Wave Political Socialization Panel Study." American Political Science/?n'w/'78:1000-8,

Kandel, W, and |, Cromartie. 2004. Nmv Paltems of Hispanic Settlemtnit in Rural Ainerica,Rural Development Research Report No. 99. Washington, DC: United StatesDepartment ofAgricuiture.

Kinder, D.R. and D.O. Seai.s. 1981, "Prejudice and Politics: Symbolic Racism versus RacialThreats to the Cood Life," foumal of Personality and Social Psychology 4:414—31.

Kish, L. 1987. Statistical Design for Research. New York: John Wiley and Sons.Key, V.O., Jr. 1949, Soiitfuni Politics in State and Nation. New York. NY: Alfred A. Knopf,Lay, J,C. 2006. "Learning about Politics in Low Income Communities: Poverty and

Political Knowledge." American PoUtirs Research 34:319-40.Lee, V.E. and A.S, Bryk. 1989, "A Multilevel Model of the Social Distiibution of High

School Achievement," Sodolo^ of Education 62:172—92.Lexin, M,L. 1961, "S<xial (Climates and Political Socialization." Public Opinion Qriarterly

25:596-606,Lipset, S.M. and E, Riiab. 1973. The Politics of Unreason: Right-Wing Extmnism in America,

1790-1970. New York, NY Harper Torchbook,Litt, E, 1963. "Civir Education, Community Norms and Political Indoctrination."

American Sociological R/nnew 28:G9—75.Maniyama, (.'>. 1998. Basics of Structural Equation Modeling. Thousand Oaks, CA: Sage.Oliver,J,E. and T. Mendelberg. 2000. "Reconsidering the Environ men ui! Determinants of

White Racial Attittides." American fotimal of Political Science 44:Ji7'l-i^9.Olzak, S, 1992. The Dynamics of Ethnic CMmpetition and CmifbcL Stanford, CT: Stanford

University Press.Phinney, J.S. and V, Chavira. 1995. "Parental Ethnic Sodalization and Adolescent Coping

with Problems Related Co Ethnicity." youma/ of Research on Adolescence 5:31—53.Pluty,er. E. 2002. "Becoming a Habitual Voter Inertia, Resotirces and Growth in Young

Adtilthood," American Political Sdence Revieio 96:57-74.Ramirez-Barranti, C.C. 1985. "The Grandparent/Grandchild Relationship: Family

Resource in an Era of V<iluniary Bonds." Family Relations 34: 343-52. Riiudenbush,S, W. and A. S, Bi yk. 2002. Hierarchical Linear Modeh: Apfilications and Data A7talysisMethods. Thotisand (iaks. CA: Sage,

Rice, T,W, and B, Steele. 2001, "Wiiite Ethnic Diversity and Community Attachment inSmall Iowa Towns." Social Sriimce Quarterly 82:397-^07.

Sanchez-Jankowski, M. I98(j. City Hound: Urban Life and Political Attitudes Amm^ ChicanoYouth. .•Mbuqtierqtie, NM: University of New Mexico Press.

Stheve, K.F. and M.J. Slaughter. 2001, " l^bor Market Competition and IndividualPreferences over Immigration Poiicj'." Revieio of Economics and Statistics 83:13.3—45.

Schuman, H., C. Steeh. and L.D, Bobo, 1985, Racial Trench in America: Trends andInterpretations. Cambridge. MA: Harvard University Press.

Sigelman. L. and S. Welch. 1993. "The ConUct Hypothesis Revisited: Black-WiiteInteraction and Positive Racial Attitudes," Social Forces 71:781-95,

Smith, A.W, 1981, "Tolerance of School DeisegregaUon: 19.'54-1977." Sodal Force.s59:1256-74.

202 Rural Sodology, Vol. 73, No. 2 June 2008

Steenbergen, M,R, and B.S, Jones. 2002. "Modeling Multilevel Data Structures," AmrricanJoumal of Political Sdence 46:218-37.

Stull, D,D., MJ, Broadway, and D. Griffith, 199.^. Any Way You Cul It: Meat Processing andSmall-Toiun America. Lawrence, KS: University of Kansas Press.

Suarez-Orozco, C, and Suarez-Orozco, M.M. 2001. Children of Immigralion. Cambridge, MA:Har\"ard University Press.

Taylor, M. 1998, "How White Attitudes Vary with Racial Composition of Local Population:Numbers Count," American Sodolo^cal fitww; 63:512-35,

\ o,ss. D,S. 1996. "Beyond Racial Threat: Failure of an Old Hypothesis iti the New South."Joumal of PoUtics ^S:il56-70.

Welch, S., L, Sigelman, T, Bledsoe, and M. Combs. 2001. Race and Place: Race Relations inan American City. New York: Cambridge University Press,

Appendix A. Factor Analysis for Dependent Variable: Opposition toImmigration-Related Diversity

Survey Question, Fall "01 Component I

Hispanic People Will Improve Life Here 0,8.SFewer Problems if Fewer ImmigiaiiLs —0.79Better Place if More Immigrants 0.75Inimigranis Here Take Jobs Away From Natives —0.68Asian People Will Improve Life Here 0.67Immigrants Work Harder than People Here 0.52Extraction Method: Principal Component Analysis.I componeJit extracted; explaining 50.-13% of variance

Survey Question, Spring "02 Componenl 1 Component 2

[lispanic People Will Improve Life HereBetter Place if More ImmigrantsFewer Problems if Fewer ImmigrantA.sian People Wilt Improve Life HereImmigranLs Here Take Jobs Away from NativesImmigranLs Work Harder Than People Here

F!!xtrdct,ion Method: Principal Component Analysis.2 components extracted;O>mponent I explains 43.5% of variance.Component 2 explains 19.6% of variance.

Appendix B. Factor Analysis for Family Longevity in the Community or"Rootedness"

Survey Question Component 1

Lived here all of my life 0.79Lived here less than 6 years -0.74Bom in same town 0.71Grandparents live in this state? 0.64Bom in foreign country -0,60

Extraction Method: Principal Component Analysis.1 component extracted, explaining 49.3% of variance.

0.760.76

-0.740.68

-0.470.48

0.38-0.12

0.280,080.740.62

Reactions to Immigration-Related Diver.-iity — Gimpel and Lay 203

Appendix C: Other Variable DefinitionsNo college plans: O=other plans after high school, I =do not intend to go to college (ofany type) alter high school.

Plans to attend 4 year college: O=other plans after hs, 1 =intend to pursue 4 yeardegree.

Estimated family income:

If you were to estimate your family's yearly (or annual) income, what would it be?

0=Le.S5than $20,000 1= $20,000-$35,000 2=$35,O0O-$50,OOO

3= $,50,O0O-$75,0O0 4= $7.'i,000-$100,000 !>=more than $100,000

Parents are full time fanners: O=neither; 1 =mother or father is; 2=father and motherboth.

ParenW are immigrants: O=neither; 1 ^mother or fatlier is; 2=fatlier and mother both.Density of acquaintanceship:

Ifyou were walking in downtown this afternoon, do you think you would you berecognized by

Converted to 0,1 0,1 Almost 0,1 Most 0,1 Some 0,1 Haidlydummy variables: Everyone People People Anyone

income defined by occupational status: Stttdents were asked an open-ended questiondescribing their parents' occupations. These were then coded by two coders into standardoccupational descriptions provided by the U.S. Census 2000. Once these occupations werecoded, we determined socioeconomic stattis by usitig the average income for thatoccupation drawn from the pooled Cimeral Sodal Surveys for 1998-2002. When bothparents worked and lived in the same household, the income amounts were sutnmed. Thecorrelation between this measure of income, and five-point income group estimatereported by the students in the question above was r = .33 (pS,OO!).

Percent Foreign Born: Percent living in census tract who were bom in foreign country,2000

Percent Farming: Percent of residents in census tract 16 years of age or older employedin full time farming, 2000.

Income: Median household income of census tract, 2000; converted to a z-scoreState: O-Maryland; l=Iowa.

204 Rural Sociology, Vol. 73, No. 2, June 2008

CTJ

32

o

40

in©

44

inOiCM

•J-CM

O

CM so•*• ©

CTl eort OO

04

CM

23

Tf

o

s

Oleoc

©

oUi I-rt oo

inTf-

61

in

Xo

51

©min

10

in

ooOi

66

inCMo

57

ineoCT)

75

Ol CM00 O

O} ono wIft

CO Tfi-;

in©

toTf

X•-.

CMCO©

oo XCM X© —

o

o

to

fiM O

S X inI-Tf

_

in

00 X

CO T i

©1ftin

Xr-r-

Tfin

CMX

Xo;in

mXXin

oO

fTj

eoC

O

mc50Tf

o

©

eo©

OlTf©

CTlooO

in

o

om©

©in©

©

o•ft

©

© ©Ift Ift

o

XCT:

mOleo

X

X

eo

CM

XoCO

OJ

ininCTJ

©

01©

X©o

©oo©

©

XOM

o

gc

CMCM

© r-- m© '- ©

CM CTl r^f r- oo© — ©

CO ^CM © X© OM O

C M ift CTJ

O rt o

eo in OMCM in toO rt o

oo 00 --1CM CTl -!fO rt O

© t- l>O rt o

© © X

© — ©

CO QO

© ©

CM 2"

© ©

© ©

^ in

c ©

m CMCM CMO ©

OO inoo ^© ©

CM Xrt CC© ©

in soTj" in

<£)CM

OOOO

—

OO

oCO—

©

<oCM©

vftO©

in

in inCM CMCM

©OO

X

CM

X5O*-

©

Ol

©

eoo©

oOM

X

X

'

oCMr-

TfTf

in

©

X

XCMTf

o

X

Tf©

in00'-

OM

•8.©

©

TfOlCM

CMin

CM

Srt

rt

I-

eo

r-eo©

©

so

XX

o

X©

glr-

XCO—

Tf

©

o

gi

5 SMOl ©X 1

$8CM ~^

an CMX 90

CTl

Tf CMX CTJ

X ©

CTl ift•ft >ftIft ©

•ft ^CM CO© ©

X -teo so

X O

Xin

27

—

-soIft

eoin

gCTJIft

97

so

in

©

49

o

o

in

o

CMO

49

o

Tf

©

f

o

50

©

o

©

©

in©

50

o

in©

©Ift

X

©

o

CM SfiX CMin

r-

ift

©

©

X XCO CMCM

XCTl

sCM

CT--Tf

©

O l©©

30

©

©

©

rt COO rt

rt inTf Ol© rt

rt IDCM rt© CM

X COCM rt© IM

X X© I-© rt

eo XCM O© CM

sg© CM

Tf

©

©

X

©

Tf©

©

Tf

©

—

X fCM CM

© ©

Ift SOCO in

© ©

Tf rt— CM© ©

Tf ID

© ©

CO OO

© ©

Tf in

© ©

O Xc^^ —© ©

X(N

OOTfrt

XcCM

inCO—

©

CM

X00rt

CM

CM

COo©

so

o

©

©

3©

©ffM

©©

CTl

CM

Tf

gX

o

©

X

©in

oOl

00XCM

©—

—

©

so

CMCTl<D

•*•

eo

Ift

OlX©

CM

—

CMtN

in

oinT*-o

©

in

©

XCM

©

OM

—

XCTl-

to

o

to

o

Tf

©

©

CTl ©X rtO> 1

X IftCTl CJTf ©

m •*•

^ o

Q in© oo^ ©

t~ ooso t-X OOi 1

X eo© COrt o

oo oOoo 00

X OCT) 1

be

"S .0 C

2 S^ „ gCMh u jB U

^ ^ B n• G "

o -- oi ' . S S ^ S ^ ^ i ^ X Z -S

0 5Cu Ca.—

O

ss: o I I I fcJ o ^ i § I SfX. Z b, 0. b. A.

s iW Srf u uV- k . b . k .

ob, pp, iro-20'i copv-righi © 2008. bl- rural sodulogjcal...

Documents