Diversity in Knoxville: An applied perspective

Madhuri Sharma

ELSEVIER

Contents lists available at SciVerse ScienceDirect

Applied Geography

journal homepage: www.elsevier.com/locate/apgeog

Diversity in Knoxville: An applied perspectiveq,qq

Madhuri Sharma*

University of Tennessee, 416 Burchfiel Geography Building, 1000 Phillip Fulmer Way, Knoxville, TN 37996-0925, USA

CrossMark

ABSTRACT

Keywords: Diversity score Metropolitan statistical area Isarithmic surface density maps Regression

Principal components analyses

Invasion-succession

Filtering

Ethnic-enclaves

Sustainable

This study contributes to the literature on applied urban geography by examining spatial patterns and processes of changing racial/ethnic diversity within the intra-urban context of Knoxville metropolitan statistical area. Knoxville embraces a diverse economic set up with opportunities in high-tech research and development, manufacturing, tertiary/service-sectors, construction, as well as entertainment industry. This serves well for its continued population growth, including minorities during 1990—2009. This paper explores how the neighborhood-level socioeconomic, demographic, and built-environment characteristics relate to tract-level racial/ethnic diversity, measured by multi-group diversity score and its components. Tools such as isarithmic surface density maps, correlations, principal components and regression analyses are used to examine processes of change. Results indicate that diversity in 1990 associates with negative change whereas diversity in 2000 associates with positive change. Though overall diversity sprawls and increases during 1990—2009, diversity among non-White declines during 2000—2009 and shows spatial confinement. Regressions suggest complicated mosaics of changing neighborhoods, providing evidence of invasion-succession, filtering and resurgence of ethnic-enclaves in specific neighborhoods. Concerning the six counties of the MSA, Knox is the most diverse whereas Union the least, though the share of Hispanics tops in Loudon and Asians in Knox. In terms of strategic planning, findings from this research can be used in creating equitable and sustainable urban communities that can improve the overall well-being of people by reducing racial/ethnic and socio-economic disparities that might occur as undesirable consequences of fast increasing diversity.

Introduction

Recent demographic and economic changes in the American South has earned it the name of The New South (e.g., McDaniel & Drever, 2009; Smith & Furuseth, 2004; Winders, 2006, 2011a, 2011b). This region, like the rest of the country, has been gaining in its racial/ethnic diversity over past two decades, particularly

q Earlier versions of this paper were presented at the annual meetings of the Association of American Geographers at Seattle, WA and New York, NY. Comments and suggestions from participants are appreciated and have been incorporated in the paper. Special thanks to Dr. Hyowan Ban of Geography, California State University-Long Beach, Dr. April Luginbuhl, freelance consultant and a friend from Ohio State University, and Dudley Bonsal (ABD) at Geography, University of Minnesota at Minneapolis, for providing valuable inputs on earlier versions of this draft. Special thanks to Will Fontanez, our department's Cartographer, and our undergraduate student assistant Chaney Paul Swiney with their assistance on making maps for this project.

qq This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. * Tel.: +1 876 974 6077; fax: +1 865 974 6025.

E-mail addresses: msharma3@utk.edu, madhursha@hotmail.com.

concerning the share of Hispanics (Smith & Furuseth, 2004; Winders, 2006, 2011a, 2011b) and Asians (Sharma, 2011a, 2011b). Much of it is driven by the growth of manufacturing and assembly plants and other service-sector opportunities moving to the southern states such as Tennessee, Alabama, South Carolina, North Carolina and Georgia, especially after the initiation of North American Free Trade Association (NAFTA) (Cobb, 2005; Perreira, 2011).

Knoxville, a southern mid-sized metropolitan statistical area (MSA) in East Tennessee had thrived as a major manufacturing and wholesaling center up until 1950s when its textile industry such as the Levi Jean, Knoxville Knitting Works, and other manufacturing industries collapsed. During past two decades though, Knoxville's economy has grown and diversified again, with several new manufacturing plants setting up their headquarters in Knoxville and in East Tennessee (Flory, 2011). These growing economic opportunities in Knoxville have contributed to its gain in racial/ethnic diversity, particularly during 1990—2009; this paper explains these changing residential mosaics at the scale of census tracts (CTs) by contextualizing them with its socioeconomic and built-environment characteristics.

Concerning its economic vibrancy, Lisega Inc. a German manufacturing company producing pipe supports and hangers

for the energy industry, opened up its only North American headquarter at Kodak City in Sevier County that was originally a part of Knoxville MSA(Flory, 20H).1 Knoxville has also thrived through the tough economic times of past two decades when the whole country encountered two recessions. In January 2012, while the national unemployment rate was 8.3% and the State of Tennessee's unemployment rate was 8.2%, Knox County's was at 6.2% while Knoxville MSA was at 6.7% (Labor Force Estimates, March 2012). Knoxville's tenacity to tread through the economically tough times have helped it gain and retain its diversity in terms of races, cultures, nationalities, classes and life cycles including the retired community (http://www.retirementplacesreport.com/tennessee_cities.html) as it also ranks high in the list of "most affordable places". Concerning industries and employability, Knoxville is headquarter to major employers such as the Aluminum Company of America (ALCOA), the Oak Ridge National Laboratory (ORNL), University of Tennessee (UT), UT-Medical Center, Blount Memorial Hospital, Covenant Health, Summit Medical Group, The Baptist Health System of East Tennessee, St. Mary's Medical Center, Home and Garden Television Network (HGTV), the Great Smoky Mountains National Park, Dollywood, Mid America Corp., Denso Manufacturing Tennessee, DeRoyal Industries, Sea Ray Boats. Inc., Philips Electronics, North America Corp., City/County of Knoxville, Boeing Defense & Space, Science Applications International Corp. (SAIC), BellSouth, Pilot Corporation, Matsushita Electronic Components Co. of America, and many more multinational corporations (http:// web.knoxnews.com/jobs/knoxville/employ.shtml).

Knoxville, a southern MSA, has 11.1% of its total population in 2009 as non-Caucasian (Table 1-A, www.census.gov). Knoxville consists of six counties — Anderson, Blount, Knox, Loudon, Sevier and Union, (2000 Census definition), with the city of Knoxville seated in Knox County, and the six counties consisting of 139 census tracts (CTs).2 Knoxville is interesting due to its mid-size college-town characteristics with the benefit of a moderate climate and beautiful landscape tucked in the Appalachian Valley (Fig. 1). It also has a diverse economic base in comparison to other MSAs of same size. Its growth in the share of Asians and Hispanics at 131.68% and 408.01% during 1990—2009 was far above the growth rates for White or Black (Table 1-A).3 Changing characteristics of

1 Concerning Knoxville and east Tennessee, as of 2011, about 67 German companies are doing business in Tennessee; prominent automaker Volkswagen recently opened an assembly plant in Chattanooga, and solar company Wacker Chemical is building a $1.5 billion polysilicon plant in Bradley County (Flory, 2011).

2 In 2010, Sevier County was removed from being considered as a part of Knoxville MSA. In this analysis, however, I use the six-county definition for Knoxville. I do this as I had already computed diversity score indices for 1990 and 2000 for a previous analysis, and I wanted to measure two decades of change in diversity. Using 2010 census data would have served better, but normalized data are available from geo-lytics only upon purchase. In 2010 Census, several tracts changed their boundary definition making them incomparable with1990 and 2000 statistics, and hence I used 2005-09 ACS five year estimates data that still follows 2000 tract boundary definitions. Thus, using the ACS 5 year estimates for 2005-09 enabled this comparison and analyses of computed indices across 1990-2000-2009. Also, ACS 5 year estimate data for extracting socio-economic and built-environment variables for the census tracts that enabled further analyses with correlations and regressions. For more detailed information on geolytics products, see www.geolytics.com/USCensus, Neighborhood-Change-Database-1970-2000 Products.asp

3 In this paper, I use the terms African-American and Black interchangeably, likewise White and Caucasian are used interchangeably. Asian refers to the combined group of non-Hispanic Asians along with Hawaiian and Pacific islanders; American Indian refers to the combined group of non-Hispanic American Indians along with Alaskans/Aleuts/Eskimos, etc.; All-Other has been used for non-Hispanic All-Other along with Some Other Races and Two or More Races. Also, the terms White, Black, Asian, American-Indian and All-Other refer to the non-Hispanic groups whereas Hispanics are a separate group. While the term Black doesn't

sound very pleasant, most segregation literature use this term very often, and I have tried to use African-American instead, but at some places I have also used Black also because of its acceptance by the larger academic community.

demographic diversity in a metropolitan area that mostly had Black diversity until 1990s may create demand of multilingual schools, multi-cultural centers, health care professionals and other civic facilities/social/welfare institutions to support and retain them.

This analysis is part of an ongoing larger project on Knoxville that examines how racial/ethnic diversity and intermixing is manifested within intra-urban contexts, and how do they get impacted by the housing market elements such as bankers/lender, realtors, builders/developers and local communities. An earlier part of this analysis discussed broad spatial patterns of diversity, intermixing and clustering during 1990—2000 that has been published. This paper moves ahead by exploring the relationship between tract-scale racial/ethnic diversity at a point in time, i.e., in 2009 and its change during 1990—2009 with their socio-economic and built-environment characteristics to explain the evolutionary process of changing diversity. In this paper, I explain these processes using multiple statistical and cartographic tools such as isarithmic surface density maps for computed diversity scores, followed up with analytical tools such as bivariate correlations, principal components and regression analyses. This paper adds to the genre of research on diversity/urban geography within the context of an emerging mid-sized southern metropolitan area, particularly in light of Knoxville's significance as the third largest new attraction center for Latinos, following Nashville in Tennessee and Birmingham in Alabama. The remainder of this paper proceeds in four sections: literature review, research design, analyses and findings, and finally discussions and conclusions.

Literature review

Racial/ethnic diversity and transitioning neighborhoods

Up until late 1900s, a large body of diversity/segregation literature has situated their empirical analysis using existing frameworks on urban ecology. These include the Classical Assimilation, Place Stratification and the Resurgent Ethnicity frameworks.4 Concerning empirical work, so far several researchers working on diversity/segregation have (i) focused on the largest MSAs/gateway cities(e.g., Charles, 2000, 2003; Clark & Blue, 2004; Ellis, Wright, & Parks, 2004; Farley & Frey, 1994; Frey & Farley, 1996; Singer, 2003, 2004; Timberlake & Iceland, 2007), and/or (ii) treated cities/metropolitan areas as the object of analysis (Brown & Sharma, 2010; Charles, 2000,2003; Clark & Blue, 2004; Ellis et al., 2004; Farley 1996; Farley & Frey, 1994; Frey & Farley, 1996; Logan, Alba, Dill, & Zhou, 2000; Singer, 2003, 2004). Those focusing on diversity/immigrants' dispersal to non-traditional urban/rural locations include Lichter, Parisi, Grice, and Taquino (2006), Singer (2003, 2004), Smith and Furuseth (2004). Intraurban analyses has been relatively rare, except for some recent ones like Acevedo-Garcia and Osypuk (2008), Grady and Darden (2012), Lobo, Flores, and Salvo (2002), Sharma (2011a), Sharma and Brown (2012), Smith and Furuseth (2004), Winders (2006, 2011a, 2011b).

4 Assimilation embraces the melting pot ideal (Alba & Nee, 2003) where immigrants, having adapted to US society and gained in socio-economic status, move into established neighborhoods that are spatially more distant from the CBD with a higher representation of Caucasians (Brown and Chung 2006, 2008). Stratification focuses on the persistence of segregation from housing discrimination, racial-stereotypes, and prejudicial preferences (Bobo & Zubrinsky, 1996; Brown and Chung 2006; Charles, 2000, 2003; Logan et al., 2000). Resurgent Ethnicity emphasizes the persistence of racial/ethnic clusters in residential choices even after the improvement of socio-economic status that could eventually lead to choosing an intermixed neighborhood, and yet, some people choose to reside in neighborhoods where a majority of their neighbors belong to their racial/ethnic background (Brown & Chung 2006; Logan, Zhang, & Alba, 2002, Logan et al., 2004).

Concerning intra-urban changing mosaics in major gateways, Lobo et al. (2002) discuss the effect of growing shares of Latina, African-Americans and Asians in the CTs of New York City during 1970—90, eventually transforming 38% of its White-dominated CTs multi-ethnic. Sandoval and Li's (2004) analysis of increasing diversity in Chicago finds their spatial dispersal beyond city centers. Concerning intra-urban context, Denton and Massey (1991) examine the 50 largest Standard Metropolitan Statistical Areas (SMSAs) plus 10 others that had substantial Hispanic population, using census data for 1970 and 1980, and find that only 31.5% of All-WhiteCTs remained so by 1980 whereas others become multiethnic. Chung's (2005) analysis of Columbus, Ohio suggests substantial dispersal of minorities beyond Interstate-270 during 1990— 2000. Dingemans and Datel's (1995) work on Sacramento, California findsinvasion-succession occurring from growing presence of Blacks and Latinos in central city locations.

Racial/ethnic diversity, changing socio-economic contexts and the New South

Concerning changing economic contexts, Florida (2004) notes that the economic diversity of a place contributes by transforming it into a "creative place" as it gains in diversity of sorts — racial, ethnic, lifestyles, etc. since such places garnish the three Ts — technology, talent and tolerance. The relationship between economic contexts and diversity can be best understood from history wherein the American Manufacturing Belt (AMB) thrived through the 1950s under Fordism, serving as magnets for African-Americans during the Great Migration of 1910—1930, as well as for other immigrants from European and the Middle East region (Geib, 1998). While racial/ethnic enclaves were important landscapes of major gateways such as Chicago, New York, Miami, Los Angeles, etc. (Charles, 2003; Clark & Blue, 2004), the growth of African-Americans during early-to-mid 1900s further facilitated the formation of Black ghettos in large-to-mid-sized MSAs, initially developing into monocentric city patterns and later through formation of edge cities/suburban enclaves, etc. The improvements in the conditions of Black in the Midwest started only after World War II when they achieved better education, enlisted in the military, got better job opportunities, eventually creating a Black-middle class (Geib, 1998). They competed for manufacturing jobs, particularly after unionization, and these collectively led to their upward mobility, and their fast growth in major Midwestern cities (Geib, 1998). Concerning the economic status of minorities, particularly as a result of the economic restructuring of the 1970s and 1980s, Levine (2000) and Sassen (1991) note that Black and Latino in the larger cities of the Midwestern region were the most adversely affected (Levine, 2000) and the minorities in the global cities were once again the worst sufferers of socio-economic and racial polarization (Sassen, 1991). Concerning residential segregation and economic contexts in other large-to-mid-sized MSAs, Brown and Sharma (2010) in their analysis of 49 MSAs (larger than 1 million population in 2000) find that when the geographic region serves as a surrogate for analyzing AMB/Rust Belt versus Sun Belt dichotomies, the AMB-MSAs suffer a heavy burden of sunk costs initially, but are soon absorbed (or written off) by 2000 and that they subsequently observe substantial shift and larger degrees of gain in intermixing over the duration 1990—2000 compared to others.

Concerning the New South, several southern MSAs have gained in their shares of immigrants and racial/ethnic diversity, triggered from economic opportunities in formal and informal sectors (Cornelius, Fitzgerald, Fischer, & Muse-orlinoff, 2010; Perreira, 2011; Sharma 2011a; Shultz, 2008; Smith & Furuseth, 2004; Winders, 2011a, 2011b). The southeastern USA has the second

largest concentration of Hispanics, and their share has increased by more than 100% during 1990—2000 across all southeastern states, with North Carolina experiencing the highest at 386%. Charlotte's Hispanic, non-Hispanic African-American and non-Hispanic White share constituted 2.6%, 35.1% and 56.8% of its total population in 1990; by 2000, these changed to 23.85%, 34.5%, and 28.4% respectively, indicating significant gain for Hispanics and huge decline for Caucasians (Smith & Furuseth, 2004: 221).

Among other mid-sized southern MSAs, Nashville, Tennessee, has drawn attention because of its significance as a "new destination" for Hispanic immigrants and refugees, driven from its vibrant music and health-care/insurance industry (Chaney, 2010; Winders, 2006). Going further south, McDaniel and Drever's (2009) analysis of Birmingham, Alabama finds that immigrant entrepreneurs are not only building ethnic enclaves, but they are also taking advantage of automobiles and advanced communication technologies, and are residentially dispersing with no more confinement to inner cities. Bobby Wilson's America's Johannesburg (2000) and later Connerly's 'The Most Segregated City in America' (2005) note that Birmingham has long since followed segregationist city planning policies through racial zoning, urban renewal, and placement of interstate highways that collectively ensured that the city remained 'the most segregated city in America'. However, recent economic growth in northeastern Alabama have employed thousands of Hispanics in the poultry/vegetable/fruits farming in Russellville, Gadsden, Attalla, Raul and Collinsville, and in the fishing/tourism industries in southern Alabama (Cobb, 2005).

Racial/ethnic diversity, health/social inequality and planning and policy

A new line of research has addressed the effects of growing diversity on rising inequalities/health risks of minorities/population residing in highly segregated spaces. Grady and Darden (2012), for example, in their analysis of Detroit MSA find that institutional racism, unfair policies and practices in the housing industry and other institutions have produced racially inequitable outcomes for black mothers while benefitting white mothers. Grady and Darden (2012: 928) find that when they control for high black segregation and very low socio-economic position indicator (SEP), there was significant reduction in the racial disparities and preterm birth. Thus, by ignoring racial segregation, one might be misestimating the effect of individual-level risk factors on health of minorities and children, particularly for Latinos and Blacks, that affect them not only in their childhood years, but lingers on throughout their life courses (Acevedo-Garcia & Osypuk, 2008; Acevedo-Garcia, Osypuk, McArdle, & Williams, 2008; Grady & Darden, 2012; Osypuk, Bates, & Acevedo-Garcia, 2010).

Very often planning and zoning policies, particularly in US urban areas, create spaces of distinction, with low quality housing structures, inaccessibility to healthy food options, lack of adequate health-care facilities, environmentally degraded locations, congestion and likewise (Acevedo-Garcia & Osypuk 2008; Acevedo-Garcia et al., 2008; Connerly, 2005; Jones, 1995; Wilson, 1992, 2000; Wilson, Hutson, & Mujahid, 2008). Unfortunately, these zoning policies more adversely affect the minorities, which challenge the communities and cities across the country plagued with racial/ethnic fragmentation, environmental injustice, and health disparities (Jones, 1995; Wilson et al., 2008).

Concerning diversity and policy making, Letki (2008) in her analyses of communities in the Great Britain, notes that diversity creates tension, mistrust, doubts, concerns, unacceptability and intolerance among people. However, she also finds that the negative consequences of diversity actually reduce, when economic variables are controlled. In short, she concludes that

diversity is good and in the British society, it has become a basis for drawing out important plans and policies to create sustainable and equitable communities. Likewise, Papillon's (2002) analysis of Canadian cities finds that diversity is an important ingredient for sustainability, and that the capacity of institutions at the national, provincial and local levels should be improved to counter the patterns of socio-economic and spatial exclusion. In their analyses of participatory planning and inclusion of local communities in a Jewish-Arab community, Shmueli and Kipnis (1998) indicate that involving the community ensures that their needs and concerns are addressed adequately, that a sense of ownership helps its implementation by creating a willingness among the community to accept alternatives and compromises. These demand integrating the new immigrants and diversity into the labor market (Papillon, 2002) and guarantee them access to various types of social/civic services, language training, educational (Jones, 1995) and housing opportunities (Vliet, 1996), and help create urban space(s) where they can build social networks and participate in various cultural and political life without giving up their own cultural and communal ties. These aspects have also been discussed by Winders (2011a, 2011b) concerning the social, cultural, and political integration of Latinos in the music city of Nashville.

Methodological approaches in measuring diversity

Among numerous measures of diversity/segregation, most commonly used are the Dissimilarity Index (Duncan & Duncan, 1955; Wong, 2008), Location Quotient (Brown & Chung, 2006; Chung, 2005; Ellis, Holloway, & Wright, 2007; Isserman, 1977; Leigh, 1970; Moineddin, Beyene, & Boyle, 2003; Sharma, 2011a, 2011b), and Exposure Index (Logan, Stults, & Farley, 2004; Wright, Holloway, & Ellis, 2011; Wong, 2003). Séguin, Apparicio, and Riva (2012) use location quotient to analyze poverty in the neighborhoods of Montreal, Canada during 1986-2006, and find that changes were minor over time, except for CTs in the gentrifica-tion trajectory where changes in poverty levels were more significant. Among those using Local Moran's I include Brown and Chung's (2006) analysis of segregation in Columbus, Ohio, Monkkonen's (2010) analyses of segregation by ethnicity and socioeconomic status in over 100 cities of Mexico, and Roberts and Wilson's (2009) analyses of new urban growth, social disparity and segregation in the largest MSAs of Latin American. Recently, a more accurate measure of clusters and geodemo-graphic segmentation was created by Grekousis and Thomas (2012), known as the Fuzzy C-Means algorithm and the Gus-tafson-Kesselalgorithm. They use this measure to examine segmentation and clusters in the prefecture of Attica, Greece. They find that the residents of eastern and northern suburbs have average to high incomes, belong mostly to working class as well as high-flying executives and/or employed in science and academia. On the other hand, residents of western suburbs have lower income and education, with several children, are unable to speak foreign languages, and work as technicians, machine operators and unskilled laborers.

Concerning conceptual notions of measuring diversity/segregation, Ellis et al. (2007) raise questions about the inaccuracies associated with these measures as most indices ignore the mixed-race households that are rising in contemporary America, particularly in the largest gateways. They find that racial mixing within households has meaningful effects on these measures, and that there is a need to re-conceptualize these measures when analyzing diversity and mixing in an increasingly multi-ethnic society. To address this issue, Wright et al. (2011 ) use Scaled Entropy to measure racial/ethnic diversity at the census tract scale,

by categorizing diversity into three levels — low, medium and high, using numerically dominant racial/groups. They demonstrate the benefits of using this measure that forces one to think beyond single-group numerical dominance, or even pairs of groups. Holloway, Wright, and Ellis's (2012) analysis of sixteen MSAs also use Scaled Entropy, and find that residential neighborhoods are becoming racially more diverse despite significant shares of urban landscapes exhibiting high levels of segregation, and that the relationship between segregation and diversity are not simply the mirror images of one another, instead diversity is becoming enfolded within segregation and vice-versa in very complicated ways.

Given the increasingly multi-racial/ethnic nature of urban America, a small group of scholars have used the Multi-group Theil Entropy Score/Index (Brown & Sharma, 2010; Ellis et al., 2007; Fischer, Stockmayer, Stiles, & Hout, 2004; Iceland, 2004; Reardon & Firebaugh, 2002; Reardon & Sullivan, 2004; Sharma, 2011a; Sharma & Brown, 2012; Timberlake & Iceland, 2007; Wright et al., 2011), which follows the conceptualization created long ago by Theil and Finezza (1972). Given the focus of this paper is changing multi-group diversity, I use the Theil Entropy/Diversity Score as a measure of analysis.

Research design

Study area and data sources

The city of Knoxville is seated in Knox County, and the MSA includes other incorporated cities/urban areas such as Alcoa, Clinton, Farragut, Lenoir City, Loudon, Maryville, Maynardville, Oak Ridge, Sevierville, Seymour, and Pigeon Forge (Fig. 1). Important freeways/state routes cutting through the six counties of the MSA include I-75, I-40, I-640,11, 441, 411,129, 321, 66, and 25W. Most of the new residential and commercial development in Knoxville follows the East-West corridor, along US-11, more commonly known as the Kingston Pike, because of availability of flatter land and the construction difficulties along the north-south orientation of the Appalachian Mountains.

In this analysis, CTs of the six-county MSA form the unit of analysis. Block group (BG) data for six population groups — All-Others (including Other-Races/Two or More Races), American Indian, Asian (including Hawaiian and Pacific Islanders), Black, Hispanic and White are used for computing diversity score (DS) and its components for each CT. The NCDB product of Geolytics for 1990 and 2000 and ACS 5-year estimates for 2005-09 constitute the data source. For correlations, principal components and regression analyses, tract-level data are used. Care is taken to assemble and create similar types of variables across demographic, socio-economic and built-environment categories since the definition of variables have changed during 1990— 2000, and for the ACS-five year estimates. Where comparative data/variables are not available, they are excluded from analyses.

Measurement statistics and methodological steps

Measurement statistics for this analysis include Diversity Score (DS) and its components. DS is computed using BG data using the specifications below (Timberlake & Iceland, 2007: 341; Sharma, 2011a: 313—314; Sharma & Brown, 2012: 326—327). Thus, the equations for computing diversity score(s) are:

DS = ]T(Pr)ln(1/Pr) (1)

DSi = ]T(Pri)ln(1/Pri) (2)

Where DS is the Diversity Score for census tract, DSi is the Diversity Score for each block group i(BGi) within that census tract, Pr is the proportion of a particular racial/ethnic group r in the census tract, Pri is the proportion of a particular racial/ethnic group r in block group i within that census tract, where there are n racial/ethnic groups.5 DS varies based on the number of groups and proportions of each group. Its lower bound is zero when only one racial/ethnic group is found, and has an upper bound when all racial/ethnic (R/E) groups are equally represented. In a simulation exercise, DS = 1.799 for 6 R/E groups (Pr = 0.167 for each), 1.609 for 5 R/E groups (Pr = 0.200 for each), 1.386 for 4 R/E groups (Pr = 0.250 for each) (see Brown & Sharma, 2010).

In this analysis, I use BG data to first compute DSi for each block group, which enables me to compute DS for the CTs, using SPSS syntax commands. I use BG data instead of tract level data as the SPSS commands enable me to compute two components of DS (Overall diversity score) for each census tract: (i) DSWNW referring to the diversity because of the two major groups — White versus non-White, and (ii) DSNW referring to the diversity among the non-White groups (i.e., the diversity occurring from the presence of other five segments — Black, Asian, American-Indian, All-Others and Hispanic). These two components provide valuable information concerning which component of diversity contributes toward larger or smaller share of overall diversity for each CT.6 Besides computing the DS for tracts, I also compute DS for the six counties of the MSA for 1990, 2000 and 2009, using Eq. (1) where Pr refers to the proportion of each racial/ethnic group in each County (see results in Table 1-B & C, Fig. 2).

Once the diversity scores and its components are computed for all 139 census tracts, these are mapped using isarithmic density surface mapping techniques, using natural breaks as the class-categories (Figs. 3 and 4), and analyzed for spatial variation across time. Next, a bivariate correlation analysis is conducted using a select list of tract-level demographic, socio-economic, and built-environment variables to identify multicollinearity and eliminate them from further use in principal components analyses (PCA).7 PCA helps extract four components that distinguish in terms of place-based characteristics. These components along with diversity score in 1990 and 2000 are used as independent variables in regression analyses whereas dependent variable consists of diversity score in 2009 and its change (2009—1990). I also conduct

5 DS and DSi values are used to compute Theil Entropy Index (measure of intermixing) for the census tracts. However, in this paper, I focus only on the diversity score and its components, honoring the length of the manuscript.

6 The SPSS syntax for computing the diversity score and its components can be written for any group-combination one is interested in analyzing (for example, Hispanic versus non-Hispanic and among non-Hispanic or Black versus non Black and among non-Black, and so on); in this paper, I focus on White vs. non-White and among non-White.

7 The variables in demographic, socio-economic, and built-environment charac-

teristics were identified based on earlier academic work done by several scholars. In

this paper, my focus is to distinguish the neighborhoods based on their race-based and socio-economic and geographic contexts, and hence I select variables that reflect these characteristics well. For example, manufacturing/blue-collar versus white-collar/professional indicate class effect, median household values, age of housing, share of newer housing etc. reflect the spatial characteristics of a place, educational achievements and income levels of population subgroups explain the overall well-being and class aspects associated with people and place. Thus, these formed my basis for identifying the variables for this analyses. Also, before conducting PCA, stepwise-backward and OLS regressions were also explored to find a fit model, but they did not yield good results and hence I conducted the PCA.

regression analyses for change in diversity score for White versus non-White group during 1990 to 2009.8

Analysis and findings

Racial/ethnic diversity at intra-urban contexts: spatial patterns for tracts and counties

During 1990—2009, Knoxville's share of Hispanics, Asians and Blacksincreased by 408%, 131.68% and 28.41% respectively, and its diversity score changed from 0.325 (1990) to 0.486 (2009) (Table 1-A). While Nashville and Memphis, Tennessee's two largest MSAs have contributed toward a larger part of Tennessee's overall racial/ ethnic diversity (Sharma, 2011b), Knoxville's diversity score continues to increase (Table 1-A, C) whereas Nashville's and Memphis's have declined during 1990—2000 (Sharma, 2011b: 312, Table 19.2). Concerning the six counties of Knoxville, Table 1-B indicates that every County has gained in its share of non-White population, particularly Knox, Loudon and Sevier (www.census.gov). For Knox County, Hispanic share has changed from 0.6% (1990) to 1.2% (2000) and 2.3% (2009); for Loudon these are 0.3%, 2.2% and 4.3%; for Sevier these are 0.7%, 1.0% and 2.4%. Concerning Asians, Knox County's share increased from 0.9 (1990) to 1.7 (2009). Anderson and Union counties also have good presence of Asians, and both are within close proximity to ORNL. Overall, Knox County has the highest share of minorities among all six counties, with 14.4% in 2009 as against 10.5% in 1990 (www.census.gov). Fig. 2 and Table 1-C indicate that diversity scores have increased for all counties during 1990—2009, with Knox at the highest (DS2oo9 = 0.576), Anderson's the second highest (DS2009 = 0.419) while Union at the lowest (DS2009=0.163) with a small decline in its white share from 99.2% (1990) to 97.1% (2009) (Table 1-B-III).

Concerning diversity at the scale of census tracts, Fig. 3 indicates spatial dispersal during 1990—2000—2009, from inside of Interstate-640 in Knox County in 1990 toward outwards into Anderson, Loudon, Blount and Sevier by 2009; highest levels of diversity (dark color) has also sprawled during 1990—2009. By 2000, diversity spread toward East Knox County and Oak Ridge; by 2009, it expanded further into Maryville, Sevierville, Loudon, Union and Sevier counties.

Diversity for White versus non-White (Fig. 4, left) shows a spatial spread-out during 1990—2000—2009, particularly in the vicinity of Knoxville, Oak Ridge, Loudon, Maryville, Pigeon Forge and Sevier-ville. However, diversity for non-White (Fig. 4, right) indicates its spread during 1990—2000 and a spatial confinement during 2000— 2009, possibly forming into ethnic enclaves(note Resurgent Ethnicity framework). Figs. 3 and 4 also suggest that the hot spots of diversity are around the downtown areas of most cities/urban areas of the MSA. In particular, Anderson, Blount, Knox, and parts of Sevier counties show hot spots for non-White diversity in 2009 whereas this was more contiguous throughout the whole MSA in 2000.9 This

8 The regression models for change in diversity score for the "among non-White" group did not generate any significant models, and hence those results are not presented here.

9 To substantiate this finding about the spatial pattern of diversity scores, the author's neighborhood reconnaissance while conducting field work with homeowners and foreclosures in Knoxville MSAsuggests that there is visible presence of Latino labor engaged in various types of service-sector jobs such as hotel cleaning personnel and restaurant services in Sevier County, in new construction projects along Alcoa Highway in South Knox and Blount counties and toward west in Loudon County. Field work also suggests that many souvenir shops, hotel businesses, and other such business catering to the tourists in Sevier County were owned by Indian and Pakistani businessmen (Asians). The author has not investigated further concerning data on race/ethnic businesses in the location, but thought it was worthwhile to mention about these observations and discussions with respondents from the field notes.

Table 1

Demographic composition and divesity score of Knoxville MSA and its six counties, 1990-2000-2009

A. Demographic composition of Knoxville MSA and its Change, 1990-2009

MSA statistics 1990 Share 2000 Share 2009 Share Pct. Change,

1990-2009

Total population 586,025 1.000 687,249 1.000 764,077 1.000 30.38

White 541,525 0.924 622,899 0.906 679,542 0.889 25.49

Black 35,126 0.060 39,628 0.058 45,104 0.059 28.41

American Indian 1536 0.003 2153 0.003 1791 0.002 16.6

Asian 4224 0.007 6256 0.009 9786 0.013 131.68

All others 180 0.000 8883 0.013 10,409 0.014 5682.78

Hispanic 3434 0.006 7430 0.011 17,445 0.023 408.01

Diversity score 0.325 0.420 0.486

B. Demographic composition of the six counties of Knoxville MSA, 1990—2009

B-I: Racial/ethnic proportions in counties in 1990 Population

Counties White Black Amer-Indian Asian All-Others Hispanic County

Anderson 0.942 0.039 0.002 0.008 0.000 0.009 68,251

Blount 0.957 0.031 0.003 0.005 0.001 0.004 85,921

Knox 0.895 0.087 0.003 0.009 0.000 0.006 33,5748

Loudon 0.982 0.012 0.001 0.002 0.000 0.003 31,321

Sevier 0.985 0.004 0.003 0.002 0.000 0.007 51,090

Union 0.992 0.000 0.001 0.002 0.000 0.006 13,694

MSA population 586,025

B-Ü: Racial/ethnic proportions in counties in 2000 Population

Counties White Black Amer-Indian Asian All-Others Hispanic County

Anderson 0.928 0.038 0.005 0.007 0.014 0.009 71,330

Blount 0.942 0.029 0.003 0.006 0.013 0.007 105,823

Knox 0.874 0.086 0.003 0.012 0.013 0.012 382,032

Loudon 0.952 0.011 0.002 0.002 0.012 0.022 39,086

Sevier 0.962 0.007 0.004 0.006 0.011 0.010 71,170

Union 0.984 0.000 0.001 0.001 0.010 0.004 17,808

MSA population 687,249

B-III: Racial/ethnic proportions in counties in 2009 Population

Counties White Black Amer-Indian Asian All-Others Hispanic County

Anderson 0.912 0.040 0.004 0.011 0.016 0.018 73,382

Blount 0.927 0.030 0.003 0.009 0.014 0.017 119,489

Knox 0.856 0.088 0.002 0.017 0.014 0.023 423,655

Loudon 0.929 0.009 0.002 0.002 0.015 0.043 45,176

Sevier 0.946 0.009 0.003 0.007 0.011 0.024 83,448

Union 0.971 0.000 0.001 0.013 0.004 0.010 18,927

MSA population 764,077

C. Population and diversity score for counties of Knoxville MSA, 1990—2000—2009

Counties 1990 2000 2009

Diversity score County population Diversity score County population Diversity score County

population

Anderson 0.278 68,251 0.356 71,330 0.419 73,382

Blount 0.220 85,921 0.299 10,5823 0.363 119,489

Knox 0.404 335,748 0.509 382,032 0.576 423,655

Loudon 0.112 31,321 0.257 39,086 0.336 45,176

Sevier 0.099 51,090 0.219 71,170 0.284 83,448

Union 0.055 13,694 0.101 17,808 0.163 18,927

pattern is further supported by the computed scores for all 3 years, where the maximum value of non-White diversity increased from 1.513 (1990) to 1.576 (2000) which declined to 1.509 (2009), even though overall diversity score increased from 0.938 (1990) to 1.09 (2009).

10 Table 2 presents correlations for a selected list of variables due to space constraints, and detailed table may be provided upon request. Variables were identified based on preliminary analyses with a larger set of variables pertaining to race-based socio-economic, demographic/change in racial groups, and built-environment characteristics. Race-based variables were selected to extract characteristics associated with the racialized aspect of neighborhoods/places reflected in the PCA components. In the PCA, a larger set of variables have been used and displayed as most of them have significant loadings that talk about place-based characteristics.

Neighborhood correlates

Bivariate correlation analysis (Table 2) indicates an interesting relationship between diversity score in 2009 (DS2009) and its change during 1990—2009 (ADS(2009—1990)) with tract-level attri-butes.10 Population size (2000 and 2009), change in population during 1990—2009 and 2000—2009, change in White (1990— 2009) and share of non-movers (1985, 1995) are all significantly but negatively associated with DS2009 whereas share of foreign-born (F-B) in 1990, 2009 and share of F-B entered during 1990— 2000 are positively associated with DS2009. Concerning socioeconomic characteristics, those positively associated with DS2009 include Bachelors/above levels of education (overall population in 1990, 2009), and for White, Black and Asian in 1990, median household incomes (1990, 2009) and per capita income for Black

K¡ loríete r«

Fig. 1. Knoxville MSA with counties, census-tracts, cities and incorporated urban areas and important roads/highways

(1990, 2009). Most education variables relate positively with ADS(2009—1990)whereas only Black per capita income (1990) relates positively with ADS(2009-1990). Concerning built-environment/ spatial attributes, it is no surprise that those employed in manufacturing occupations (1990, 2000) relate with low DS2009 and have no association with ADS(2009—1990). Interesting to note, however, is that race-based occupation in either manufacturing or managerial in 2009 are all positively associated with DS2009 whereas they have no association with change in diversity. Other interesting finding is that homeownership by White, Black and

Asian in 1990 and White, Black and Hispanic in 2009 are positively associated with diversity in 2009 and insignificantly with change whereas Hispanic homeownership (2009) and Hispanic professional occupations (2009) are both positively associated with change in diversity. Median year of housing structure built (2009) and the share of newer homes built, i.e., share of homes built during 1990—2000 and 2000—2009 out of 2009 total housing stock are all associated with lower diversity in 2009. Concerning diversity score and its components, the surprising negative relationship between diversity score among non-White

Fig. 2. Diversity scores and population size in six counties of Knoxville metropolitan area over 1990-2009

(1990, 2000) with diversity score in 2009 is indicative of minority-clusters within intra-urban context of Knoxville.

Principal components analyses: place-based perspective

The four components cumulatively explain 66.50 percent of total variation (Table 3). These components along with DS1990 and DS2000are used as independent variables in regression analyses whereas dependent variable(s) consist of: (i) Diversity Score in 2009 (DS2009), (ii) Change in Diversity Score during 1990—2009 (ADS(2009— 1990)), and (iii) Change in Diversity Score for White versus non-White during 1990—2009 (DDSWNW(2009—1990)) Table 4, panels A, B and C respectively). Though using DS1990 and DS2000 as predictors may/will increase the R-square values substantially, I use these to measure their independent effects on the predictability of place-based characteristics. The following briefly explains the characteristics of places/neighborhoods captured in the four components.

Component I (PC-I) gets positive loadings on median household values (3 years), median year of housing structure built, population engaged in managerial/professional occupations in 2000 and 2009, White and Black population in managerial/professional occupation (2009), median household income (three years), per capita incomes for White, Black, Asian, and Hispanic in 1990, per capita incomes for Black, Asian and Hispanic in 2009, and educational achievements of Bachelors/above for overall population in all three years. Negative loadings occur on high school/lesser educated in all 3 years, White with high school/lower educated in 1990, and change in Hispanic share during 1990—2009. This component, thus, characterizes with creative-class/high-income/ newness of a place.

Component II (PC-II) gets positive loadings on median household values(2000, 2009), share of housing structure built (1990— 2000, 2000—2009), median years of housing structure built (1990, 2000, 2009), White homeownership (2009), manufacturing/blue

collar (2009), per capita income (Black, 2009), median household income (1990, 2000, 2009), population (1990, 2000, 2009) and change in total population and for White during 1990—2009. Negative loadings occur on share of foreign-born (F-B) in 2000 and 2009, F-B entered (1990—2000, 2000—2009) out of total 2009 F-B population, Hispanics in managerial (2000), manufacturing/blue collar occupation in 2009 (overall and Black, Asian and Hispanic), and Hispanic homeownership (1990). This component, thus, identifies with newer/medium-to-high income White-Black neighborhood versus others with newly arriving F-B (Black, Latino and Asian) engaged in blue or white collar jobs.

Component III (PC-III) gets positive loadings on non-movers (1985, 1995), high school/lesser educated (1990, 2009), median household income (2000, 2009), blue collar/manufacturing (2000, 2009) and White in manufacturing (2009), White homeownership (1990, 2009), and share of housing structure built during 1990— 2000. Negative loadings occur on Bachelors and/or above educated (all three years), for White and Asian (1990) and for Black and Asian (2000), F-B share (1990, 2000) and F-B entered during 1990—2000. This component, thus, characterizes with stable/blue-collar/white versus educated/Black-Asian-diverse neighborhoods.

Component IV (PC-IV) gets positive loadings on Black with high school/lesser educated (1990), per capita income for Hispanic (2009), managerial occupations (2000), Hispanics in managerial occupations (2009), and Black homeownership (1990, 2009); negative loadings occur on White homeownership (1990,2009). This component characterizes with racially segregated distinctness of Knoxville-Black/Latino diversity versus White only neighborhoods.

Regression analyses

The generic regression equation for dependent variable Y becomes:

Y[DS - 2009, DS(2009 - 1990), DSWNW(2009 - 1990)]

b02*DS2000 + bi*PC - I + b2*PC - II

+b3*PC - III + b4*PC - IV

Where Y is the dependent variable of Diversity Score in 2009 (Y[DS-2009]) or Change in Diversity score during 1990—2009 (YDDS(2009—1990)) or Change in Diversity Score for White vs. Non-white during 1990—2009 (YDDSWNW(2009—1990)).

a is the intercept;

bm is the coefficient on DS1990 when used in the model,

otherwise 0, (Table 4, panel I);

b02 is the coefficient on DS2000 when used in the model,

otherwise 0, (Table 4, panel II); and

bi, b2, b3, b4 are the coefficients associated with PC-I, PC-II, PC-III

and PC-IV in the model (Table 4, panel III).

On the right hand side of the equation,

Using the above specifications to analyze intra-urban variation of diversity in 2009 and its change during 1990—2009, higher values of the dependent variable at a point in time (YDS-2009) represent higher diversity, and positivevalues for change in diversity during 1990 to 2009 [i.e., YDS(1990—2009) and YDSWNW (1990—2009)] represent gain in diversity.

Diversity at a point in time: 2009

Table 4-A suggests that DS2000(Beta = 1.257) serves as a better predictor of DS2009 compared to DS1990(Beta = 0.491), but the role of the four components get overshadowed in the presence of either

b01 *DS1990

Fig. 3. Diversity score-overall in Knoxville MSA, 1990, 2000 and 2009.

of these two variables and the direction of predictability for the components also changes. In panel-III, Table 4-A (four components-models), PC-II, PC-III and PC-IV are significant predictors, with Betas of -0.574, -0.314 and 0.462 respectively. This indicates that neighborhoods with newer/medium-to-high income White-Black neighborhood have lower diversity (Beta = -0.574) as against others with newly arriving F-B engaged in blue/white collar jobs that are more diverse. PC-III (Beta = -0.314) associated with stable/blue-collar/white neighborhoods are less diverse compared to others that have educated, diverse population. Finally, PC-IV (Beta = 0.462) suggests that the diverse (Black/Latino) neighborhoods are more diverse in 2009 whereas white-only/sluggish-low-income are less diverse. Even among lower-income groups, there are pockets of

diverse versus homogeneous clusters such as White-poor clusters in South Knox County and Black/Latino poor clusters toward East of downtown in Knox County.

This analysis also suggests that while diverse neighborhoods of 2000 become more diverse in 2009, they are also transitioning into minority enclaves, as diversity among non-White (1990, 2000) relate with lower diversity in 2009.This is also evident from regression analyses as minority clusters are forming, particularly in blue-collar neighborhoods (PC-III Beta = -0.314). At the same time, there are other neighborhoods with Hispanics in managerial occupation (2009), higher Hispanic per capita income (2009)along with Black homeownership (1990, 2009) that is contributing toward greater diversity (PC-IV Beta = 0.462). These processes

Fig. 4. Diversity score-white versus non-white (left) and among non-white (right) in Knoxville MSA, 1990, 2000 and 2009.

reflect transitioning neighborhoods where filtering and/or invasion-succession (Buzar et al., 2007; Rerat, 2011) occurs with newly arriving/diverse population when they buy/rent in relatively diverse neighborhoods, increasing temporary diversity, and later move elsewhere after gain in socio-economic status (assimilation framework of Alba, Logan, Stults, Marzan, & Zhang, 1999). This analysis also suggests that Hispanics and Blacks are likely co-residing adding to diversity whereas White areas are difficult to change, irrespective of their class — such as affluent White western suburbs (except well off Asians) and poorer white south Knox neighborhoods. At the same time, neighborhoods in east Knoxville (e.g., Fourth & Gill, the Holstein Hills) have gone through gentrifi-cation and now attract newer/younger/relatively more diverse, yuppie and academic community.

Changing diversity as a process: 1990—2009

Models (Table 4-B and C) suggest that PC-II is the only significant predictor (Beta = -0.544) with DS1990 in the model, whereas the model with DS2000 suggests all four components and DS2000 as significant predictors. In the components-only model (panel III, Table 4-B), PC-III is the only significant predictor (Beta = 0.413). These models suggest that neighborhoods with higher diversity in 1990 have negative change (Beta = -0.544, Table 4-B, panel I) whereas those with higher diversity in 2000 have positive change. PC-III, the strongest of all four (Table 4-B, panel II) relates with stable/blue-collar versus newer/dynamic/diverse neighborhoods, and a Beta of 1.124 indicates that stable/blue collar neighborhoods are also becoming more diverse in the recent times, probably from invasion and succession. While this sounds contradictory to what was noted about stable/sluggish neighborhoods pertaining to diversity in 2009 (Beta on PC-III = 0.359, Table 4-A-panel II and -0.314 for panel III), when it comes to change, the diversity

levels in 2000 has a stronger effect on change during the two decades. Also interesting to note is that Beta on PC-III = 0.413 (Table 4-B, panel III) is far lower than Beta = 1.124 (panel II) which indicates that the degree of change during 1990—2009 is much higher in neighborhoods that were diverse in 2000 compared to those that were not. This process can be explained from the Community Norm perspective that offers an alternative way of explaining this change, illustrated as an overall catching-up phenomenon at the scale of MSA (Brown & Sharma, 2010) and within intra-urban context (Sharma & Brown, 2012) such that the neighborhoods that are diverse try to catch up and gain more to keep up with the norm, whereas neighborhoods that are not diverse maintain their inertia, per Community Inertia perspective (Brown & Sharma, 2010; Sharma & Brown, 2012). Concerning the components of diversity, regression tests concluded that models for change in diversity for "non-White" were not significant. Concerning change in diversity for White vs. non-White (Table 4-C), the best model (panel II) has three components as significant predictors and that DS2000 has the most effect on the direction and degree of change during 1990—2009.

Discussions and conclusions

During the last decade (2000s), the whole country has gone through recessions, job losses, housing foreclosures, increase in poverty and unemployment, and Knoxville is no exception. At a time when jobs are very difficult to find, the engagement of minorities (e.g., Blacks, Asians, Hispanics) in blue-collar and/or white-collar jobs (as indicated by PC-II and IV) contributes to overall diversity at neighborhoods irrespective of their homeownership or renter-ship status. At the same time, this analysis also indicates that minority clusters have formed during 1990—2009 in specific

Table 2

Correlations of diversity score and change with neighborhood characteristics and computed indices.

A: Correlations with demographic characteristics DS-2009 DDS (2009—1990)

Total population, 2009 -0.238** 0.117

Change, population, 1990—2009, share -0.344** 0.014

Change, White, 1990—2009, share -0.227** 0.019

Change, Black, 1990—2009, share 0.008 0.384'

Change, Asian, 1990—2009, share 0.005 0.218

Change, Hispanic, 1990—2009, share 0.053 0.263

Proportion foreign born, 1990 0.314** 0.01

Proportion foreign born, 2009 0.521** 0.498

Foreign-born entered during 1990—2000, as share of total F-B in 2000 0.300** 0.061

Foreign-born entered during 2000—2009, as share of total F-B in 2009 0.124 0.262

Population lived in same house, 1985 (share 1990 population) -0.402** -0.078

Population lived in same house, 1995 (share 2000 population) -0.427** -0.102

B: Correlations with socio-economic characteristics DS-2009 DDS (2009—1990)

Bachelors and/or graduate degree 1990, proportion) 0.263** 0.159

Bachelors and/or graduate degree 2009, proportion) 0.194* 0.175'

Bachelors and/or graduate degree, White, 1990, proportion) 0.187* 0.178

Bachelors and/or graduate degree, Black, 1990, proportion) 0.187* 0.178

Bachelors and/or graduate degree, Asian, 1990, proportion) 0.192* -0.089

Median household income, 1999 -0.235** 0.085

Median household income, 2009 -0.254** 0.052

Per capita income, African American (1990) 0.260** 0.192

Per capita income, African American (2009) -0.317** -0.007

C: Correlations with built-environment characteristics DS-2009 DDS (2009—1990)

Manuf./warehouse/trans.-empl. 1990, as share of total empl. -0.386** -0.018

Manuf./warehouse/trans.-empl. 2009, as share of total empl. -0.313** -0.047

Manuf./warehouse/trans.-black, empl. 2009, as share of total empl. 0.328** 0.112

Managl./prof.empl. Black, 2009, as share of total empl. 0.565** 0.058

Manuf./warehouse/trans.-Asian, empl. 2009, as share of total empl. 0.282** 0.170'

Managl./prof.empl. Asian, 2009, as share of total empl. 0.095 0.084

Manuf./warehouse/trans.-Hispanic, empl. 2009, as share of total empl. 0.213* 0.178

Managl./prof.empl.-Hispanic, empl. 2009, as share of total empl. 0.251** 0.322

Homeownership, White, 1990 -0.504** 0.036

Homeownership, African-American, 1990 0.476** -0.03

Homeownership, Asian, 1990 0.251** -0.064

Homeownership, White, 2009 -0.659** -0.134

Homeownership, African-American, 2009 0.502** 0.023

Homeownership, Hispanic, 2009 0.394** 0.464

Median year housing, 2009 -0.425** -0.063

Share, housing built 1990—2000, of 2000 HH-Stock -0.489** -0.056

Share, housing built, 2000—2009, of 2009 HH-Stock -0.262** 0.006

D: Correlations with computed indices DS-2009 DDS (2009—1990)

Diversity score, 1990 0.773** -0.103

Diversity score, among non-White, 1990 -0.239** 0.003

Diversity score, White vs. non-White, 1990 0.762** -0.104

Diversity score, 2000 0.821** 0.123

Diversity score, among non-White, 2000 -0.410** 0.02

Diversity score, White vs. non-White, 2000 0.818** 0.102

Diversity score, 2009 1 0.552

Diversity score, among non-White, 2009 -0.039 0.310

Diversity score, White vs. non-White, 2009 0.973** 0.463

Change in diversity score, 2009—1990 0.552** 1

Change in diversity score, 2009—2000 0.354** 0.738

Change in diversity score-among non-White, 2009—1990 0.174* 0.277

Change in diversity score-among non-White, 2009—2000 0.379** 0.299

Note: Bold for correlations >±0.250.

DS-2009 is Diversity Score in 2009; DDS (2009—1990) is Change in Diversity Score from 1990 to 2000.

** Correlation is significant at the 0.01 level; * Correlation is significant at the 0.05 level (2-tailed).

pockets of the metropolitan area, in accordance with the Resurgent Ethnicity framework (Brown & Chung, 2006; Charles, 2003). Diversity that was confined inside Knox County in 1990 has spread to Knox, Loudon, Anderson, and Blount counties by 2000, and further into Union and Sevier counties by 2009 with the hot-spots of diversity visible all across the metropolitan area. In particular, diversity in the adjoining counties are most evident in Alcoa, Sevierville, Pigeon Forge, Gatlinburg, Loudon and Oak Ridge — though the reasons for gain in diversity in these cities range from historical presence of Black in Alcoa to availability of service sector

economy in Sevierville, Gatlinburg, and Pigeon Forge, and newer urban development and proximity to ORNL for Loudon, and hightech, research and development jobs in Knoxville. Also, gain in diversity occurs for neighborhoods that were more diverse in 2000, whereas those with higher diversity in 1990 become less diverse by 2009(r = -0.103 with ADS(2009—1990). These processes provide evidence of Community Norm and Community Inertia perspectives at work. This study also provides evidence of invasion/succession and filtering (Buzar et al., 2007) as inner-city locations that were previously occupied by white segments get emptied over time with

Table 3

Principal components analyses.

Rotated components matrix

Components

PC-III

Demographic characteristics

Total population, 1990 Total population, 2000 Total population, 2009 Change, population, 1990—2009, share Change, White, 1990—2009, share Change, Black, 1990—2009, share Change, Asian, 1990—2009, share Change, Hispanic, 1990—2009, share Proportion foreign born, 1990 Proportion foreign born, 2000 Proportion foreign born, 2009

Foreign-born entered during 1990—2000, as share of total F-B in 2000

Foreign-born entered during 2000—2009, as share of total F-B in 2009

Population lived in same house, 1985 (share 1990 population)

Population lived in same house, 1995 (share 2000 population)

Socio-economic characteristics

High school or less education (1990, proportion)

Bachelors and/or graduate degree (1990, proportion)

High school or less education (2000, proportion)

Bachelors and/or graduate degree (2000, proportion)

High school or less education (2009, proportion)

Bachelors and/or graduate degree (2009, proportion)

High school or less education, White, (1990, proportion)

High school or less education, Black, (1990, proportion)

High school or less education, Asian, (1990, proportion)

Bachelors and/or graduate degree, White (2000, proportion)

Bachelors and/or graduate degree, Black (2000, proportion)

Bachelors and/or graduate degree, Asian (2000, proportion)

Bachelors and/or graduate degree, Hispanic (2000, proportion)

Median household income, 1989

Median household income, 1999

Median household income, 2009

Per capita income, White, 1990

Per capita income, African American, 1990

Per capita income, Asian, 1990

Per capita income, Hispanic, 1990

Per capita income, African American, 2009

Per capita income, Asian, 2009

Per capita income, Hispanic, 2009

Built-environment characteristics

Manuf./warehouse/trans.-empl. 2000, as share of total empl.

Managl./prof.empl. 2000, as share of total empl.

Manuf./warehouse/trans.-empl. 2009, as share of total empl.

Managl./prof.empl. 2009 as share of total empl.

Managl./prof.empl. White, 2009, as share of total empl.

Manuf./warehouse/trans.-white, empl. 2009, as share of total empl.

Managl./prof.empl. Black, 2009, as share of total empl.

Manuf./warehouse/trans.-Black, empl. 2009, as share of total empl.

Managl./prof.empl. Asian, 2009, as share of total empl.

Manuf./warehouse/trans.-Asian, empl. 2009, as share of total empl.

Managl./prof.empl. Hispanic, 2009, as share of total empl.

Manuf./warehouse/trans.-Hispanic, empl. 2009, as share of total empl.

Homeownership, White, 1990

Homeownership, African-American, 1990

Homeownership, Asian, 1990

Homeownership, Hispanic, 1990

Homeownership, White, 2009

Homeownership, African-American, 2009

Homeownership, Asian, 2009

Homeownership, Hispanic, 2009

Median year housing, 1990

Median year housing, 2000

Median year housing, 2009

Share, housing built 1990—2000, of 2000 HH-Stock Share, housing built, 2000—2009, of 2009 HH-Stock Median value of owner occupied housing, 1990 Median value of owner occupied housing, 2000 Median value of owner occupied housing, 2009 Percent variance accounted for, by component Percent variance accounted for, cumulatively

0.058 0.176 0.186 0.240 0.183 0.273 0.259 -0.376 0.040 0.043 0.293 -0.150 -0.343 -0.215 0.026

-0.780 0.779 -0.759 0.874 -0.823 0.899 -0.717 -0.231 -0.122 0.919 0.405 -0.017 0.380 0.784 0.701 0.640 0.900 0.712 0.747 0.609 0.537 0.534 0.263

-0.628 0.506 -0.373 0.909 0.857 -0.673 0.400 -0.124 0.328 0.326 -0.116 0.060 0.147 -0.186 -0.070 0.429 -0.109 -0.046 -0.080 0.013 0.464 0.385 0.278 0.220 0.249 0.877 0.811 0.833 25.830 25.830

0.523 0.696 0.747 0.737 0.787 0.337 0.077 0.260 -0.161 -0.393 -0.610 -0.675 -0.530 0.089 0.317

-0.108 0.031 -0.224 0.056 -0.174 0.093 -0.085 -0.070 -0.221 -0.056 0.024 -0.134 -0.592 0.461 0.536 0.580 0.238 0.277 0.152 0.021 0.549 0.289 0.075

-0.158 -0.476 0.182 0.316 0.159 0.128 -0.668 -0.005 -0.439 -0.572 0.007 0.239 -0.198 -0.067 -0.417 0.443 0.155 0.129 -0.120 0.690 0.699 0.767 0.716 0.615 0.335 0.412 0.380 16.584 42.414

0.067 0.185 0.164 0.306 0.246 0.279 0.197 0.162 -0.927 -0.835 -0.230 -0.388 -0.223 0.880 0.801

-0.582 -0.071 -0.427 0.455 -0.360 0.595 -0.082 -0.816 -0.219 -0.445 -0.873 -0.276 0.311 0.375 0.386 0.170 0.134 0.059 -0.076 0.220 0.090 0.019

0.094 0.371

0.033 0.138 0.443 -0.344 0.248 -0.060 -0.208 0.206 0.097 0.473 -0.156 -0.894 0.142 0.386 0.000 0.205 0.061 0.174 0.196 0.117 0.390 -0.033 0.113 -0.103 0.125 14.848 57.262

-0.224 -0.234 -0.176 -0.048 0.003 -0.077 -0.065 -0.105 0.102 0.046 -0.282 0.046 -0.106 0.155 0.121

-0.053 -0.030 0.023 -0.023 0.226 -0.094 -0.259 0.942 0.037 0.086 0.123 -0.031 -0.255 0.027 -0.024 -0.101 0.149 0.117 0.055 0.007 -0.117 0.149 0.764

-0.050 0.589

-0.122 -0.067 -0.131 -0.027 -0.006 -0.106 -0.114 -0.125 -0.126 0.813 -0.809 0.921 0.110 0.092 -0.656 0.874 -0.063 0.154 -0.060 -0.248 -0.229 -0.156 -0.105 -0.120 -0.038 0.018 9.240 66.502

Note: A cutoff value of ±0.35 was considered significant when gauging Component Loadings for Interpretation. Cells > +0.35 indicated by bold. Cells < -0.35 indicated by italics.

Table 4

Regression models for diversity score in 2009, change in diversity during 1990—2009, and change in diversity score for white vs non-white during 1990—2009.

A. Dependent variable: diversity score in 2009 (DS2009)

Panel I: with DS-1990 Panel II: with DS-2000 Panel III: only four components

Independent Y = DS-2009 Independent Y = DS-2009 Independent Y = DS-2009

Variables b Beta t-Value p-Value Variables b Beta t-Value p-value Variables b Beta t-Value p-Value

PC-I 0.028 0.152 1.140 0.269 PC-I 0.046 0.250 2.946 0.009 PC-I 0.024 0.129 0.962 0.348

PC-II -0.074 -0.401 -2.108 0.049 PC-II 0.030 0.160 1.053 0.306 PC-II -0.107 -0.574 -4.278 0.000

PC-III -0.002 -0.011 -0.039 0.970 PC-III 0.067 0.359 2.505 0.022 PC-III -0.058 -0.314 -2.336 0.031

PC-IV 0.029 0.155 0.564 0.579 PC-IV -0.025 -0.133 -1.006 0.328 PC-IV 0.086 0.462 3.441 0.003

DS-1990 0.491 0.494 1.269 0.221 DS-2000 1.127 1.257 5.725 0.000 None x x x x

Constant 0.425 x 3.162 0.005 Constant 0.036 x 0.370 0.715 Constant 0.592 x 24.307 0.000

r-Value 0.828 r-Value 0.937 r-Value 0.811

r-Squared 0.686 r-Squared 0.879 0.658

B. Dependent variable: change in diversity score from 1990 to 2009, DS (2009- 1990)

Panel I: with DS-1990 Panel II: with DS-2000 Panel III: only four components

Independent Y = DS (2009-1990) Independent Y = DS (2009-1990) Independent Y = DS (2009-1990)

Variables b Beta t-Value p-Value Variables b Beta t-Value p-Value Variables b Beta t-Value p-Value

PC-I 0.028 0.206 1.140 0.269 PC-I 0.050 0.366 2.417 0.026 PC-I 0.033 0.238 1.305 0.207

PC-II -0.074 -0.544 -2.108 0.049 PC-II 0.065 0.477 1.754 0.096 PC-II -0.041 -0.299 -1.638 0.118

PC-III -0.002 -0.014 -0.039 0.970 PC-III 0.154 1.124 4.389 0.000 PC-III 0.056 0.413 2.259 0.036

PC-IV 0.029 0.211 0.564 0.579 PC-IV -0.116 -0.850 -3.593 0.002 PC-IV -0.030 -0.221 -1.210 0.241

DS-1990 -0.509 -0.697 -1.318 0.204 DS-2000 0.878 1.330 3.388 0.003 None x x x x

Constant 0.425 x 3.162 0.005 Constant -0.183 x -1.412 0.175 Constant 0.251 x 10.248 0.000

r-Value 0.649 r-Value 0.783 r-Value 0.605

r-Squared 0.422 r-Squared 0.613 r-Squared 0.367

C. Dependent variable: change in diversity score for white vs. non-White from 1990 to 2009, DS-WNW (2009—1990)

Panel I: with DS-1990 Panel II: with DS-2000 Panel III: only four components

Independent Y = DS-WNW (2009-1990) Independent Y = DS-WNW (2009-1990) Independent Y = DS-WNW (2009-1990)

Variables b Beta t-Value p-Value Variables b Beta t-Value p-Value Variables b Beta t-Value p-Value

PC-I 0.020 0.198 1.163 0.260 PC-I 0.034 0.341 2.181 0.043 PC-I 0.023 0.235 1.339 0.196

PC-II -0.052 -0.522 -2.152 0.045 PC-II 0.040 0.404 1.439 0.167 PC-II -0.024 -0.243 -1.385 0.182

PC-III -0.002 -0.015 -0.044 0.966 PC-III 0.106 1.064 4.027 0.001 PC-III 0.047 0.472 2.693 0.014

PC-IV 0.021 0.210 0.597 0.558 PC-IV -0.081 -0.807 -3.305 0.004 PC-IV -0.028 -0.283 -1.616 0.123

DS-1990 -0.425 -0.795 -1.600 0.127 DS-2000 0.534 1.107 2.734 0.014 None x x x x

Constant 0.294 x 3.191 0.005 Constant -0.114 x -1.173 0.256 Constant 0.149 x 8.703 0.000

r-Value 0.700 r-Value 0.767 r-Value 0.646

r-Squared 0.489 r-Squared 0.588 r-Squared 0.417

Note: None of the regression models were significant when explaining change in diversity among the non-white groups during 1990—2009, hence I don't show those models here.

gain in SES, and are filled in by newly-arriving foreigners and minorities, per Assimilation framework of Alba et al. (1999).

Earlier research has suggested that systemic factors, such as lack of recognition of foreign credentials (Papillon, 2002), in this case the educational degrees of immigrant populations and Hispanics being treated as "the outsiders", racial discrimination and prejudice in the work environment, as well as lack of access to affordable housing and suitable language training may contribute to social exclusion of more vulnerable newcomers (Jones, 1995; Papillon, 2002; Shmueli & Kipnis, 1998). This, I believe, needs to be addressed in context of Knoxville and other small-to-mid-sized metropolises in USA that dream of creating sustainable urban entities. As indicated earlier (Séguin et al., 2012; Shmueli & Kipnis, 1998; Sharma, 2011a; Watson, 2006; Wilson et al., 2008), such research on diversity and sustainability must be used to conceptualize and implement comprehensive strategies focused at mobilizing residents at the grassroots level to address public policy.

This study provides useful information about changing contexts of Knoxville's neighborhoods that can be useful for participative planning and strategic community development. In a society where racial/ethnic diversity continues to increase, it has its own implications on changing attitudes, trust and relationships toward each other, as indicated by Letki (2008). In particular, Knoxville is still 88.9% White, but is fast changing, and field reconnaissance by the author indicates that Hispanics, in particular, are being looked down upon as "the outsiders". This requires creating a safe and welcoming environment for all, including Hispanics, Asians and

other groups. Knoxville is home to a diverse set of economy, and with holistic planning, it can become a creative class city, in accordance with Florida's (2004) recipe about making a community progressive and sustainable in the long run.

From the perspective of planning and zoning, earlier research indicates the discomfort of staying in a heterogeneous community creates segregated spaces, eventually forming spaces of difference, disinvestment, urban blight and poverty concentration (Jones, 1995; Séguin et al., 2012; Smith, 1986; Weaver & Baghchi-Sen, 2013). Drawing parallels with Knoxville, investors have continued to pour in big dollars in commercial and residential development in the west, whereas south Knox County remains the most dilapidated and poverty stricken part of the MSA (field reconnaissance by the author). The city continues to invest in its renovated Market Square and high-rise expensive condominiums in the downtown area whereas just few blocks away to the east remains the ghetto, struck with the metro's highest rates of crime (http:// www.spotcrime.com/tn/knoxville). Such issues demotivate private and public investors that eventually create spaces of difference, exacerbating social and economic inequality. Planners and policy makers in Knoxville must invest in equitable and sustainable development to accommodate the demands of growing diversity in terms of public infrastructure, social, cultural, health, economic and human capital needs, as has also been suggested by other scholars (see Acevedo-Garcia et al., 2008; Jones, 1995; Letki, 2008; Papillon, 2002; Shmueli & Kipnis, 1998; Vliet, 1996; Watson, 2006; Weaver & Baghchi-Sen, 2013). To improve the overall well-being of

communities, and to reduce racial, ethnic and socio-economic disparities, one must consider comprehensive strategies that integrate the best urban planning approaches such as the smart growth, sustainability, new urbanism, and socially just practices.

References

Acevedo-Garcia, D., & Osypuk, T. L. (2008). Invited commentary: residential segregation and health — the complexity of modeling separate social contexts. American Journal of Epidemiology, 168(11), 1255—1258. Acevedo-Garcia, D., Osypuk, T. L., McArdle, N., & Williams, D. R. (2008). Toward a policy-relevant analysis of geographic and racial/ethnic disparities in child health. Health Affairs, 27(2), 321—333. Alba, R. D., Logan, J. R., Stults, B. J., Marzan, G., & Zhang, W. (1999). Immigrant groups in the suburbs: a re-examination of suburbanization and spatial assimilation. American Sociological Review, 64, 446—460. Alba, R., & Nee, V. (2003). Remaking the American mainstream: Assimilation and

contemporary immigration. Cambridge: Harvard University Press. Bobo, L., & Zubrinsky, C. L. (1996). Attitudes on residential integration: perceived status difference, mere in-group preferences, or racial prejudice? Social Forces, 74(3), 883—909.

Brown, L. A., & Chung, S.-Y. (2006). Spatial segregation, segregation indices and the

geographical perspective. Population, Space and Place, 12,125—143. Brown, L. A., & Chung, S. (2008). Market-led pluralism: re-thinking our understanding of racial/ethnic spatial patterning in US cities. Annals of the Association of American Geographers, 98,180—212. Brown, L. A., & Sharma, M. (2010). Metropolitan context and racial/ethnic intermixing in residential space: US MSAs 1990—2000. Urban Geography, 31,1—28. Buzar, S. N., Philip, O., Hall, R., Haase, A., Kabisch, S., & Steinfuher, A. (2007). Splintering urban populations: emergent landscapes of reurbanisation in four European cities. Urban Studies, 44, 651—67 . Chaney, J. (2010). The formation of a Hispanic enclave in Nashville, Tennessee.

Southeastern Geographer, 50(1), 17—38. Charles, C. Z. (2000). Neighborhood racial-composition preferences: evidences from

a multiethnic metropolis. Social Problems, 47, 379—40 . Charles, C. Z. (2003). The dynamics of racial residential segregation. Annual Review

of Sociology, 29,170—20 . Chung, S.-Y. (2005). Intra-urban segregation changes: an evaluation of three segregation frameworks with a case study of Columbus, Ohio MSA, 1990 and 2000. Ohio State University, Department of Geography. PhD Dissertation. Clark, W. A. V., & Blue, S. (2004). Race, class, and segregation patterns in U.S.

immigrant gateway cities. Urban Affairs Review, 39, 667—688. Cobb, J. C. (2005). Beyond the 'y'all wall: the American south goes global. In J. C. Cobb, & W. Stueck (Eds.), Globalization and the American south (pp. 1—18). Athens: University of Georgia Press. Connerly, C. E. (2005). The most segregated city in America, city planning and civil rights in Birmingham 1920-80. Charlottesville, VA: University of Virginia Press. Cornelius, W. A., Fitzgerald, D., Fischer, P. L., & Muse-orlinoff, L. (2010). Mexican migration and the U.S. economic crises: A transnational perspective. Regents: University of California at San Diego. Denton, N. A., & Massey, D. S. (1991). Patterns of neighborhood transition in a multiethnic world: U.S. metropolitan areas, 1970—1980. Demography 28(1), 41—63.

Dingemans, D., & Datel, R. (1995). Urban multi-ethnicity. Geographical Review, 85, 458—477.

Duncan, D., & Duncan, B. (1955). A methodological analysis of segregation indexes.

American Sociological Review, 20, 210—217. Ellis, M., Holloway, S., & Wright, R. (2007). The effects of mixed-race households on

residential segregation. Urban Geography, 28(6), 554—577. Ellis, M., Wright, R., & Parks, V. (2004). Work together, live apart? Geographies of racial and ethnic segregation at home and at work. Annals of the Association of American Geographers, 94, 620—637. Farley, R. (1996). Black-White segregation: the views of Myrdal in the 1940s and trends of the 1980s. In O. Clayton (Ed.), An American dilemma revisited: Race relations in a changing world. New York: Russell Sage Foundation. Farley, R., & Frey, W. H. (1994). Changes in the segregation of Whites from Blacks during the 1980s: small steps toward a more integrated society. American Sociological Review, 59, 23—45. Fischer, C. S., Stockmayer, G., Stiles, J., & Hout, M. (2004). Distinguishing the geographic levels and social dimensions of U.S. metropolitan segregation, 1960—2000. Demography 41(1), 37—59. Florida, R. (2004). The rise of the creative class: And how it is transforming work,

leisure, community and everyday life. Basic Books. Flory, J. (June 4, 2011). Lisega dedicates Sevier Plant. last accessed 10.07.12 http://

www.knoxnews.com/news/2011/jun/04/lisega-dedicates-sevier-plant/. Frey, W. H., & Farley, R. (1996). Latino, Asian, and Black segregation in U.S. metropolitan areas: are multiethnic metros different? Demography, 33, 35—50. Geib, P. (1998). From Mississippi to Milwaukee: a case study of the southern Black migration to Milwaukee, 1940—1970. The Journal of Negro History, 83, 229—248.

Grady, S., & Darden, J. (2012). Spatial methods to study local racial residential segregation and infant health in Detroit, Michigan. Annals of the Association of American Geographers, 102(5), 922—931.

Grekousis, G., & Thomas, H. (2012). Comparison of two fuzzy algorithms in geo-demographic segmentation analysis: the fuzzy C-means and GustafsoneKessel methods. Applied Geography 34,125-136.

Holloway, S. R., Wright, R., & Ellis, M. (2012). The racially fragmented city? Neighborhood racial segregation and diversity jointly considered. The Professional Geographer, 64(1), 63-82.

http://www.spotcrime.com/tn/knoxville, last accessed on May 12, 2013.

http://www.retirementplacesreport.com/tennessee_cities.html, last accessed on January 30, 2013.

http://web.knoxnews.com/jobs/knoxville/employ.shtml, last accessed on May 12, 2013

Iceland, J. (2004). The multi-group entropy index. Working paper, U.S. Census Bureau www.census.gov/hhes/www/housing/housing_patterns/multigroup_entropy. pdf.

Isserman, A. M. (1977). The location quotient approach to estimating regional economic impacts. Journal of the American Planning Association, 43(1), 33-41.

Jones, H. (1995). The continuing ethnic-origins dimension of Australian immigrantion policy. Applied Geography, 15(3), 233-244.

Labor force estimates. March 2012. http://www.wbir.com/news/article/213613/2/ Unemployment-rates-fall-in-most-Tennessee-counties, last accessed on May 2, 2013.

Leigh, R. (1970). The use of location quotients in urban economic base studies. Land Economics, 46(2), 202-205.

Letki, N. (2008). Does diversity erode social cohesion? Social capital and race in British neighbourhoods. Political Studies, 56, 99-126.

Levine, M. (2000). The economic state of black Milwaukee. In S. F. Battle, & R. Hornung (Eds.), The state of black Milwaukee. Milwaukee, WI: Milwaukee Urban League:171-193.

Lichter, D. T., Parisi, D., Grice, S. M., & Taquino, M. (2006). National estimates of racial segregation in rural and small town America. Paper Presented at the Population Association of America, Los Angeles meeting.

Lobo, A. P., Flores, R J. O., & Salvo, J. (2002). The impact of Hispanic growth on the racial/ethnic composition of New York City neighborhoods. Urban Affairs Review, 37, 703-727.

Logan, J., Alba, R. D., Dill, M., & Zhou, M. (2000). Ethnic segmentation in the American metropolis: increasing divergence in economic incorporation, 19801990. International Migration Review, 34, 98-132.

Logan, J. R., Stults, B. J., & Farley, R. (2004). Segregation of minorities in the Metropolis: two decades of change. Demography, 41,1-22.

Logan, J. R., Zhang, W., & Alba, R. D. (2002). Immigrant enclaves and ethnic communities in New York and Los Angeles. American Sociological Review, 67(2), 299-322.

McDaniel, P. N., & Drever, A. I. (2009). Ethnic enclave or international corridor? Immigrant businesses in a new south City. Southeastern Geographer, 49(1), 3-23.

Moineddin, R., Beyene, J., & Boyle, E. (2003). On the location quotient confidence interval. Geographical Analysis, 35(3), 249-256.

Monkkonen, P. (2010). Measuring residential segregation in urban Mexico: Levels and patterns. IURD Working Paper Series. Institute of Urban and Regional Development, UC Berkeley. Last accessed 05.09.13 http://escholarship.org/uc/item/ 6hf1q7hg.

Osypuk, T. L., Bates, L. M., & Acevedo-Garcia, D. (2010). Another Mexican birth-weight paradox? The role of residential enclaves and neighborhood poverty in the birthweight ofMexican-Origin Infants. Social Science & Medicine, 70(4), 550. [NIH Public Access Manuscript].

Papillon, M. (2002). Immigration, diversity and social inclusion in Canada's cities. Discussion Paper F/27 family Network. Canadian Policy Research Networks Inc.

Perreira, K. M. (2011). Mexican families in North Carolina: the socio-historical contexts of exit and settlement. Southeastern Geographer, 51 (2), 260-286.

Reardon, S. F., & Firebaugh, G. (2002). Measures of multigroup segregation. Sociological Methodology, 32, 33-6 .

Reardon, S. F., & Sullivan, D. (2004). Measures of spatial segregation. Sociological Methodology, 34(1), 121-162.

Rerat, P. (2011). The new demographic growth of cities: the case of reurbanisation in Switzerland. Urban Studies, 49,1107-1125.

Roberts, B. R., & Wilson, R. H. (Eds.). (2009). Urban segregation and governance in the Americas. New York: Palgrave-MacMillan.

Séguin, A.-M., Apparicio, P., & Riva, M. (2012). Identifying, mapping and modelling trajectories of poverty at the neighbourhood level: the case of Montréal, 19862006. Applied Geography, 35, 265-274.

Sandoval, J., & Li, S. (2004). The multicultural metropolis: changing segregation patterns in the city of Chicago from 1990 to 2000. (Unpublished Paper presented at the Annual meeting of the American Sociological Association, San Francisco, CA).

Sassen, S. (1991). The global city. Princeton, NJ: Princeton University Press.

Sharma, M., & Brown, L. A. (2012). Racial/ethnic intermixing in intra-urban space and socio-economic context: Columbus Ohio and Milwaukee Wisconsin. Urban Geography, 33(3), 317-34 .

Sharma, M. (2011a). A geographic perspective on intra-urban racial/ethnic diversity, segregation and clustering in Knoxville, Tennessee: 1990-2000. Applied Geography 32, 310-323.

Sharma, M. (2011b). Asians and Hispanics in the American South: the case of Tennessee. In J. W. Frazier, E. L. Tettey-Fio, & N. F. Henry (Eds.), Race, ethnicity and place in a changing America (pp. 307-328). Albany: State University of New York Press.

Shmueli, D. F., & Kipnis, B. A. (1998). Participatory planning and ethnic interaction in Ma'alot—Tarshiha, a Jewish—Arab community. Applied Geography, 18(3), 225—241.

Shultz, B. J. (2008). Inside the gilded cage: the lives of Latino immigrant males in rural central Kentucky. Southeastern Geographers, 48(2), 201—218.

Singer, A. (2003). At home in the nation's capital: Immigrant trends in metropolitan Washington. Washington: The Brookings Institution, Metropolitan Policy Program.

Singer, A. (2004). The rise of new immigrant gateways. Washington: The Brookings Institution, Metropolitan Policy Program.

Smith, N. (1986). Gentrification, the frontier, and the restructuring of urban space from gentrification of the city (pp. 128—137). New York: Routledge. This chapter adopted from The Urban Geography Reader. Eds. Nicholas R. Fyfe and Judith T. Kenny. 2005.

Smith, H. A., & Furuseth, O. J. (2004). Housing, Hispanics and transitioning geographies in charlotte, North Carolina. Southeastern Geographer, 44, 216—235.

Theil, H., & Finezza, A.J. (1972). A note on the measurement of racial integration of schools by means of informational concepts. Journal of Mathematical Sociology, 1,187—194.

Timberlake, J. M., & Iceland, J. (2007). Change in racial and ethnic residential inequality in american cities, 1970—2000. City and Community, 6, 335—365.

Vliet, W. V. (1996). Sustainable development, global restructuring and immigrant housing. Habitat International, 20(3), 349—358.

Watson, V. (2006). Deep difference: diversity, planning and ethics. Planning Theory, 591), 31—50.

Weaver, R. C., & Baghchi-Sen, S. (2013). Spatial analyses of urban decline: the geography of blight. Applied Geography, 40, 61—70.

Winders, J. (2006). New Americans' in a 'New-South' city? Immigrant and refugee politics in the Music City. Social and Cultural Geography, 7(3), 421—435.

Winders, J. (2011a). Re-placing southern geographies: the role of latino migration in transforming the south, its identities, and its study. Southeastern Geographer, 51(2), 342—358.

Winders, J. (2011b). Representing the immigrant: social movements, political discourse, and immigration in the U.S. South. Southeastern Geographer, 51 (4), 596—614.

Wilson, B. (1992). Structural imperatives behind racial change in Birmingham, Alabama. Antipode, 24(3), 171—202.

Wilson, B. (2000). America's Johannesburg. Lanham, MD: Rowman & Littlefield.

Wilson, S., Hutson, M., & Mujahid, M. (2008). How planning and zoning contribute to inequitable development, neighborhood health and environmental injustice. Environmental Justice, 1(4), 211—216.

Wright, R.. , Holloway, S., & Ellis, M. (2011). Reconsidering both diversity and segregation: a reply to Poulsen, Johnston and Forrest, and to Peach,. Journal of Ethnic and Migration Studies, 37(1), 167—176.

Wong, D. W. S. (2003). Spatial decomposition of segregation indices: a framework toward measuring segregation at multiple levels. Geographical Analysis, 35(3), 179—194.

Wong, D. W. S. (2008). Conceptual and operational issues in incorporating segregation measurements in Baranzini. In A. Jose Ramirez, C. Schaerer, & P. Thalmann (Eds.), Hedonic price modeling in hedonic methods in housing market: Pricing environmental amenities and segregation. Springer.

Diversity in Knoxville: An applied perspective Academic research paper on "Social and economic geography"

Abstract of research paper on Social and economic geography, author of scientific article — Madhuri Sharma

Similar topics of scientific paper in Social and economic geography , author of scholarly article — Madhuri Sharma

Academic research paper on topic "Diversity in Knoxville: An applied perspective"