Sie sind auf Seite 1von 18

ARTICLE IN PRESS

Health & Place 13 (2007) 1431 www.elsevier.com/locate/healthplace

Social space, social class and Bourdieu: Health inequalities in British Columbia, Canada
Gerry Veenstra
Department of Anthropology and Sociology, University of British Columbia, 6303 N.W. Marine Dr., Vancouver, BC, Canada V6T 1Z1

Abstract This article adopts Pierre Bourdieus culturalstructuralist approach to conceptualizing and identifying social classes in social space and seeks to identify health effects of class in one Canadian province. Utilizing data from an original questionnaire survey of randomly selected adults from 25 communities in British Columbia, social (class) groupings dened by cultural tastes and dispositions, lifestyle practices, social background, educational capital, economic capital, social capital and occupational categories are presented in visual mappings of social space constructed by use of exploratory multiple correspondence analysis techniques. Indicators of physical and mental health are then situated within this social space, enabling speculations pertaining to health effects of social class in British Columbia. r 2005 Elsevier Ltd. All rights reserved.
Keywords: Social space; Social class; Economic capital; Cultural capital; Social capital

Introduction Three general approaches to the conceptualization of social class can be identied in the public health literature: (i) social class is equated with socio-economic status, i.e., income, educational attainment and/or occupational prestige, (ii) social class is a social group dened primarily by the nature of its relationship to the economic mode of production, a Marxist perspective, and (iii) social class is a social group dened relationally in social space by its possession and utilization of various capitals such as economic, cultural and social capital, a perspective based in the work of the French theorist Pierre Bourdieu.
E-mail address: Gerry.Veenstra@ubc.ca. 1353-8292/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.healthplace.2005.09.011

While the body of work attributable to the rst two approaches to social class is immense, very few health scholars have applied Bourdieus relational perspective on classes to health inequalities (see Williams, 1995; Poland, 2000; Frohlich et al., 2001; Gatrell et al., 2004). After describing some of the ways in which Bourdieus culturalstructuralist approach to conceptualizing social class differs from the other two approaches, this article utilizes original questionnaire survey data from the province of British Columbia to provide an exploratory empirical application of Bourdieus perspective to health inequalities in Canada. The rst approach to class is located analytically at the level of the individual, equating social class with the social standing of individuals in various status hierarchies. (A Weberian focus on status, distinct from his conceptualization of class, is apparent here.) Although the number of status

ARTICLE IN PRESS
G. Veenstra / Health & Place 13 (2007) 1431 15

hierarchies operative in a given society are undoubtedly close to innite, health researchers have primarily focused on measures of income, educational a and/or occupational prestige, each of which enables straightforward comparisons between individuals within linear status hierarchies. Hypothesized causal pathways from socio-economic status to health and well-being are also primarily individual-level in character, including such phenomena as stress, shame, learned effectiveness, health behaviours and material circumstances that often differ by status level. Of the three approaches to class and health, this approach has generated the largest body of empirical research (see Kaplan and Keil, 1993; Feinstein, 1993; Adler et al., 1994; Adler and Ostrove, 1999; Mirowsky and Ross, 2003 for reviews of the literature). From this literature it is clear that educational experiences and credentials, economic wealth and occupational prestige are persistent correlates of health and wellbeing in most or all developed nations, including Canada. Some health researchers (e.g., Muntaner and Lynch, 1999; Coburn, 2000; Scambler and Higgs, 2001) have taken exception to the use of the term class in the aforementioned body of research, however, instead advocating for a conception of social class embedded in the ontology of Marxist scholarship. For these scholars, a social class is a social group with an identity and existence above and beyond any one individual, implying that an aggregate of individuals who share a given characteristic is not necessarily equal to a social class. The most important social phenomena for health inequalities are group-level relations, therefore, wherein the health of individuals ow from (are a by-product of) class relations. In most Marxist scholarship, social classes are groups dened primarily by the nature of their relationship to the economic mode of production: thus economic ownership and control over productive facilities, budgets and other employees are some of the analytical criteria that serve to delimit the outlines of social classes (see the work of Wright, 1978, 1979, 1985, 1998 for instance). In contrast with the socioeconomic hierarchies described above, class positions are not always arrayed in linear fashion from highest to lowest (Grabb, 1997). The body of work applying this approach to social class and health is also impressively large, with an especially long lineage in the United Kingdom (e.g., Bartley et al., 1996; Hattersley,

1997; Borooah, 1999; Fitzpatrick and Dollamore, 1999; Prandy, 1999; Chandola and Jenkinson, 2000). Far less work of this kind can be found in the United States, however (although see Schwalbe and Staples, 1986; Boffetta et al., 1997; Muntaner et al., 1998), and almost none pertains to Canada (but see Veenstra, 2006). At the risk of oversimplifying the insights produced by this complex and wide-ranging body of work, I believe that this research has yet to clearly demonstrate the existence of social classes and illuminate the inter-class dynamics that inuence health. Rather, some of this research has shown the relevance of control in the workplace for healthbut likely because control over oneself and other employees is less stressful and thus more health promoting than being under the control of others. Other research has documented relationships between economic ownership for healthbut likely because owners tend to be wealthier and exert more control in the workplace than employees. In short, the empirical relationships delineated by much of the social class and health research appear to demonstrate the relevance for health of various characteristics of occupations and the individuals that inhabit them, but do not obviously and clearly identify social classes (as groups) and the class relations (group dynamics) within which they operate. Bourdieus social class framework In the early 1990s, Clark and Lipset (1991) proposed that social class is steadily declining in importance in post-industrial societies, a claim that has since sparked a massive dialogue regarding the currency of social class. (See Hout et al., 2001 for an enthusiastic counter-argument.) At the very least, Clark and Lipset (1991) are perceptive when they note that workplace divisions may not be the classdening criteria in modern society that they were during the industrial era. Bourdieus approach to conceptualizing social classes has caught the imagination of many social scientists because, while it shares afnities with Marxist perspectives, it does not rely primarily upon economic labour market criteria to delineate social groups that might be social classes. In this article I draw primarily upon Bourdieus vision of social classes as presented in Distinction: A Social Critique of the Judgement of Taste (1984) and Practical Reason: On the Theory of Action (1998).

ARTICLE IN PRESS
16 G. Veenstra / Health & Place 13 (2007) 1431

Here, his social class framework is based upon his conceptualization of social space, a space of positions dened in relation to one another. All agents are located in space in such a way that the closer to one another in those two dimensions, the more they have in common; and the more remote they are from one another, the less they have in common (Bourdieu, 1998, p. 6). Within social space, groupings of similarly located agents have the potential to be social classes. Thus classes are dened relationally in his vision, and are groups sharing similar circumstances: One must therefore construct the objective class, the set of agents who are placed in homogenous conditions of existence imposing homogenous conditionings and producing homogenous systems of dispositions capable of generating similar practices; and who possess a set of common properties, objectied properties, sometimes legally guaranteed (as possession of goods and power) or properties embodied as class habitus (and, in particular, systems of classicatory schemes) (Bourdieu, 1984, p. 101). Bourdieus scheme incorporates multiple criteria for distinguishing among positions, such that every form of power or resource in social life, e.g., economic capital, educational capital and social capital, can be deemed a principle of distinction and thus a class-relevant capital. Of the many forms of capital, however, cultural capital (encompassing educational capital, social background and the cultural tastes fostered in personal and parental educational experiences) and economic capital are the fundamental principles of distinction in modern societies, according to Bourdieu, and serve to frame social space more prominently than do other forms of capital. Cultural tastes and practices are particularly important for the manifestation of social groupings in Bourdieus vision of the social space. To each class of positions there corresponds a class of habitus (or tastes) produced by the social conditioning associated with the corresponding condition and, through the mediation of the habitus and its generative capability, a systematic set of goods and properties, which are united by an afnity of style (Bourdieu, 1998, p. 8).

Bourdieu essentially argues that the social space based upon the possession of capitals serves to frame cultural tastes and practices, and that these tastes then serve to manifest social class inequalities. Within the social space, dominant groups purportedly maintain boundaries between themselves and lesser groups by delimiting the nature of tastes legitimate, middle-class and popular, or high-brow, middle-brow and low-browand then utilizing familiarity with these cultural forms to exclude others. The groups thus manifested within the social space, while not yet social classes (still requiring some form of political mobilization, perhaps, in order to become an active class), are predictive of social classes, or predisposed to become classes in the Marxist sense of the word, according to Bourdieu. Always arguing for the necessary inseparableness of theoretical and empirical exploration, Bourdieu utilized survey data from 1960s France and correspondence analysis to craft his own visual representation of the French social space (Bourdieu, 1984, pp. 128129, 1998, p. 5), grafting together a wide range of indicators from several questionnaire surveys of the public. He included variables assessing demographic characteristics, economic and cultural capital (assessed by income and educational attainment in particular), voting preferences and various occupational categories, lifestyle practices and cultural tastes. The lifestyle practices included activities such as golf, piano, chess, skiing, hiking, shing and sailing, and the cultural tastes included beverage choices such as ordinary red wine, sparkling white wine, mineral water, whisky and champagne, musical tastes (Bach, Stravinsky, modern jazz, Petula Clark) and favourite actors. The exploratory analysis produced three primary dimensions of note: the rst and most important dimension seemingly represented total volume of capital (referring to the sum of economic and cultural capital in particular), the second the relative composition of economic and cultural capital, and the third changes in these over time, or social trajectory. Bourdieu claims to have overlaid several spaces upon one another, a rst composed of the space of social conditions, a second composed of the space of life-styles, and a third (imagined) representing the theoretical space of habitus, i.e., the space of generative formula by which social positions and conditions are transformed into lifestyles practices and cultural tastes. The resultant visualization of the social space of France identied

ARTICLE IN PRESS
G. Veenstra / Health & Place 13 (2007) 1431 17

groupings of survey variable categories that, in Bourdieus opinion, represented groups of positions that were potentially predictive of social classes. I refer to Bourdieus social class scheme as culturalist because, like the work of Raymond Williams, it accords a signicance to cultural tastes, sensibilities and practices that scientic Marxist scholarship, relegating culture to the superstructure, does not. It is structuralist because, like Saussure in linguistics and Levi-Strauss in structural anthropology, it focuses primarily upon relations among elements, adopting an ontology of intersubjectivity versus one focused on atomistic individuals and never obviously prioritizing one element over another. In addition, it has a realistic quality because it is embedded in empirical observations in a particular context (France), an objectivist character because the real is revealed in relations, but also a constructivist quality given that perceptions relating to cultural tastes and dispositions are an integral part of his scheme. I believe that Bourdieus culturalstructuralist vision of social classes has the potential to make a meaningful contribution to the social class and health literature. If critiques to modern scientic Marxist perspectives are to be taken seriously, the framework may better represent the operative classes of modern-day society. It also incorporates explicit theoretical consideration of cultural and lifestyle practices embedded within social structure, some of which may inuence health directly (and are generally under-theorized in the social determinants of health literature). In addition, it provides a means of utilizing individual-level survey data to present social groupings that may be predictive of social classes. Thus, in an analytical situation in which multilevel modelling (the only valid statistical way of attributing variance in an individual-level dependent variable to a second, higher group level) is inappropriate because the number of higher order groups in class analysis is small, the identication of heath relevant social groupings via the application of multiple correspondence analysis is still possible. Surprisingly, very few health researchers have engaged in this line of exploration to date. An empirical application of Bourdieus framework My review of the health literature produced only one example of quantitative research explicitly utilizing Bourdieus analytical framework to explicate health inequalities. Anthony Gatrell and

colleagues (2004) applied Bourdieus notion of the social space to inequalities in psychological morbidity in two cities in northwest England. Specially, they sought to resurrect Bourdieus analytical method for describing a social space, noting in so doing that his quantitative work has been mostly neglected by social theorists. Although the authors were not explicitly concerned with identifying social classes, seeking instead to investigate the overlap between social space and geographical space, their analysis does address, at least indirectly, the third approach to conceptualizing social class. In order to assess the contribution of my analysis to the health literature, it is important to understand what Gatrell et al. accomplished in northwest England. Gatrell et al. chose a high- and low-income locality from each of Lancaster and Salford and conducted a questionnaire survey in these localities. The survey achieved a response rate of 44% and a sample size of 777 for the analysis described in the article. Along with basic demographics (age, gender, marital status), they assessed numerous aspects of material circumstances and social relationships: economic capital (e.g., income, car ownership, satellite television installed, home ownership), educational capital (personal educational qualications), occupational status, and social capital (sense of loneliness, desire to move, meeting with neighbours, sense of community). Upon entering these variables into a multiple correspondence analysis, the two primary dimensions produced by the analysis were interpreted by the authors to represent degree of economic capital and degree of social capital, respectively; degree of educational capital apparently adhered closely to the economic capital dimension. The authors then used logistic regression models incorporating these and other variables to identify the most persistent correlates of psychological morbidity; presence of a long-standing illness, loneliness, perceived troubles managing nancially and age were then entered into another multiple correspondence analysis. The rst dimension presented in the resultant correspondence map was indicated by nancial worries and loneliness, and was most closely allied with psychological morbidity, whereas the second dimension was indicated primarily by age. Thus economic and social capitals were deemed to be closely intertwined when it came to the social space manifesting psychological morbidity. Finally, the authors engaged in exploration of the overlap between their conception of social space and geographical space

ARTICLE IN PRESS
18 G. Veenstra / Health & Place 13 (2007) 1431

(locality or neighbourhood of residence), but found surprisingly little correspondence between the two. Gatrell and colleagues make an important contribution to the health literature by explicitly addressing Bourdieus notion of social space, i.e., by thinking relationally and then utilizing an analytical technique that supports a relational perspective, and also by exploring the interconnectedness of geographical and social space, a connection that Bourdieu discussed (Bourdieu, 1989) but not incorporate into his own empirical analyses in France. It is especially interesting to note that their multiple correspondence analysis of various capitals produced a social space with its two primary dimensions delineated by economic capital (linked with educational capital) and social capital, not total capital (economic and educational capital summed together) and composition of (economic and educational) capital as reported by Bourdieu in his analysis of the social space of France. The analysis by Gatrell et al. is not entirely consistent with Bourdieus theoretical vision of inequalities, however, because it engages in multivariate logistic regressions to identify candidates for inclusion in subsequent multiple correspondence analyses, thereby presuming strictly dened relations of causality, and because it ignores cultural tastes and practices (that importantly produce the afnities of style that contribute substantially to the nature of the social space, according to Bourdieu). As such their representations of the social space are not truly representations of the social spacebased upon the possession of capitals and cultural tastes and practices to delineate the positions and conditions of agentsenvisioned by Bourdieu in Distinction and Practical Reason. The authors note, of course, that there is no such thing as the social space, as every rendering of a social space is entirely dependent upon the selection of survey variables by the researcher. Still, the exclusion of cultural tastes and lifestyle practices means that the social space presented by Gatrell et al. may not represent a social space displaying social class boundaries. This article adopts Bourdieus culturalstructuralist approach in order to identify the health effects of social class, the latter not dened by economic ownership and control over the means of economic production but instead by the possession of various cultural tastes and dispositions, lifestyle practices, parental educational background, educational capital, economic capital, social capital and occupational type. This investigation is accomplished by

virtue of the inclusion of a wide range of variables drawn from an original questionnaire survey of randomly selected adults into multiple correspondence analyses in order to present two visual mappings of social space, one for employed survey respondents and one for all survey respondents (employed and retired, older and younger). Also utilizing respondents own self-assessed social class identications, interpretations of social groupings that may be predictive of social classes are presented. Finally, indicators of physical and mental health and well-being are situated within these social spaces of British Columbia, thereby enabling the very rst culturalstructuralist investigation of social class inequalities in health and wellbeing. I believe that the novelty of my analysis lies in its further resurrection of a strictly relational way of conceptualizing classes, its simultaneous consideration of multiple forms of capital and a wide variety of cultural tastes and practices, and its use of exploratory multiple correspondence analysis techniques to visualize the empirical manifestation of social (class) groupings in this modern-day Canadian context. Data Twenty-ve communities on the coast of British Columbia were selected in order to facilitate multilevel modelling of the health effects of social capital (see Veenstra, 2005a). A mailed questionnaire survey of randomly selected adults living in these communities was administered by the author in 2002. To obtain the random sample, a selection of households was drawn from the most current telephone listings using a systematic random sampling technique. In order to obtain a randomly selected adult household member, the letter recipient was asked to give the questionnaire to the resident of the household aged 18 and over with the rst birthday in the year. There were N 1435 respondents in total, representing an overall response rate of 56.5% (response rates by community are available in Veenstra, 2005a). Table 1 describes the distributions of responses to various demographic survey items and ve indicators of health. To assess physical health, respondents were asked if they had a recent limiting injury and whether or not they had a long-term limiting illness. In addition, body-mass index scores were calculated from respondents height and weight and then overweight and obese respondents were

ARTICLE IN PRESS
G. Veenstra / Health & Place 13 (2007) 1431 Table 1 Characteristics of survey respondents Survey item or variable What is your gender? In what year were you born? Categories Female Male Ageo25 2544 4564 6574 75 and older Married Widowed Separated Divorced Common-law Single Canada Other Apartment/condo House/townhouse Moveable dwelling Same place Moved recently None One or more First nations Other Yes No Yes No BMIo25 Moderately overweight Obese (BMI430) Depressiona Some depressive symptomsb No depressive symptoms Excellent Very good Good Fair Poor Distribution 677 747 27 450 626 182 136 883 101 45 122 138 136 1102 332 164 1168 90 1322 96 905 495 65 1370 304 1110 343 1069 629 537 235 109 329 997 217 509 491 163 24 (47.5) (52.5) (1.9) (31.7) (44.1) (12.8) (9.6) (62.0) (7.1) (3.2) (8.6) (9.7) (9.5) (76.8) (23.2) (11.5) (82.1) (6.3) (93.2) (6.8) (64.6) (35.4) (4.5) (95.5) (21.5) (78.5) (24.3) (75.7) (44.9) (38.3) (16.8) (7.6) (22.9) (69.5) (15.5) (36.3) (35.0) (11.6) (1.7) 19

What is your current status?

In what country were you born? In what type of dwelling do you live?

Where did you live one year ago? Calculated variable: number of children at home To which ethnic or cultural group(s) did your ancestors belong? In the past 12 months, did you have any injuries that were serious enough to limit your normal activities? Do you have any long-term illness, health problem or handicap which limits daily activities or the work that can be done? Calculated variable: body-mass index

Calculated variable: depressive symptoms

In general, how would you rate your health?

Three or more depressive symptoms. One or two depressive symptoms.

distinguished from the others. To assess emotional well-being, eleven questions1 assessed recent depressive symptoms; respondents who reported extensive experience (most of the time or all of the time) with three or more of these symptoms in the past week were distinguished from those reporting only
During the past 7 days, how often did you feel: downhearted and blue? really happy? quite nervous? calm and peaceful? worthless? restless and dgety? hopeless? depressed? that everything you did was an effort? that people disliked you? that your sleep was restless? /none of the time, almost never, some of the time, most of the time, all of the timeS
1

one or two symptoms and those reporting none. Lastly, respondents were presented with a standard self-rated health question to assess overall health and well-being. Table 2 describes the variables used to categorize occupations and assess various capitals and cultural tastes and practices. Three survey questions pertaining to the respondents main job were used to classify occupational types: What is your occupational title? What is the name of the business or company you work for? How would you describe your activities or duties? Occupations were coded

ARTICLE IN PRESS
20 G. Veenstra / Health & Place 13 (2007) 1431 Table 2 Occupational categories and measures of capital Survey item or variable Pineo occupational classication Categories Self-employed professional Professional High level management Semi-professional Technicians Middle managers Supervisors Foremen/women Skilled clerk/sales/service Skilled craft/trade Farmers Semi-skilled clerk/sales/service Semi-skilled manual Unskilled clerk/sales/service Unskilled manual Farm labourers Less than $20,000 Between $20,000 and $29,999 Between $30,000 and $39,999 Between $40,000 and $59,999 Between $60,000 and $79,999 More than $80,000 Owns home fully Owns home partially Rents home Less than high school High school Some postsecondary Community college University Less than high school High school Some postsecondary Community college University Smokes daily Smokes occasionally Non-smoker Never drinks copiously Drinks copiously sometimesa Heavy drinkerb No religion Protestant Roman Catholic Other religion Trusts communityc Mistrusts communityd Trusts politiciansc Mistrusts politiciansd Sense of belongingc Little belongingd Meets neighbours often Meets neighbours sometimes Never meets neighbours Yes No Yes No Distribution 13 76 33 85 25 51 73 24 52 113 7 71 68 45 67 12 131 111 190 318 234 368 584 624 209 164 269 380 327 273 423 343 164 206 195 234 70 1117 935 113 268 481 589 205 125 975 430 345 1056 1106 304 466 616 301 669 755 1207 213 (1.6) (9.3) (4.0) (10.4) (3.1) (6.3) (9.0) (2.9) (6.4) (13.9) (0.9) (8.7) (8.3) (5.5) (8.2) (1.5) (9.1) (8.2) (14.1) (23.5) (17.3) (27.2) (41.2) (44.0) (14.7) (11.6) (19.0) (26.9) (23.1) (19.3) (31.8) (25.8) (12.3) (15.5) (14.7) (16.5) (4.9) (78.6) (71.0) (8.6) (20.4) (34.4) (42.1) (14.6) (8.9) (69.4) (30.6) (24.6) (75.4) (78.4) (21.6) (33.7) (44.5) (21.8) (47.0) (53.0) (85.0) (15.0)

What is your best estimate of the total income from all household members from all sources during the last 12 months (including income from savings, investments, pensions, rent, and unemployment insurance as well as wages)?

Calculated variable: owns or rents home

Calculated variable: highest personal educational attainment

Calculated variable: highest parental educational attainment

At the present, how often do you smoke cigarettes?

How many times in the last month have you had more than ve drinks on one occasion? What is your religion, if any?

Most people in my community can be trusted. Politicians are generally good, capable people. I feel that I belong in my community. How often do you meet socially with neighbours?

Have you volunteered regularly in the last year? Did you vote in the last federal election on November 27, 2000?

ARTICLE IN PRESS
G. Veenstra / Health & Place 13 (2007) 1431 Table 2 (continued ) Survey item or variable Calculated variable: number of voluntary associations Categories More than one association One association No associational involvement Always feels alone Sometimes feels alone Seldom feels alone Lower class Working class Middle class Upper class Distribution 598 735 102 90 431 898 29 501 822 11 (41.7) (51.2) (7.1) (6.3) (30.4) (63.3) (2.1) (36.8) (60.3) (0.8) 21

How often do you feel alone?

There is quite a bit of talk these days about different social classes. People generally feel they belong to one of the lower, working, middle or upper classes. Which of the following classes best describes you?
a

One time in the month. Two or more times in the month. c Agree strongly and agree responses collapsed together. d Neutral, disagree and disagree strongly responses collapsed together.
b

using the Statistics Canada Standard Occupational Classication and Pineo occupational categories were assigned to employed respondents. To assess economic capital, respondents reported their household income and whether they owned their home. To assess educational capital and social background, respondents reported their own and their mothers and fathers highest levels of educational attainment; personal attainment and highest credential attained by a parent are recorded in Table 2. To assess cultural knowledge, a series of 71 cultural knowledge items were presented to respondents. The clustering of these knowledge items in a principle components analysis are described elsewhere (Veenstra, 2005b). The eleven items that loaded highest on the eleven resultant factors respectively were utilized in this analysis, and included familiarity with various sports gures, artists, novelists and books, and magazines.2 I also included in my analysis three items that did not load highly on any of the eleven factorsfamiliarity with Architectural Digest, Scientic American and sports gure Don Bradman. Some of the cultural knowledge items were taken from Erickson (1991, 1996), who studied relationships between cultural tastes and social class elsewhere in Canada, whereas others were simply concocted by the author. As
Please indicate whether you have read or heard of the following books: Oliver Twist, Kamouraska, The Wealthy Barber. Please indicate which of the following magazines you have read in the past six months: Time, The New Yorker, Chatelaine, Penthouse. Please indicate which of the following artists you recognize: A. J. Casson, Andy Warhol. Please indicate which of the following athletes you recognize: Brett Favre, Michael Jordan.
2

representatives of the factors derived by the principle components analysis, these cultural knowledge items best represent the entire body of 71 knowledge items, but are not assumed to necessarily represent the high-brow and low-brow tastes that maintain the boundaries of social classes. (In follow-up work, qualitative methods of inquiry are needed to identify better cultural knowledge measures for social class comparisons.) To assess cultural practices, a list of 22 activities were presented to respondents; those with a suitably large N (450) were used in the analysis.3 Respondents were also asked about smoking habits, alcohol consumption and religious afliation. Various sentiments and activities that may represent forms of social capital were assessed for survey respondents. One survey item measured trust in community members, another assessed perceptions regarding the trustworthiness of politicians, a third measured sense of belonging in the community and a fourth assessed loneliness. Other items measured degree of networking with neighbours and civic action in the form of volunteerism and of voting. Lastly, responses to a survey item pertaining to associational involvement4 enabled
Have you done any of the following in the last month? running/jogging; walking long distances; hiking; bicycling; kayaking; gardening, yard work; bowling; swimming; baseball; popular or social dance; weight-training; home exercises; yoga/ tai-chi; golng; exercise class/aerobics 4 Which of the following types of organizations do you currently participate in? political parties; trade unions; environmental groups; parent-teacher associations, school service associations; tenants groups; neighbourhood watch or neighbourhood improvement associations; youth groups; church or other
3

ARTICLE IN PRESS
22 G. Veenstra / Health & Place 13 (2007) 1431

the calculation of a variable assessing breadth of participation within different types of voluntary associations. Finally, respondents self-assessments of their own social class standing were collected in order to investigate the overlap between perceptions of class standing (an indicator of class consciousness?) and the social groupings identied by the multiple correspondence analyses. Methods In the spirit of relational thinking, this article does not explicitly identify dependent variables, instead utilizing multiple correspondence analysis.5 Correspondence analysis (CA) is a technique for analysing the association between two categorical variables by representing the categories of the variables as points in a low-dimensional space. It does not specify a dependent variable, and it is primarily used for exploratory rather than conrmatory purposes. In essence, CA produces a visual representation of the relationship between two categorical variables that is more easily interpreted than a contingency table, especially when the table has many cells and the correspondence of variable categories cannot be readily discerned. In the visual representation, categories with similar distributions will be represented as points that are close in space, and categories that have very dissimilar distributions will be positioned far apart (Clausen, 1998, p. 2). While the number of dimensions can be large, generally only the two dimensions that explain the most variance are displayed visually in a correspondence map. The technique has very few data assumptions: the data elements must be non-negative numbers, an
(footnote continued) religious groups; charitable or volunteer associations, service clubs; art or music groups; evening classes; social clubs (e.g., Rotary Club, Womens Institute, Lions Club); sports teams, gyms, exercise classes; professional associations; literary societies; seniors clubs; other groups or organizations. 5 Three other articles investigating social determinants of health in a traditional causal manner also utilize this data set. Veenstra (2005a) used multilevel modelling to discern the health effects of community of residence, economic capital, educational capital and social capital. Veenstra (2005c) utilized regression modelling to describe zero-order and multivariate relationships between perceptions of relative social status, objective status and health, and Veenstra (2006) used regression modelling to assess relationships between neo-Marxist conceptualizations of class position, socio-economic status and health.

association should exist to be described, generally assessed by use of the Chi-square statistic, and the contingency table ought to be relatively large (Clausen, 1998). Although the technique has a long history in French scholarship, it was seldom utilized in Anglo-American scholarship prior to the publication of the English translation of Bourdieus Distinction. Multiple correspondence analysis (MCA) is the extension of CA to the analytical situation with three or more variables. Like CA (and unlike log linear models which analyze interrelationships between a set of variables), MCA examines the relations between the categories of variables (Clausen, 1998). It produces dimensions that individually explain some of the variance in the Nway table, and also produces a correspondence map that visually depicts the two dimensions that explain the most variance. As with CA, the loading of points on dimensions can be calculated, enabling the researcher to interpret the dimensions. Like CA, variable categories can be active (used to compute the dimensions that produce the correspondence map) or treated as supplementary (not used to compute the dimensions, but then overlaid upon the existent map). In addition, entire variables can be designated supplementary. In this article, two social space mappings are enacted utilizing the multiple correspondence analysis routine in SPSS 13.0. The rst mapping includes the Pineo occupational classication as an active variable and so pertains to the social space of employed people only. The second excludes the Pineo measure and as such pertains to the social space of all respondents, adults aged 18 and over. For each mapping, the variables described in Tables 1 and 2 are entered into the multiple correspondence analysis as active variables that contribute toward shaping the social space, excepting age, gender, self-assessed social class and the health measures which are entered as supplementary variables instead. The two dimensions that explain the most variability overall are depicted visually in a correspondence map, and interpretations or speculations regarding category groupings that might be predictive of social classes are presented. The spatial positioning of the health indicators in this social space is then discussed, enabling further speculation pertaining to the relationship between social (class) groupings and health inequalities in British Columbia, Canada. Substitution with the mode was used to accommodate missing values. In total, the data set of 1435 respondents contained 59 variables of interest for

ARTICLE IN PRESS
G. Veenstra / Health & Place 13 (2007) 1431 23

this analysis, one of whichthe Pineo occupational classicationhad a particularly small N (815). Substitution with the mode is problematic when there are many missing cases, and so a process for identifying and then deleting particularly problematic cases and variables before imputing the mode was devised. First, the number of variables with a missing value was identied for each case. The greatest number of variables for any one case was quite low (only 10) and so no cases were deleted for this reason. Second, alcohol consumption (8.3%), highest parental educational achievement (7.2%), household income (5.8%) and self-assessed social class (5.0%) were the variables with the highest proportions of non-responses, excepting the Pineo classication63 cases had missing values for two or more of these four variables and so were deleted. Among the remaining cases, only alcohol consumption possessed a sizeable proportion of missing values (6.8%)of the 93 non-responses, 28 also failed to respond to one of the other variables in the data set, excluding the Pineo occupational classication, and were also deleted. In total, 91 especially problematic cases were deleted, leaving a data set containing relatively few missing values prior to utilizing multiple correspondence analysis. The nal sample for employed persons was 763 and the nal sample for all respondents was 1344. Results The two-dimensional visual mappings of the social spaces of employed persons and all survey respondents are shown in Figs. 1 and 2, respectively. The reader should seek to identify variable categories that cluster together on paper, but should not give too much credence to the spatial distance between any two given variable categoriesthese depictions of social space can only display general patterns. A total of 763 cases were utilized in the analysis that produced Fig. 1, 595 of which did not have any missing values. The variable principle normalization option was utilized in this multiple correspondence analysis. dimension 1 (D1) explained 6.27% and dimension 2 (D2) explained 5.35% of the total variance. D1, displayed vertically, was inuenced primarily by educational attainment and occupational classication (with discrimination measures of .324 and .320, respectively), secondarily by associational involvement (.256) and in smaller part by household

income (.142). The dispersal of educational attainment along D1 is apparently linear, with less than high school to the bottom and university degree to the top (the latter category most clearly distinguished from the others). The occupations also present an identiable pattern along D1. The lowermost occupations are the manual jobs (unskilled, semi-skilled and farm labourers), followed by unskilled clerks and salespeople, then technicians, skilled trades people and semi-skilled clerks and salespeople, next farmers and foremen/women, then skilled clerks and salespeople, supervisors and middle managers, next high-level managers and semi-professionals, and lastly professionals and selfemployed professionals at the top of Fig. 1. Household income is also seemingly linearly dispersed along D1, with the lowest incomes to the bottom. Associational involvement is not linearly dispersed on D1, however (although it is on D2), making its contribution to the mapping somewhat ambiguous. Several of the outermost categories possessed small Ns, i.e., The New Yorker (43), Architectural Digest (40), o$20,000 (32), Kamouraska (23), o25 yrs (15) and self-employed professional (13). These variable categories did not play an integral role in dening the resultant dimensions and so should not play an important role in interpreting them either. I interpret the rst dimension to represent the key social class dimension of this social space, constructed primarily on the basis of workplace distinctions and educational attainment and secondarily on the basis of economic capital. This interpretation corresponds very closely with Bourdieus interpretation of ndings from 1960s France, wherein total capital (educational and economic) purportedly served to structure social space more than any other principle, and where occupational distinctions (class fragments in Bourdieus lexicon) displayed interpretable patterns along these lines. D2, displayed horizontally in Fig. 1, was inuenced strongly by marital status (.395), home ownership (.347) and type of dwelling (.344) and was moderately inuenced by income (.141) and the presence of children at home (.134). On the left of the mapping are separated, divorced and (especially) single respondents, with married respondents on the right. Also on the left are (transient) renters living in apartments and condos. Household income is also apparently linearly arrayed along D2, with poorer respondents on the left of the mapping and wealthier ones on the right. Lastly, the presence of children at home are on the right of D2.

24

total capital (educational and economic) +

The New Yorker self-employed professional

professional aerobics university Kamouraska Scientific American kayaking A.J. Casson semi-professional high level management Wealthy Barber runner/jogger > one club swimming weight-training SRH excellent home exercises bowling hikes

Architectural Digest yoga/tai-chi

parent university

> 75 yrs

life course and family structure


separated

ARTICLE IN PRESS

First Nations

> $80,000 Oliver Twist Time magazine volunteer Don Bradman trusts politicans other religion Chatelaine middle class $60-79,999 kids at home golfing parent CC owns home 65-74 Andy Warhol married townhouse/house parent some PS Brett Favre Michael Jordan Protestant Penthouse SRH very good middle managers has injury partial home-owner supervisors 45-64 25-44 Roman Catholic foremen/women sometimes feels alone skilled clerk/sales/service

G. Veenstra / Health & Place 13 (2007) 1431

divorced

single

renter

semi-skilled clerk/sales/service obese has illness no kids at home SRH good no religion moved recently community college mistrusts community some PS parent < HS parent HS skilled craft/trade never meets neighbours smokes occasionally did not vote non-volunteer working class technicians common-law little belonging heavy drinker $30-39,999 high school $40-59,999 SRH fair always feels alone $20-29,999

depression
one club farm labourers

unskilled clerk/sales/service

SRH poor < 25 yrs


moveable dwelling smokes daily unskilled manual semi-skilled manual < high school

apartment/condo

< $20,000

total capital (educational and economic) -

Fig. 1. Social groupings in the social space of employed persons.

ARTICLE IN PRESS
G. Veenstra / Health & Place 13 (2007) 1431 25
total capital (educational and economic) +

kayaking runner/jogger The New Yorker baseball weight-training Kamouraska university yoga/tai-chi hikes Penthouse no clubs parent university swimming bicycling aerobics

SRH excellent
Scientific American Architectural Digest The Wealthy Barber dancing golfing > $80,000 A.J. Casson

25-44
kids at home Brett Favre home exercises parent CC long walks partial home-owner single moved recently

< 25 yrs

social relationships -

volunteers occasional drinker parent some PS > one club $60-79,999 no religion separated Time magazine SRH very good heavy drinker other religion Oliver Twist Andy Warhol First Nations has injury middle class trusts politicans smokes occasionally Michael Jordan married townhouse/house common-law social community college did not vote parent HS mistrusts community never meets neighbours some post-secondary $40-59,999 little belonging working class

relationships +

45-64
Don Bradman

obese SRH good


high school Protestant

renter always feels alone

depression
one club

divorced

owns home $30-39,999

smokes daily apartment/condo moveable dwelling

has illness
parent < HS

lower class

65-74
$20-29,999

SRH poor

SRH fair

< $20,000 < high school

> 75 yrs

widowed

total capital (educational and economic) -

Fig. 2. Social groupings in the social space of all respondents.

ARTICLE IN PRESS
26 G. Veenstra / Health & Place 13 (2007) 1431

This depiction of D2 corresponds somewhat closely with the second dimension reported by Bourdieu, who claimed that relative possession of educational and economic capital dened the second most important dimension of the French social space. While I report a similar phenomenon note the movement of increasing incomes from the lower left to the upper right and increasing education credentials from the lower right to the upper leftother factors apparently represent stronger inuences upon D2. Specically, I interpret D2 to represent life-course and family structure more than anything else: married respondents aged 45 and older living in non-rental housing are contrasted with (generally younger) non-married respondents living in temporary accommodations. (I would argue that Bourdieus second dimension may also reect the nature of life-course and family structuremiddle-aged married respondents with home equity are undoubtedly likely to have a sizeable proportion of their total capital provided by economic capital.) Next, I attempt to identify groupings of variable categories that may actually represent social groupings of people that are predictive of social classes. It is important to consider both dimensions simultaneously when looking for groupings, although I have prioritized D1 over D2 in this interpretive process. It is also important to note that the middle of the mapping represents characteristics of survey respondents not well explained by D1 and D2 (but possibly explained by other dimensions), and so variable categories located near the intersection of the axes may not play an important role in dening social groupings in this social space. (Although Figs. 1 and 2 do not contain all of the variable categories that were used to construct these social spaces, all of the categories not presented in the gures belong to the middle portions of the mappings.) Professional class: One distinctive grouping pertains to the collection of variable categories to the top of D1 and slightly to the left of the vertical axis. This group is apparently well-educated (as are their parents) and wealthy, although educational credentials are the more important structuring principle of the two, and includes professionals and high-level managers. Many distinctive cultural practices pertaining to physical activity accumulate in this grouping, e.g., aerobics, kayaking, yoga/tai-chi, running and weight training. Distinctive cultural tastes in this group are familiarity with magazines

such as The New Yorker, Architectural Digest and Scientic American, books such as Kamouraska and The Wealthy Barber and artists such as A.J. Casson.6 Members of this group are also likely to participate in a range of clubs and associations. I refer to this group as the professional class. In the spirit of Bourdieus framework, this group can lay claim to a variety of distinctive cultural tastes and practices that may be used by group members to maintain class boundaries between themselves and others. Middle class: Moving down the vertical axis, the second grouping encompasses the dense central portion of the correspondence mapping. This group includes skilled clerks and salespeople, supervisors and middle managers, and is moderately well educated (many members possessing a community college or university credential). Their household incomes are quite high. With respect to cultural tastes, they are often familiar with Andy Warhol, Brett Favre7 and Michael Jordan, may have glanced at Penthouse magazine recently, like to engage in an occasional bout of drinking but generally choose not to smoke, trust other members of the community and politicians, voted in the last federal election and like to spend time socializing with neighbours. Members of this group are also quite likely to refer to themselves as middle class. Working class: A third grouping pertains to the collection of variable categories on the bottom end of D1. Members of this group generally have few educational credentials, tend to work in unskilled
In a principle components analysis of cultural tastes (Veenstra, 2005b), the factor represented well by The New Yorker was also well represented by Harpers and Saturday Night magazines. The Wealthy Barber is a popular American book devoted to nancial strategies and represented a factor that also included The Economist and Financial Times publications. Kamouraska is a classical piece of French literature from the Canadian province of Quebec, and represented a factor containing a series of literary works such as Sophies World and Whats Bred in the Bone. A.J. Casson, sometime member of the Group of Seven, was a Canadian landscape painter from the early half of the twentieth century. Familiarity with A.J. Casson represented a factor comprised of artists such as Harold Town, Cornelius Kriefhoff, Mary Pratt and Henry Moore. 7 The factor well represented by Andy Warhol contained other popular artists such as Salvadore Dali, Pablo Picasso and Robert Bateman. Brett Favre is a professional football player in the US and represented a factor mostly composed of sports gures such as Joe Thornton (a hockey player in the NHL), Shannon Stewart (baseball player and former member of the Toronto Blue Jays) and Danny McManus (football player in the Canadian Football League).
6

ARTICLE IN PRESS
G. Veenstra / Health & Place 13 (2007) 1431 27

and semi-skilled manual occupations and are far from wealthy. Smoking and drinking heavily are the only cultural practices contained in this data set that accrue to this group. Group members often live in moveable dwellings, relationships with neighbours are often non-existent and participation in the civil space via volunteering is negligible. I refer to this group as the working class, noting that members of this grouping also tend to refer to themselves in this way. Class of solitude: The nal potential grouping refers to the lower left quadrant of the social space mapping. This is a poor group with members who tend to rent accommodation in temporary living situations and are likely to be young and single, separated or divorced. This grouping is not obviously associated with any specic occupational categories. I refer to this group as the class of solitude, noting the degree of loneliness and lack of belonging to the community evident among the members of this sector, but am least condent that this section of social space represents a real social grouping. And are these exploratory conceptions of social (class) groupings associated with health? Regarding the placement of my health indicators within this social space of employed people, the excellent selfrated health category appears to be located near the professional class, fair and poor self-rated health fall squarely within the working class portion of the space and the indicator of depressive symptoms is close to both the class of solitude and the working class. The three assessments of physical wellbeingpresence of injuries, long-term limiting illness and obesityfall squarely within the middle portion of the mapping, however, indicating that they belong to the middle class (or, more accurately, that they do not obviously belong to any of the other classes). In general, it seems that mental health and psychological interpretations of overall well-being are better explained by these social (class) groupings than are any of the measures of physical well-being. The two-dimensional visual mapping of the social space of all survey respondents, again using variable principle normalization, is shown in Fig. 2. A total of 1344 cases were utilized in the analysis, 983 of which did not have any missing values. dimension 1 (D1) explained 6.62% of the total variance and dimension 2 (D2) explained 5.43%. D1, arrayed vertically, was explained primarily by educational attainment (discrimination measure .267),

income (.235), age (.222) and parental education (.194). As in Fig. 1, education, parental education and income are linearly arrayed along D1. The extreme education categories are the most sharply distinguished from the other categories. Unlike Fig. 1, age clearly plays an important role in the primary dimension, with older respondents falling on the lesseducated and poorer end of D1. Thus, as might be expected by the inclusion of a large number (n 407) of retired people in the sample, issues pertaining to life-course and age cohort play a bigger role in the distribution of capitals, cultural tastes and lifestyle practices in social space. I interpret this dimension to reect primarily the degree of overall possession of the most important capitals (educational and economic), just as Bourdieu reported for French society, but also note the manifestation of life-course along this dimension. D2 was explained primarily by home ownership (discrimination measure .367) and marital status (.307), secondarily by age (.208) and dwelling type (.202) and in smaller part by indicators of social capital (community trust .145, associational involvement .142, sense of belonging to the community .124). I interpret this dimension to represent the presence and quality of social relationshipsfamily structure, life-course and social capitalwith a slightly larger role played by social capital in D2 of Fig. 2 than in D2 of Fig. 1. A version of the highest class similar to that displayed in Fig. 1 manifests itself in Fig. 2, a grouping again focused specially on various aspects of a physical lifestyle but with golng and dancing also included in the realm of the professional class. Familiarity with Penthouse magazine, a middle-class cultural taste in Fig. 1, also falls within sight of this class grouping. The middle class of Fig. 2 closely resembles the middle class of Fig. 1. The working class of Fig. 2, while quite similar to its counterpart in Fig. 1, has a pronounced age component that is not as clearly evident in Fig. 1, with elderly respondents in particular more likely than not to be included in this class. Lastly, the class of solitude manifests itself in the middle left portion of this mapping, but in this instance includes heavy drinking, a cultural practice seemingly attributed to the working class in Fig. 1. With respect to the indicators of health, Fig. 2 shows patterns similar to those manifested in Fig. 1: excellent self-rated health is located near the professional class, fair/poor health falls within the working class, depression falls within the class of

ARTICLE IN PRESS
28 G. Veenstra / Health & Place 13 (2007) 1431

solitude, and both obesity and the presence of injuries are located within the middle portion of the map. In contrast with Fig. 1, however, the presence of long-term limiting illnesses falls within the realm of the working class, again perhaps a function of age. (Thus only in Fig. 2 is a measure of physical health located outside of the middle portion of social space.) Discussion Clearly the analysis has important limitations. First, the statistical analysis is entirely explorative and so does not take advantage of the randomness inherent to the survey samplethere are no tests of statistical signicance facilitating extrapolations to the entire population of residents of these communities. Conrmatory testing of the exploratory ndings presented in this article is certainly warranted for the future. Second, in contrast with Gatrell et al. (2004), geographical space, indicated in this instance by community of residence, was not explicitly incorporated into my analysis. Gatrell et al. found little overlap between social space and geo-spatial locale in the Lancaster area anyway, and my own multilevel investigation into communitylevel determinants of health utilizing this same data set reported little variability in health attributable to community of residence (Veenstra, 2005a). Also, Bourdieu discussed the nature of overlap between the two kinds of space (Bourdieu, 1989), noting that people who are far from one another in social space can still meet and interact, however briey, in physical space. Regardless, further investigation of the overlap between conceptions of geo-political and social space in this and other contexts is probably warranted as well. Third and most importantly, Bourdieu recommends the mobilization of numerous methods of observation and measurement when engaging in investigations of social space and social class. Ethnographic exploration of the connections between social class and culture in British Columbia that identies the perceptions that produce social distancing in everyday life would certainly supplement the analysis presented here, as would a multi-method inquiry into the cultural tastes and practices that most readily differentiate social (class) groupings. Clearly my analysis can only represent the rst step in a wide-ranging, longitudinal and multi-method research program devoted to uncovering and understanding the nature of health inequalities embedded

in social class dynamics in British Columbia, Canada. The degree of consonance between the primary dimensions generated in this modern western Canadian context and those produced by Bourdieu for French society of the 1960s is remarkable. In both analyses, possession of educational capital and economic capital taken together represented the foremost dimension of social space; in both analyses, the relative possession of these two key resources represented a second important principle of distinction. The linear nature of the distribution of occupational categories along the rst dimension in my conception of social space is also a remarkable nding, but loses some lustre upon noting that the Pineo occupational classication utilized to distinguish occupational types from one another in this analysis is in fact a measure of occupational prestige based primarily upon the salaries generated by and the educational credentials required for occupations. Thus educational capital and economic capital possibly represent the key framing principles for the social space of these British Columbian communities, just as they did for 1960s France (but perhaps not for modern-day Lancaster, England). However, educational capital was relatively more important than economic capital in my depiction of social space, and social relationships pertaining to family structure, life-course and social capitalalso played a crucial role in shaping this social space. The incorporation of social capital in this work and in the social spaces produced by Gatrell et al. (2004) represents an important renement of this discourse, validating the notion that many forms of capitalany social, cultural, economic or political resourcecan distinguish social groupings that may be predictive of social classes and inuence health and well-being via class dynamics. Because the analysis is purely exploratory and the variables are indicators of individual-level phenomena, I cannot be condent that collections of variable categories represent actual social groupings in British Columbia, let alone social classes with a sense of shared identity. Still, it is interesting to note that self-assessed social class adhered to my interpretations of groupings to some degree, suggestive of a class consciousness along working class and middle class lines. It is also important to note that distinctive cultural tastes and practices manifested themselves among the higher classes in this social space far more readily than they did among the

ARTICLE IN PRESS
G. Veenstra / Health & Place 13 (2007) 1431 29

lower ones (excepting heavy drinking and smoking, tastes that fell within the class groupings near the bottom of the social space mappings). This insight resonates with Bourdieus contention that some cultural tastes and practices come to be dened as exclusively high-brow and are then used by elites to maintain boundaries between themselves and lesser classes. The tastes and practices with the potential to play this role in class dynamics seemingly describe the famous west coast lifestyle. This lifestyle, manifested in its purest form in the urban cities of Vancouver and Victoria and based in large part upon outdoor activities, includes such activities as running for exercise, commuting to work by bicycle (rather than in an environmentpolluting car), kayaking in the nearby waterways, hiking in the surrounding mountain ranges, skiing in the mountains at Cypress and Whistler, taking yoga classes, visiting spas, hitting the links and playing organized sports like baseball or soccer. According to my depictions of social space, highbrow magazines such as The New Yorker and Architectural Digest may also belong to the realm of west-coast elite cultural tastes. By contrast, the only cultural tastes assessed in this data set that inhered to the working class were smoking and excessive alcohol consumption. (I expected knowledge of sports gures such as Michael Jordan and Brett Favre, participation in bowling as an extracurricular activity and familiarity with Chatelaine and Penthouse magazines to represent distinctively working class tastes. These expectations were not manifested in my depictions of social space.) Qualitative exploration of the role of such sophisticated and low-brow tastes and practices in everyday social interactions throughout these Canadian communities would help to elucidate the exclusionary role they play among classes. Poor relationships with neighbours and the loneliness that comes from a paucity of close and meaningful social relationships located themselves within the class groupings near the bottom of social space. In and of itself, this nding illustrates the profound effects of class on the lives of these Canadians. The psychological and interpretive measures of health also adhered closely to class lines in this social space, supporting Mildred Blaxters (1989) contention that perceptions of well-being reect social class experiences more than objective measures of physical heath reect class inequalities (although illnesses also became pertinent when non-employed respondents were included

in the analysisperhaps a function of age as much as class). In my depictions of social space, poor selfrated health and depressive symptoms were located in the lower portions of social space near or between the working and solitude classes and excellent selfrated health was located near or between the elite professional and middle classes. The location of these health indicators suggests that educational attainment and income or wealth are closely associated with self-rated health, certainly not a new insight. (For instance, Mirowsky and Ross, 2003 argue that the relationship between educational attainment and health is primarily based upon an individuals ability to problem-solve in difcult circumstances, and represents the basis of the class-health relationship). The location of excellent self-rated health close to a portion of social space containing many physical activities (such as running, kayaking, aerobics, etc.) suggests that exercise also inuences self-rated health directlyagain, nothing new here. The novelty of my analysis, however, comes from the speculations that arise from adopting a structuralist vision of social space and the activities, tastes and perceptions that inhere within it, and from the practical utilization of multiple correspondence techniques to assess relations among elements. Like Bourdieu, in this analysis I am fundamentally concerned with the structure versus agency and objectivism versus subjectivism dichotomies. With respect to these dichotomies, Bourdieu summarizes the gist of his analysis in Distinction in this way: On the one hand, the objective structures that the sociologist constructs, in the objectivist moment, by setting aside the subjective representations of the agents, form the basis for these representations and constitute the structural constraints that bear upon interactions; but on the other hand, these representations must also be taken into consideration particularly if one wants to account for the daily struggles, individual and collective, which purport to transform or to preserve these structures. This means that the two moments, the objectivist and the subjectivist, stand in a dialectical relationship (Bourdieu, 1989, p. 15). In essence, Bourdieu seeks here to re-address the distinction between life chances and life choices described by Weber. A relational interpretation of social space encompasses or even transcends these

ARTICLE IN PRESS
30 G. Veenstra / Health & Place 13 (2007) 1431 Bartley, M., Carpenter, L., Dunnell, K., Fitzpatrick, R., 1996. Measuring inequalities in health: an analysis of mortality patterns using two social classications. Sociology of Health and Illness 18 (4), 455475. Blaxter, M., 1989. A comparison of measures of inequality in morbidity. In: Fox, J.J. (Ed.), Health Inequalities in European Countries. Gower Publishing, Aldershot. Boffetta, P., Kogevinas, M., Westerholm, P., Saracci, R., 1997. Exposure to Occupational Carcinogens and Social Class Differences in Cancer Occurrence. IARC Scientic Publications 138, 331341. Borooah, V.K., 1999. Occupational class and the probability of long-term limiting illness. Social Science and Medicine 49, 253266. Bourdieu, P., 1984. Distinction. A Social Critique of the Judgement of Taste. Harvard University Press, Cambridge (originally published in French in 1979, tr. by Richard Nice). Bourdieu, P., 1989. Social space and symbolic power. Sociological Theory 7 (1), 1425. Bourdieu, P., 1998. Practical Reason. On the Theory of Action. Stanford University Press, Stanford (originally published in French in 1994). Chandola, T., Jenkinson, C., 2000. The new UK National Statistics Socio-Economic Classication (NS-SEC); investigating social class differences in self-reported health status. Journal of Public Health Medicine 22 (2), 182190. Clark, T.N., Lipset, S.M., 1991. Are social classes dying? Reprinted (2001). In: Clark, T.N., Lipset, S.M. (Eds.), The Breakdown of Class Politics: A Debate on Post-Industrial Stratication. The Johns Hopkins University Press, Baltimore, pp. 3954. Clausen, S.-E., 1998. Applied Correspondence Analysis: An Introduction. Sage Publication, Inc., Thousand Oaks, California. Coburn, D., 2000. Income inequality, social cohesion and the health status of populations: the role of neo-liberalism. Social Science and Medicine 51, 135146. Erickson, B., 1991. What is good taste good for? Canadian Review of Sociology and Anthropology 28 (2), 255278. Erickson, B., 1996. Culture, class, and connections. American Journal of Sociology 102 (1), 217251. Feinstein, J., 1993. The relationship between socioeconomic status and health: a review of the literature. Milbank Quarterly 71, 279322. Fitzpatrick, J., Dollamore, G., 1999. Examining adult mortality rates using the National Statistics Socio-Economic Classication. Health Statistics Quarterly 2, 3340. Frohlich, K.L., Corin, E., Potvin, L., 2001. A theoretical proposal for the relationship between context and disease. Sociology of Health and Illness 23, 776797. Gatrell, A.C., Popay, J., Thomas, C., 2004. Mapping the determinants of health inequalities in social space: Can Bourdieu help us? Health and Place 10, 245257. Grabb, E.G., 1997. Theories of social inequality: classical and contemporary perspectives, 3rd ed. Harcourt Brace and Co., Toronto. Hattersley, L., 1997. Expectation of life by social class. In: Drever, F., Whitehead, M. (Eds.), Health Inequalities. The Stationary Ofce, London. Hout, M., Brooks, C., Manza, J., 2001. The persistence of classes in post-industrialist societies. In: Clark, T.N., Lipset, S.M. (Eds.), The Breakdown of Class Politics: A debate on

dichotomies by identifying the real with relations rather than with substances. Applying a relational interpretation to health inequalities in social space means that a spatial association between commuting to work by bicycle and excellent self-rated health, for example, need not be necessarily a (direct) causal relationship. Rather, the accoutrement of cycling (the helmet, the backpack, the rain gear) and the terminology of architectural discourse (modern and country, apse and balustrade) may induce social distances in everyday interactions by providing subtle messages to members of other classes to keep their distance and stay in their place. Cultural tastes and practices such as these form the routinized logic of the daily life of social space contexts, and social distancing of this kind, undoubtedly consonant with the possession of material resources (via personal wealth) and highly prestigious occupations, may serve to isolate communities of people from others, all of whom know where they belong, and then produce health inequalities via other means by fact of the segregation. In short, a graphical depiction of social space provides pointers towards social groupings (that may be social classes) where group members share an afnity of style and reservoir of resources, and so is suggestive of the shape and character of the group-level relations that accord health-enhancing privilege and opportunities to sub-sections of society. Acknowledgements This research project was funded under the auspices of a New Investigator Scholar Award (20002005) from the Canadian Institutes of Health Research. Ana Sandrin graphically designed and administered the survey questionnaire, Shona Kelly classied the occupations and Xuekui Zhang provided statistical consulting with regard to multiple correspondence analysis and treatment of missing values.

References
Adler, N.E., Ostrove, J.M., 1999. Socioeconomic status and health: what we know and what we dont. Annals of the New York Academy of Sciences 896, 315. Adler, N., Boyce, T., Chesney, M., Cohen, S., Folkman, S., Kahn, R., Syme, S., 1994. Socioeconomic status and health: the challenge of the gradient. American Psychologist 49, 1524.

ARTICLE IN PRESS
G. Veenstra / Health & Place 13 (2007) 1431 Post-industrial Stratication. The Johns Hopkins Press, Baltimore, pp. 5576. Kaplan, G., Keil, J., 1993. Socioeconomic factors and cardiovascular disease: a review of the literature. Circulation 88, 19731998. Mirowsky, J., Ross, C.E., 2003. Education, Social Status, and Health. Aldine de Gruyter. Muntaner, C., Lynch, J., 1999. Income inequality, social cohesion, and class relations: a critique of Wilkinsons neoDurkheimian research program. International Journal of Health Services 29 (1), 5981. Muntaner, C., Eaton, W.W., Diala, C., Kessler, R.C., Sorlie, P.D., 1998. Social class, assets, organizational control and the prevalence of common groups of psychiatric disorders. Social Science and Medicine 47, 20432053. Poland, B., 2000. The considerate smoker in public space: the micropolitics and political economy of doing the right thing. Health and Place 6 (1), 114. Prandy, K., 1999. Class, stratication and inequalities in health: a comparison of the Registrar-Generals Social Classes and the Cambridge Scale. Sociology of Health and Illness 21, 466484. Scambler, G., Higgs, P., 2001. The dog that didnt bark: taking class seriously in the health inequalities debate. Social Science and Medicine 52, 157159. 31 Schwalbe, M.L., Staples, C.L., 1986. Class position, work experience, and health. International Journal of Health Services 16, 583602. Veenstra, G., 2005a. Location, location, location: contextual and compositional effects of social capital in British Columbia, Canada. Social Science and Medicine 60, 20592071. Veenstra, G., 2005b. Can taste illumine class? Cultural knowledge and forms of inequality. Canadian Journal of Sociology 30 (3), 247279. Veenstra, G., 2005c. Social status and health: absolute deprivation or relative comparison, or both? Health Sociology Review 14 (2), 121134. Veenstra, G., 2006. Neo-Marxist class position and socioeconomic status: distinct or complementary determinants of health? Unpublished paper. Williams, S.J., 1995. Theorizing class, health and lifestyles: Can Bourdieu help us? Sociology of Health and Illness 17, 577604. Wright, E.O., 1978. Class, Crisis, and the State. New Left Books/ Verso UK, London. Wright, E.O., 1979. Class Structure and Income Determination. Academic Press, New York. Wright, E.O., 1985. Classes. Verso, London. Wright, E.O., 1998. Class Counts. Cambridge University Press, Cambridge.

Das könnte Ihnen auch gefallen