Sports Science(Quality work, no plagarism, A+)

QUEST. 19%. 48.221-242 © 1996 National Association for Physical Education in Higher Education

Approaches to Social Inequality in the Sociology of Sport

Peter Donnelly

sport, by its very nature, produces and reveals inequalities in terms of physi- cality and athletic perfonnance. [n social terms, however, sport has often been considered the great social leveller. This paper considers the production and reproduction of social inequality in sport in regard to the development of ideas and research about social inequality in the sociology of sport. Specifically, inequality in sport is considered in terms of the progression from reflection to reproduction to resistance theses and the progression from categorical to dis- tributive to relational levels of analysis. The paper concludes with a consider- ation of some ways in which sport may be involved in the production of social equality.

Competitive sport, by its very nature, produces and reveals inequalities in terms of physicality and athletic performance. Pierre de Coubenin referred to Olym- pic spon as a "democracy of ability," prestnnably in reference to his idealized view of an Olympics based on participation rather than outcome.' If democracy is con- sidered a social condition of equality, then the condition of having been selected as an Olympic athlete apparently makes one an equal among one's fellow Olympi- ans. More realistic, and perhaps more modem views of sport recognize not only that Olympic athletes are already part of a very select/elite group in terms of athleticism and physical giftedness, but also that the whole purpose of the Olym- pics and other sport competitions is to reveal inequalities among athletes and to rank those athletes in terms of athleticism and physical giftedness.

Sociological views of social inequality are far more complex. They tend to reflect the entire poiitical spectrum, from the right wing view that social inequality is inevitable (social Darwinism) or necessary to create incentive (conservatism), to the leî wing view that socia) inequality is wrong and should be eradicated by democratic (social democracy) or revolutionary (Marxist) means (cf. Figler, 1981, P- 4]). In focusing on equalization as a process, Williams (1983) identified two distinct meanings that reveal the complexities involved in studying and eradicat- ing socia] inequalities. In the first meaning, equalization is seen as a "normative principle" entailing a process "in which any condition, inherited or newly created,which sets some men above others or gives them power over others, has to

is a substantially revised version of a paper thac appeared in French as "Les Sociales Dans le Spon," in Sociologie et Sociites, 27(1). 1995, pp. 91-104.

Peter Donnelly is with the Departments of Kinesiology and Sociology ai McMaster University, Hamilton, ON L8S 4K1, Canada.

22i

222 DONNELLY

be removed or diminished" (p. 118). In the second overl^ping meaning, equaliza- tion is seen as a meritocratic process based on the premise that everyone should "start equal," and that "any subsequent inequalities [are] seen as either inevitable or right" (p. 118). As Williams wryly notes, the second meaning represents the "equal opportunity to become unequal" (1983. p. 118).

It is the fact that equality and inequality are measurable characteristics when comparing humans in terms of factors such as height, athletic ability, and possibly intelligence, that results in complications with regard to social inequality. As Williams notes,

what needs to be shown is that the measurable difference is relevant to the particular inequality, in a social sense: height would not be. though colour of skin has been held to be; energy or intelligence might be, and this is where most serious contemporary argument now centres. (1983, p. 119)

This paper focuses on the nexus between the measurable inequalities that form the basis of modem competitive sport and various social characteristics that have been socially constructed as the bases for human inequality.

The consequences of social inequality in sport are parallel to its consequences in the larger society—various forms of discrimination and self-discrimination, and the development of ideological views that justify the discrimination. Sociological approaches to social inequality are clearly reiterated in the sociology of sport; and in North American sociology of sport, research and theory show a progression of thought that characterizes changing views of social inequality in sport. This pro- gression is evident in the shift fi'om reflection to reproduction to resistance theses, and in the development from categorical to distributive to relational levels of analy- sis. The two main parts of this paper are concemed with examining the underlying theses and the levels of analysis as they bave impinged on the development of views of social inequality in the sociology of sport. The final part deals with the production and reproduction of social inequality in terms of the nexus between measurable inequalities and socially constructed inequalities; includes a brief ex- amination of the debate between culturalist and stmcturalist explanations for so- cial inequality in sport; and considers some ways in which sport has been, and has the potential to become, an agency for the production and reproduction of social equality.-

Underlying Theses

The sequence of theses in the development of sociology of sport, from re- flection to reproduction to resistance, was first proposed in a prelitninary way by Ingham (1980) who made note of the refiection (or representation) and resistance theses. The reproduction thesis emerged from Bourdieu and Passeron (1977). and the three were put together in terms of the development of sociology of sport by Ingham and Loy (1980) in a conference paper.'

Reflection

The most passive of the theses is the reflection thesis—"sport is a mirror [or microcosm] of society." It was most popular in the 1960s and 1970s during the early development of the sociology of sport, although it is still reiterated in some text books. The reflection thesis is connected with structural fiinctionaiism. the

APPROACHES TO SOCIAL INEQUALITY 223

predominant theory at the time in American sociology and the most significant influence in the origins of North American sociology of sport (cf Luschen, 1967).

In the case of functionalism (and its structural-functional variants), a presump- tion is made that spon reflects society . . . it is axiomatic that the material structures and ideological dimensions of "society" and those of sport in some way correspond. (Ingham, 1980, p. 1)

The reflection thesis permitted both positive and negative thinking about sport, depending on the analyst's political position with regard to dominant values. Sport was either an institution that socialized individuals into conformity with approved dominant values or, for the more critical functionalists, especially those concemed with social inequality and social control—racism, sexistn, tnilitarism, and so forth (e.g., Hoch, 1972)—was an institution that socialized individuals w i ± unaccept- able values. It quickly became apparent that the reflection thesis did not explain anything; it just described an obvious situation—^that sport was a part of society and must therefore reflect that society. But it was an important stage for sociology of sport to pass through because of the need to overcome the concurrent thinking about sport—that it was a distinct sphere somehow separate and different from social life, even transcending social life (e.g.> Novak, 1976).

Reproduction

The 1970s saw the emergence of the reproduction thesis, a more dynamic thesis that suggested sport was not just a passive reflection of society, but a force actively involved in the reproduction of a status quo in society from generation to generation. Notions of social reproduction are rooted in more critical aspects of sociology and indicate that sport is actively involved in maintaining an unequal status quo in which more powerful groups in society retain their power over subor- dinate groups. This thesis became relatively popular in the sociology of sport through the Westem countercultural critique of sport (see Mandeil, 1990) and from the availability in English of the worlcs of such European neo-Marxist sport sociolo- gists as Brohm (1978). Rigauer (1981), and Vinnai (1973). The reflection thesis continued to be prominent throughout the i 970s; however, it became apparent that the reproduction thesis could also be used to characterize functionalist thinking in which sport is seen to play an active role, through the process of socialization, in contributing to social system needs such as adaptation, integration, goal attain- ment, pattem maintenance and tension management—all of which are concemed with the maintenance of a status quo view of the social order.

Resistance

The early 1980s marked a key turning point in North American sociology of sport. A new level of theoretical sophistication was evident at the Queen's Univer- sity (Canada) conference on "Sport and the State" (Cantelon & Gruneau, 1982); the inaugural meeting of the North American Society for the Sociology of Sport (NASSS) in Denver was marked by criticism of extant positions: theoretical devel- opments continued through 1981 at the 1st Regional Symposium of the Intema- tional Committee for the Sociology of Sport in Vancouver (Ingham & Broom, 1982) and the NASSS Fort Worth conference (Dunleavy, Miracle, & Rees, 1982; Rees & Miracle, 1986); and the feminist critique of sport was prominent at the

224 DONNELLY

1982 NASSS Toronto conference (Theberge & Donnelly, 1984). By this time, the shortcomings of the reproduction thesis were beginning to become apparent. Indi- viduals were rendered as passive agents in this active but one-way process. They were characterized as falsely conscious consumers of the new "opiate of the masses" (sport), unaware of the forces involved in producing and reproducing inequality and maintaining their subordinate status; or they "were characterized as passive leamers "molded' and 'shaped' by 'society'" (Coakley, 1993, p. 170)."

In the resistance thesis, sport is seen as "contested terrain," and individuals are seen as active, self-reflexive agents (a) who "might quite consciously value sports as meaningful and beneficial aspects of their lives, while at the same time being aware that ruling groups attempt to use sport as an instrument of control" (Hargreaves, 1982, p. 43; Gruneau, 1983, pp. 151-152); (b) who have the capacity to change the conditions under which they practice sport and recognize and change the conditions that maintain their subordinate status; and (c) whose attempts at resistance sometimes have an opposite effect, serving to reinforce the conditions of their subordination. The resistance thesis does not deny or replace the reproduc- tion thesis, but rather supplements it by introducing the notion of agency to stand in a dialectical relationship with the structural aspects of the reproduction thesis. The resistance thesis is also evident in Coakley's view of "socialization-as-interaction" in which "the tensions, negotiations, misunderstand- ings, and resistances that characterize the social relationships associated with a person's entry into and participation in [sport]" need to be taken into account—the sport participant is seen "as an active agent, as someone with self-reflective abili- ties and creative potential, as someone seeking autonomy and affirmation of iden- tities, as someone who is not simply a passive leamer" (1993, p. 170).

Development through the tliree theses has had a profound effect on our think- ing about social inequality and how it is produced, reproduced, and resisted in sport. Initially, social inequalities in sport were seen simply as a reflection of the social inequalities evident in society. Subsequently, social inequalities came to be seen either as a consequence of hegemonic forces attempting to maintain (repro- duce ) an unequal status quo from which they were the principle beneficiaries, or as an incentive to motivate individuals to seek social mobility (cf Davis & Moore, 1945). Most recently, social inequalities have come to be seen as a part of the contested terrain of modem sport in which what is at stake is a more democratized, less unequal form of sport. This outcome is, of cotirse, opposed by those who find privilege in the status quo and who benefit from the present conditions of social inequality (e.g., white, heterosexual, able-bodied males from the middle and upper classes); and such opposition may, in tum, be resisted by those seeking to overtum the status quo.

Levels of Analysis

The theoretical progression in the sociology of sport has been paralleled by three levels of analysis: categorical, distributive, and relational.^ It should be noted that these levels of analysis are not strictly parallel to the reflection, reproduction, and resistance theses respectively (i.e., research assuming the reproduction thesis was not all conducted under the distributive level of analysis). However, there is a temporal relation, and both typologies reflect the increasingly sophisticated level of interiH-etation and analysis in the sociology of sport.* As with the theses, these

APPROACHES TO SOCIAL INEQUALITY 225

levels of analysis are representative of the ideological positions taken by research- ers examining social inequality in sport. However, they provide an even more im- portant guide to interpreting how we have come to understand the production and reproduction of social inequality in sport. In the following discussion, most atten- tion is paid to categorical analysis because it establishes the socially constmcted bases for social inequality in sport; less attention is given to distributive analysis, although it does provide empirical support for the effectiveness of the social cat- egories; and least attention is paid to relational analysis because its assumptions and critiques are implicit throughout the paper.

The following sections describe each of the levels of analysis and consider how each has been used to examine three bases for social inequality in sport: so- cial class, gender, and race/ethnicity. The selection of these three is purely a reflec- tion of the amount of research that has been conducted on social inequality under these categories and is not an indication that they represent the more important forms of discrimination. Any social characteristics that have, at any time and in any place, been interpreted as barriers to involvement or participation (e.g., age, religion, mental and physical disabilities, sexual orientation, region of origin) could be considered in an examination of social inequality in sport. It should also be noted that distinguishing social characteristics in this way adds a level of distor- tion to the analysis because none of the characteristics are discrete. Each of us has a gender, a social class background, a racial/ethnic affiliation, and a variety of other social characteristics that serve as advantages or disadvantages in the various stmctural and cultural circumstances of our lives, and we relate to each other on the basis of "ours" and "theirs." The articulation between the three sets of social relations that have produced the most research is only now beginning to be explored.

Categorical Analysis

Categorical analysis is concemed with identifying differences in character- istics and/or behavior between categories of individuals that have been established and defmed as having some significance. Thus, categorical research constructs the categories of interest (e.g., skin color but not hair color) and attempts to discover and explain differences between representatives of different social classes, racial/ ethnic groups, males and females, or any other serial categories. For the purposes of sociology of sport, such differences would be in terms of sport practices or preferences, or in terms of the characteristics presumed to predispose individuals to different sport practices or preferences (e.g., biological, psychological, geo- graphical, or other characteristics used to describe individual differences).

Social Class. A great deal of research shows that sport preference varies by social class, and a number of theories have been put forward in an attempt to explain this finding on the basis of psychosocial differences between the classes. For example, Luschen (1969) proposed that the upper-class preferences for nov- elty (new experiences and objects), individual achievement, and exclusivity are reflected in a preference for "new" sports, individual sports, and sports dependent on organization into clubs. Yiannakis (1976) used sport stmcture, cost, publicity, and physical contact as the distinguishing characteristics in his finding that higher status individuals showed a preference for individual and expensive sports (reflecting autonomy in higher status occupations and the ability to pay) and

226 DONNELLY

shunned publicity (in favor of sports that are out of the limelight, reflecting a pref- erence for and an ability to ensure privacy) and aggressive physical contact (an- other occupational reference, in this case to mental work in contrast to the physical work of the lower classes). Renson (1976) attempted to show that "the sports piayed by a particular social class serve as symbols of their status and function within society" (Loy, McPherson, & Kenyon, J978, p. 366). This status symbolism is evident in higher class sports characterized by their use of "status sticks" (cf maces and batons); upper-middie-class interest in "nature" sports; Jower-naiddle-class in- terest in sports using targets, nets, and balls; and lower-class involvement in indi- vidual and contact sports.

Such explanations represent middle-range theorizing—in their attempt to explain observed events, these middle-range theories do not take into account the larger cultural and political economic issues involved in the idea of class. They are coDcemed with identifying the apparent characteristics of sports and social classes, and they impute class differences in choice of sport—^but these choices must be seen to occur in a structural situation in which the power to choose varies enor- mously. For example, Donnelly (1993) has identified four processes by which sports democratize—a situation of apparent choice. In three of these (recruitment, the doctrine of good works, and modelling) the element of choice is clearly controlled by the dotninant class. Only in the process of acquisition is some choice evident, and even the acquisition of access to sports takes place in a constraining environ- ment that limits the choices available. The middle-range theories also appear to describe particular local situations rather than explain them. For example, the same sport may represent quite different levels of social status in different societies— "gyrtmastics is low in Belgium . . . at the middle in Germany . . . and at the upper-middle in the United States" (Loy et al., 1978, p. 366).

Bourdieu (J 978) also appears to engage in categorical analysis in his depic- tions of aristocratic/upper-class sports as "disinterested" practices, tniddle-class sports as representative of the "developmental" ethic, and working-class sports as expressive of an "instrumental" relationship with the body. However, these cat- egories are much more rooted in historical notions of class, and they lead to dis- tributive and relational levels of analysis (Bourdieu, 1984). His work gives a much more clear sense that

spOTting fffaaices . . . are articulated and negotiated within a set of historically develc^wd and socially prcxfaiced limit^OTis and possibiJiiies. . . . They are im- pOTtant sites where the social relations of power and privilege are negotiated, developed and givai particular meaning and represairaoCTi (Dewar, 1991, p. 18)

Research on biological differences has been largely absent from consider- ations of social class. Feminist research has been most significant in drawing our anention to the reflection of ideology through biological science—namely, the desire to establish difference (i.e., categories; cf Birrell & Cole, 1990; Davis & Delano, 1992). Feminist research, however, has tended to focus on the political effects of biological differences between males and females, and it has also sensi- tized us to the effects of emphasizing racial/ethnic difference. Notions of social mobility have tended to limit emphases on biological/psychological differences between social classes—^Bl^ks cannot become Whites, nor females become males without biological intervention,^ but lower-class individuals can sometimes become middle-class individuals without such interveDtion.

APPROACHES TO SOCIAL INEQUALITY 227

There are examples, however, of "scientific" attempts to discover such bio- logical differences between social classes and a great deal of evidence of popular belief in such differences. Social Darwinism, for example, imputes biological/psy- choiogicaJ differences between classes. In the case of sport, the achievements of working-class athletes have, in the past, often been attributed to characteristics of body size, shape, and habitus (such as the "naturalness" of hard work to such individuals). The amateur-professional distinction was, in part, a result of such beliefs: working-class cricketers were usually assigned to the task of bowling (pitching), defined in the late 19th century as more menial/manual than the gentle- manly skill of batting; Victorian mountaineers attempted to explain the climbing skill of their guides in terms of the physiological adaptation of their feet (Wherry, 1896); and even in the 1950s, when working-class rock climbers were beginning to make their mark on the sport, their achievements were often attributed to their small size and powerful build (prestunably the result of some Lysenkian inherit- ance of characteristics acquired in several generations of manual work in factories and mines).

Gender. In contrast to social class, categorical analyses of gender have focused almost excltisively on "discoveries" of the biological and psychological sciences regarding gender differences. These were, and still are in many places, regarded as sufficient explanation of the differences in sport performance and prac- tices between men and women. Notions about the "apparently" greater height, strength, and speed of men, and about the vulnerability of women's reproductive organs, are presented as "facts" regarding "natural" differences between men and women. Similarly, masculinity and femininity have been regarded as personality traits, and these, together with other traits (dependence, dominance, etc.), have been assessed as determinants of difference between men and women. When the physiological characteristics and personality traits are combined with the presumed requirements for most modem sports, it comes as no surprise that a close fit is found between male biology and psychology and those requirements. Thus, fe- male athletes were presumed to experience not only lower levels of performance (biology as destiny) but also role cotiflict resulting from going against their "natu- ral femininity" in the pursuit of athletic excellence. When such arguments are pushed to their extreme, the achievements of female athletes may be attributed to their "masculinity," resulting in ideas concerning the sexuai orientation of female ath- letes.

A great deal of research and theoretical argument in the sociology of sport for the last 10 years has been devoted to deconstructing and discrediting these arguments (see, for example, Birrell, 1988; Coakley, 1994; Hall, Slack, Smith, & Whitson, 1991; and Willis, 1982, for summaries of these critiques). Dewar (1991) has specifically shown how such categorical analysis is based on a set of decontextualized "facts" that take male ^hievement as the standard of compari- son. But the categories of masculinity and femininity are stubbornly persistent as a basis for understanding sport inequality, and as Theberge (1985) has noted, sport inequality becomes an imponant site for the reproduction of social inequality by transforming women's presumed physical inferiority into social inferiority.

Race/Ethnicity. Such is the power and ingenuity of white, male, middle/ upper-class hegemony that racial subordination is based on an opposite sei of bio- logical parameters to gender subordination. If women are presumed to be biologi- cally inferior (in everything but childbearing and nursing) and therefore socially

DONNELLY

inferior, Blacks are presumed to be biologically suf^rior (in terms of certain sport-specific skills) and therefore socially inferior! Once again biologically de- terminist arguments have been employed to "explain" the supposedly superior jumping and sprinting ability of Blacks and the supposed inferiority of Blacks in sports such as swimming."* Psychological arguments regarding differences in in- telligence and other traits have also been employed to justify pattems of positional segregation in team sports. Thus, while lower levels of athletic achievement by women in certain sports are employed to justify their social subordination, the high levels of achievement by black athletes in certain sports have been attributed to natural ability and physiological advantages—^presumably in contrast to the hard work and intelligence of white athletes (cf Davis. 1990).

Again, a great deal of work in the sociology of sport has been applied to deconstructing and discrediting these examples of bad science (for summaries of these critiques see Coakley, 1994, and Phillips. 1993), but the idea of difference and advantage persists and appears to have a distinct effect on sport selection— in terms of counseling by coaches and physical education teachers, and in terms of self-selection by young athletes. In contrast to the idea of gender difference, which has a long history with regard to sport participation, the idea of racial dif- ference and advantage became a widely accepted post hoc explanation of observed events once sport had begun to integrate, and black athletes began to win. A people who had been enslaved and maintained in a subordinate social position for so long could not be permitted the same explanations for athletic achievement as white athletes—hard work, persistence, intelligence, and so forth. The biological explanations, derived from earlier attempts to search for difference once race be- came established as a socially significant category (cf. Hoberman, 1992),'' served to demean the achievements of black athletes and maintain their social subordi- nation.

Social arguments sitnilar to those developed for social class have frequently been employed to establish ethnic categories and explain observed differences in spon selection and participation rates, even in playing .styles.'" In the Canadian case, most research has been concemed with the categories of anglophone and fi-ancophone, and observed differences. Hall et a!. (1991) have summarized vari- ous explanatiotis proposed to account for positional segregation by language group in professional ice hockey (from cultural difference to discrimination). Tliey also summarize the most prevalent debates conceming the underrepresentation of francophones on Canadian national teams and the different patterns of sport par- ticipation between the two language groups by asking whether these result fi-om a lower emphasis on sport in francophone culture (reflected in the underdevelop- ment of amateur spon until the 1980s and the lack of physical education and interschool sport programs in many Quebec schools) or from prejudice and dis- crimination. Though it seems likely that both culttirai differences and discrimina- tion combine to produce the observed pattems, adopting one position or the other has produced some heated debate (cf. Curtis & White, 1992; Laberge & Girardin, 1992; McAll. 1992; White & Curtis. 1990a. 1990b).

The decision to establish categories based on ascribed characteristics; the decision to attribute social significance to some of those categories (e.g., weight more so than height); and the decision to seek support for the establishment of those categories by seeking differences employing bio-, psycho-, and socioscientific means—each of Uiese decisions is ideologically loaded, inflecting the power and

APPROACHES TO SOCTAL INEQUALITY 229

the ideology of the decision makers. And the results of such decisions provide powerful support for W.I. Thomas' dictum that, if people defme a situation as real, it has real consequences. It is difficult to imagine what sport or sport sociology research would be like if such categories had not been established or if others had been constructed. Whatever their form, such categories provide the basis for re- search employing the distributive level of analysis.

Distributive Analysis

Dewar has noted that distributive analysis is concemed with "the nature and extent of opportunities available to different categories of individuals" (1991, p. 19), and that such analysis is particularly concemed with the nature of social in- equality. Categorical analysis is concemed with the individual bases of inequality, whereas distributive analysis draws our attention to social stmcture; and within liberal democracy it emphasizes Williams' (1983) second meaning of eqtiality (viz., equality of opportunity, or meritocracy). Dewar notes that research questions con- cem "who gets what, and why" in terms of "the numbers of programs, the funding, facilities, and opportunities for mobility to leadership positions" (1991, p. 19). Following are brief examples of the results of this type of research with respect to the three selected bases for inequality.

Social Class. Hargreaves summarizes a large body of research showing that "the higher the class, the greater the rate of panicipation" in sport and physical activity (1986, p. 96). Even the most recent stirvey data from American high schools show strong evidence of differential rates of involvement by social class (Fejgin, 1994). These findings are particularly evident at the level of high-performance sport in Canada, and they appear to persist even after government efforts to shift to a more meritocratic basis for selection. For example, Gmneau (1976), using 1971 data, showed an overrepresentation at the Canada Wmter Games of athletes with fathers from the highest occupational categories and an underrepresentation of those from the lower occupational categories. Despite the implementation of fed- eral government programs to overcome the barriers to participation (primarily de- fined as financial). Beamish (1990) found that similar pattems persisted over 15 years later.

Gender Recent data indicate that in "advanced" societies, adult men and women show similar rates of participation in recreational physical activity—for example, Canada (Stephens & Craig, 1990) and Switzerland (Lamprecht & Stamm, in press). However, the typical individuals most likely to participate in at least one competitive sport are still "young, well-educated, white males who belong to the professional or managerial occupational categories, and who have relatively high incomes" (Hall et al., 1991, p. 156)." And female athletes still show a marked decline iti participation during adolescence:

In the 10-14 age group, 49% of the female population are involved in sport, versus 72% of the male population. Beginning at age 12, involvement in girls declines steadily until only 11 % are involved in physical activity by grade 11 (age 16-17). (Best, Blackhurst, & Makosky, 1992, p. 149)

These results are also evident in Fejgin's (1994) study in which there are signifi- cantly lower rates of panicipation among American high school females than males. There are also unintended consequences of government policy evident with regard

230 DONNELLY

to gender. Since the passage of Title IX of the Educational Amendments Act in the United States, a law intended to increase opportunities for female athletes (and which has, to a certain extent, had that effect), there has been a very significant decline in the number of postparticipation opportunities for women athletes in terms ofcoaching and administrative positions (Stangl &Kane, 1991).

Race/Ethnicity. A great deal of data collected in the United States (sum- marized, for example, by Coakley, 1994; Curry & Jiobu, 1984; and Phillips, 1993) provide empirical support for the observation that Blacks are overrepresented in some sports (e.g., American football, baseball, basketball, and certain events in track and field) and at some positions in the team sports (noncentral). Those stud- ies rareiy go on to make the point that Blacks are underrepresented in the vast majority of sports; but recent concems have shown that, even in the sports in which Blacks are overrepresented, the opportunities to move on into leadership positions (coaching and administration) are extremely rare in comparison to those available to the white athletes. There are very few studies that have considered representa- tion on the basis of other racial or ethnic categories. In the francophone/anglophone Canadian examples noted previously, pattems still appear to show lower represen- tation of francophones on national teams (Laberge, 1989) and lower levels of par- ticipation in competitive sports among francophone males in comparison to anglophone males (Curtis & White, 1992).'-

By focusing on the idea of opportunity to participate, distributive analysis has a number of limitations because increased opporttmity does not automatically lead to increased participation. For example, upper- and upper-middle-class indi- viduals usually enjoy the broadest possible opportunity set in any given society, and yet there are examples of their withdrawal from participation in various activi- ties (e.g., boxing), and of activities in which they have shown little interest in participation in recent years (e.g., bowling). Similarly, there are inexpensive ac- tivities that do not enjoy high levels of participation from low income individuals (e.g., hiking). Also, in liberal democratic terms, it is the provision of opportunities to piay that is considered to be important, not the encouragement of a more com- plete democratization that would necessitate ftill involvement in an activity—that is, the power and the right to determine the forms, circumstances, and meanings of participation (Donnelly, 1993). A further limitation to the idea of opportunity in- volves the observation that, in order for any individual to take the opportunity to participate in any given activity, that activity has to have the capacity to become meaningftil to the individual. What is the point of increasing the opportunity to participate in a sport for an individual who is enmeshed in a set of social relations that make such participation meaningless?

Relational Analysis

Relational analysis shifts beyond the notion of discrete categories and the distribution of resources among those categories to consider whole sets of social relations—among men and women, social classes, and racial/ethnic groups; and it examines such relationships in their social context (e.g., among men and women who belong to social class and racial/ethnic groupings in a particular time and place). In relational analysis, sport and other cultural forms are seen as "cultural representations of social relations;" and such analyses "begin with the assumption that sporting practices are historically produced, socially constmcted and cultur-

APPROACHES TO SOCIAL INEQUALITY 231

ally defitied to serve the interests and needs of powerful groups in a society" (Dewar, 1991, p- 20). Such power is not expressed in a simplistic conspiratorial sense, rather, it represents the power to create cultural forms to one's own tastes; the power atid motivation to maintain the status quo that supports one's privileged position; and the power to make such hegemony (in a Gramscian sense) appear to be natural and normal to the dominated.

"Understanding how this hegemony is negotiated and contested is crucial in relational analyses" (Dewar, 1991, p. 20) because such analyses also begin with the assumption that hegemony "is never total or exclusive" (Vfliliams, 1977, p. 113). Relational analyses are frequently conducted employing ethnographic methods in order to determine the meaning of sports to individuals representing various social categories, and the relationships among those individuals.'^ However, although Dewar (1991) argues that distributive analyses are frequently conducted from a liberal democratic ideological position, the data provided frequently make a sig- nificant contribution in developing a powerful critique of practices in modem sport and to supponing relational levels of analysis.

Because relational analysis is the most recent to emerge, and because stud- ies are time consuming to conduct, there is little representative research for the three selected bases for social inequality. However, studies by Gruneau (1983) and Hargreaves (1986) provide examples of the analysis of class relations in sport; studies by Birrell and Cole (1990), Birreli and Richter (1987), and the collection of work edited by Messner and Sabo (1992)''' provide examples of gender rela- tions in sport; and studies by Birrell (1989), Majors (1992), and Foiey (1990) all examine racial/ethnic relations in sport. Because relational analysis is most likely to be sensitive to context, it has been at the forefront in recognizing that distortion is introduced by separate analyses of class, gender, and race/ethnicity, and in com- bining such analyses in research. For example, in Birreil's (1992) study of gender and race she notes that "The most effective blending would highlight not only class relations and gender relations, but racial relations as well" (p. 185). Also Coakley and White's (1992) study of English adolescetits blended class and gen- der analyses.'^

Relational Analyses, Democratization, and Resistance

The Case for Relational Analyses

Tlie citius, altius, fortius nature of modem sport is designed to expose in- equalities between human beings. Some will be faster, stronger, and/or more skill- fiii than others, a status they will have achieved through a combination of genetic inheritance and training. Such measurable differences between human beings may, in the case of high levels of success in certain sports, result in upward social mobil- ity. Sport success can lead to wealth and access to education that may reduce so- cial inequality for some (and at the same time perhaps produce even greater sociaJ inequalities for many others). However, although measurable differences in alh- letic performance may result in less inequality based on socia! class (cf. Page, 1973, p. 28, on "the embourgeoisement of the jocks"), the same situation is not so apparent for gender and racial inequality (always given that the articulation of the three conditions of social inequality under consideration makes discrete arguments of this type rather artificial).

232 DONNELLY

At first, continuing racial and gender inequality and the lack of social mobil- ity for the majority of athletes (and the temporary social mobility for some) was seen in the sociology of sport as a straightforward reflection of the societal status quo. Then it began to become apparent that sport was modeling inequalities in the larger society and, along with education, could be seen as a part of the apparatus to prepare young people to accept such inequalities. Competition in sport revealed physical inequalities, and these served to confirm the rightness of social inequali- ties—though always holding open the possibility of social mobility for those suc- cessful in the competition. The growing awareness of such modeling at a time when long-entrenched social inequalities were being questioned (i.e., the civil rights and women's movements) led to the view in the sociology of sport that, even more than modeling social inequality, sport was an active agent in the production and reproduction of social inequality.

To retum to Williams' question, "is . . . the measurable difference . . . rel- evant to the particular inequality, in a social sense?" (1983, p. 119). Or, to put it more bluntly, does athletic achievement make an individual more valuable as a human being? When the meaning of sport is seen in terms of entertaitmiem and commercial enterprise, the material and symbolic (prestige, status) rewards given to successful athletes suggest that physical superiority may sometimes be tratis- formed into a certain degree of social superiority. But sport is clearly an active agent in the production of racial and gender inequality. In both of these cases, science and ideology have formed a powerful alliance to attribute meanings to measurable physical and performance characteristics. Both "inferior" and "supe- rior" levels of performance are constructed as indicators of social inferiority. It is possible then to identify, in general terms, the relationship between physicality and social status within each of the three sets of social relations under consider- ation:

Soda] class: physical superiority = social superiority (for those originating in a subordinate social class)

Gender: physical inferiority = social inferiority (for women)

Raca'Ethnicity: physical superiority = social inferiority (for those with some African heritage living in a mixed race society)

The contradictory interpretations of black and female athletic accomplish- ments, and perhaps the persistent search for fiaws in those who have achieved upward social mobility through sport, suggest that the nexus between the measur- able inequalities that form the basis for modem sport and the various social char- acteristics that have been socially constructed as the bases for human social in- equality is far more complex than we previously believed and needs a g r e ^ deal more analysis.

Such analyses both require and affinn a relational level of analysis. Con- siderations of separate categories (e.g., lower class/tniddle class; mEJe/female; black/white) and the distribution of resources or opponunities among those cat- egories, fall far short of engaging with the complex social reality implied by the social construction and definition'^ of, and relations between, the established cat- egories. The determination, for example, that men are stronger or more aggres- sive (on average) than women (categorical analysis), or that men receive far tnore opportunities and resotirces than wotnen (distributive analysis), leaves open a

APPROACHES TO SOCIAL INEQUALITY 233

great many questions. What value is placed on strength and ag^ession and why? How and why have men constructed and maintained "sport" as a gender distin- guishing activity? How have men been able to maintain the greater share of op- portunities and resources? And. how and why are changes in the relationships between men and women occurring, and what part does sport play in such changes?"

Sport and Democratization

A categorical level of analysis h ^ not really lent itself to considerations of process and change, although distributive and relational levels of analysis have been concemed with changes in the distribution of opportunities and resources and with changes in social relatiotis, respectively. However, raising the issue of process and change indicates that much of the preceding discussion, and much of the work on social inequality in the sociology of sport, has failed to distinguish concepttially between two overlapping processes—democratization through sport and democratization of sport. Democratization through sport involves a process whereby involvement in sport is presumed to ^sist in reducing overall social in- equality; the democratization of sport involves the reduction of inequalities within the sport environment. Of course, it is frequently assumed that the latter will lead to the former, although the processes by which this might occur have not been fully explored."*

Democratization Through Sport. In the early years of North American so- ciology of sport, democratization through sport was frequently expressed in terms of sport as the great (social) leveller (Goodhart & Chataway, 1968. ch. 3). Sage described this presumed process as follows:

Spon has often been responsible for breaking down social class lines, racial and ethnic barriers, and sex role stereotypes. Sport often is the common de- nominator for members of different social classes—they may use sport as a focus for conversation or may band together to participate or suppon a spon team. Thus members of different social classes may unify in the cause of spon. . . . [Spon} has played a prominent role in American life in assimilating differ- ent races and ethnic minorities into American culture . . . [and T]he liberation of the female has its linkage with spon. (1974. p. 398)

Such ideas lie behind the notion of sport as "war without weapons" (Goodhart & Chataway, 1968; see also Lorenz, 1967) and de Coubenin's ideas about the Olympic family and the democracy of ability. And though it is true that many examples exist of good fellowship and racial/ethnic, gender, and class harmony and democratization being facilitated by sport, it is equally true that racial/eth- nic, class, gender, and other forms of hostility might equally well be engendered t>y sport.

The evidence does not support any widespread democratization through sport. In terms of social class, there has been a clear expansion of the tniddle class in Westem societies in the last 30 years, with many who formerly identified them- selves as skilled working class now identifying themselves as middle class and embracing middle-class values. This in itself may be interpreted as an example of democratization, but it does not appear that sport has had any role in the process. The expansion of the middle class appears to be leading to a democratization of

DONNELLY

formerly middle- and upper-middle-class sports such as golf (Bragg, 1995). The increased interest in golf, however, does not appear to be driving any reduction in social inequality. Evidence in the last 15 years points to a growing gap between the rich and poor in many Westem societies. A recent report from the Joseph Rowntree Fotmdation, Income & Wealth, identifies rising inequality in a number of Westem societies (e.g.. New Zealand, UK, USA; Droban, 1995); and recent United Na- tions statistics show that North America has the highest rates of child poverty among advanced societies. Poverty represents a real class barrier to participation in sport.

With regard to race, there is also an expanding black middle class in North America (and it is here that the artificial distinction, for purposes of analysis, be- tween class, race, and gender becomes apparent). Although some of this expan- sion may be a result of highly paid and scholarship athletes (as it is with the social mobility of white lower-class athletes), it is much more obviously a result of equal opporttmity and affirmative action programs in education and the workforce. How- ever, there is evidence of a form of racial democratization through sport in South Africa, where the refusal to participate had an impact on democratization (cf. Harvey &HouIe, 1994; Kidd, 1991).

Recent data from Statistics Canada (Gadd, 1995) show that the gap between men and women is closing, but greater Involvement and achievement in sport by women appears to be a part of a whole pattem of greater involvement and success in ail aspects of society rather than an engine driving the change. And, as noted above, in the cases of both race and gender, sport participation may actually have served to slow down the process of democratization. In terms of race, sport soci- ologists are familiar with the various arguments suggesting that black youths are encouraged to see greater opportunities for success through sport rather than through education; the tendency to define black success in sport as a result of natural abil- ity tends to affirm traditional racial notions and close off other avenues of opportu- nity. In terms of gender, the continuing image of women adopting traditional male models of sport, and failing to achieve the same levels of performance as males, may model and serve to affirm a subordinate status. Thus, sport has rarely suc- ceeded as a great leveller, and, as outlined above, sport is now frequently seen as an active agency for the production and reproduction of social inequality (see, for example, various articles in Harvey & CanteJon, 1988).

Democratization of Sport. With regard to the overlapping process of de- mocratization of sport, there is evidence of an apparent reduction of social in- equality through increased involvement in sport and physical activity by groups whose access had in the past been limited (e.g., women, iower-ciass individuals, older persons, and persons with disabilities). The similar rates of recreational par- ticipation for men and women in Canada and Switzerland are examples of this (Lamprecht & Stamm, in press; Stephens & Craig, 1990). This democratization has been a result of various social movements promoting social equality, human rights (including the right to participate in sport that lies at the basis of the Sport for All movement), and health. However, there is evidence of a backlash in some cases (e.g., the current moves to reform Title IX legislation in the US) and indica- tions that any apparent democratization is only partial. As Donnelly noted,

A fully democratized spon and leisure environment would include both the right to participate, regardless of one's particular set of social characteristics,

APPROACHES TO SOCIAL INEQUALITY M5

and the right to be involved in (tetenninmion of the forms, circumstances, and meanings of participation. (1993, p. 417)

The current democratization of sport appears to have included increases only in the right to participate.

Lamprecht and Stamm (in press) go even further than this in their contribu- tion to a current debate in European sociology of sport over the issue of democra- tization. Although there is widespread acknowledgement of a process of expan- sion and inclusion in sport and physical activity, a contradictory process of greater distinction and differentiation (cf. Bourdieu, 1978; Gebauer, 1986) is also identi- fied;

Thus, new opportunities and meanings of sport contain new forms of social distinctions and social limitations. . . . There are differences between certain sports, between certain sport places, between motives and, quite generally, between "real sport" and "nonsporting sport" which also render possible the demonstration of social differences. (Lamprecht & Stamm, in press)

The debate concems whether the process of democratization of sport is neutral- ized by the process of differentiation through sport. Thus, even the production of new opportunities in sport may be accompanied by the production of new grounds for social inequality.

This discussion of democratization supports the claims of the resistance the- sis to be used in conjunction with a relational level of analysis. The process of democratization stands in a dialectical relationship with the production and repro- duction of social inequality. Although sport is clearly involved in the reproduction of socia] inequality, the existence and (partial) success of democratizing move- ments—(such as the women's, antiracist, and antipoverty movements)—or sport-specific democratizing agents (such as Title IX or the Sport for All move- ment) attempting to reduce social inequalities are clear indications that the pro- duction and reproduction of social inequality do not go unchallenged (i.e., are resisted). Therefore, this must be taken into account in any analyses of sport and social inequality.

Culturalism vs. Structuralism and Resistance

In further support of the need for more relationaJ analyses, even when it is accepted that social inequality is produced and reproduced in sport, a conflict ex- ists between culturalist and structtiraiist arguments. From a culturalist position, choice of sport participation is assumed to be voluntary, and different cultures and values are presumed to result in the different choices that are made regarding sporl involvement. Culturalism may take a conservative fonn in which some values are deemed superior to others (cf Simpson, 1987; Vedlitz, 1988), or a more liberal form in which no value judgements are made about consumer, lifestyle, or physi- cal activity choices (cf White & Curtis, 1990b). Unfortunately, those adopting a culturalist explanation sometimes fail to consider the social structure within which a particular culture was produced—a stmcture which, for subwdinate groups, niay constrain opportunity or inclination to become involved in certain physical activities. Structuralism is characteristic of neo-Marxist work in the sociology of spon (e.g., Brohm, 1978). From a structtiraiist position, choice is determined ^ d limited by one's location in the social structure, and discrimination against

236 DONNELLY

subordinate groups is the major determining factor in restricting choice of sport involvement.

The problem with adopting a purely culturalist or a purely stmcturalist posi- tion is that both may be quite deterministic. Bourdieu (1984) identifies the twin determinations of culture and structure respectively as (a) class (and presumably racial/ethnic and gender) habitus as a determinant of taste, lifestyle, and consumer choices; and (b) a social space (such as the space of sport practices) characterized by power relations and ideologies that discriminate and therefore restrict access. Individuals make their choices about participation from a position that involves both cultural and structural constraints. But they may also step outside their par- ticular habitus and invade restricted social space—for example, in the case of girls wirming the right to play on boys' teams. This involves great difficulties and may not always be successftil, but without such resistance there would be no social development in sport.

A young Afro-Canadian male from a public housing project may face a double struggle in his desire to become a figure skater: the structural constraints of pov- erty and the potential racism of a majority-White environment, and the cultural constraints of class, racial, and gender cultures in which figure skating is not seen as a legitimate choice of sporting practice. That same male choosing basketball would find an enabling structural and cultural environment in which poverty and racism are less constraining; in which the class, racial, and gender cultures would deem the choice appropriate; and in which there would be a positive expectation of skilled performance and success. But it is important that, however difficult, the choice between figure skating and basketball is seen to exist. Cultures are pro- duced within structural environments that both enable and constrain; those envi- ronments change; and the dynamics of cultural production form a dialectical rela- tionship with the structural envirormient to produce choices that frequently con- form to cultural expectations, but that may sometimes resist both the cultural ex- pectations and the stmctural environment. The resistance thesis and a relational level of analysis endorse this articulation between structure and agency and pro- vide a sound basis for the study of social inequality.

In addition to the potential for individual challenges to social inequality in sport, there are a number of agencies and movements that have both enabling and constraining eifects with regard to overcoming social inequality in sport. Following is an outline of a number of examples of these, including, the cotnmer- cialization of sport, govemment involvement, educational institutions, and social movements.

Commercialization of Spon. The market is having a significant effect in overcoming inequality through the commodification and commercialization of sport. Money has no gender, race/ethnicity, or class; and since everyone is a potential customer for the goods and services of the sport industries, it is in the interests of the sport industries to expand their cotisumer base. Thus, the market can have the effect of reducing social inequalities and increasing the democratization of sport. And although the market tends to be govemed by the status quo and is more likely to follow trends that are the products of culture than it is to itiitiate them (e.g., greater women's involvement in a variety of activities), the provision of inexpen- sive sports outfits, equipment, and facilities has brought a number of activities within reach of individuals who would not have been able to afford them in the past. Of course, the sporting goods industry also seeks and creates

APPROACHES TO SOCIAL [NEQUALiTY 237

markets; and sportswear, equipment, and facilities may just as easily be used to create distinctions.

Govemment Involvement. Govemments, through the provision of facili- ties and services, also have an effect on reducing inequalities. Former federal and provincial policies in Canada tended to gear provision to the high-performance sector, and policies of gender and racial equity had a small impact on reducing inequalities. Government policies are begitming to shift toward the promotion of physical activity because of its perceived potential to reduce health care costs. As a result, the greater involvement in physical activity by, for example, the aging and the disabled is becoming apparent.

Educational Institutions. Educational institutions, through their physical education and sport programs, also have a great deal of potential for reducing inequalities. Unfortunately, such programs have far too often reflected the rather conservative ideologies of white, male, middle-cJass physical education teachers and coaches. Demands by women for a greater share of school resources; increas- itig multiculturalism in many parts of North America; and a new generation of teachers, many of whom are sensitized to issues of inequality in schools, may produce further changes.

Social Movements. Finally, various social movements are having a pow- erful impact on reducing inequality in sport, particularly the human rights move- ments (working through various nongovernmental organizations and action groups) that are dedicated to reducing inequalities in all sectors of social life—feminism, gay liberation, and movements by the aging and disabled.''^ Similarly, the sport for all, fitness, and health movements are having an effect on increasing access to sport and physical activity for many sectors of the population, and the envi- ronmental movement is having a sensitizing effect regarding forms of sport involvement.

Conclusion

Although sport has been an important agent in the production and reproduc- tion of social inequality, democratizing actions on the part of individuals and orga- nizations have sensitized us to sport's potential to be an equally important agent of social transformation for the production of social equality. Such transformations are usually thought of in individual terms conceming health and quality-of-life issues. We have also long held (though with little evidence) that sport participation has the capacity to transform the character of individuals, particularly in terms of self-confidence and attitudes. I have suggested previously, however (Donnelly, 1991, 1993), that the struggle to achieve more fully democratized forms of sport and leisure—that is, involvement both as a participant and in the determination of the forms, circumstances, and meanings of panicipation—^might result in the ca- pacity to transform communities. People could leam initiative, community en- deavor, collective rather than individual values, self determination, and so forth, that could permit them to begin to take charge of their own lives and communities. Thus sport may clearly be seen as contested terrain, providing fertile ground for ttie production and reproduction of social inequality, undergoing change (some- times progressive) as a result of the various processes and ageticies noted above, and having the potential to transfomi individuals and communities in ways that seriously reduce social inequalities.

238 DONNELLY

References

Beamish, R. (1990). The persistence of inequality: An analysis of participation patterns among Canada's high performance athletes. International Review for the Sociology ofSpon,25{2), 143-156.

Best, J . C , Blackhurst, M., & Makosky, L. (1992). Sport: The way ahead (Report of the Minister's task force on federal sport policy). Ottawa: Sport Canada.

Birrell, S. (1988). Discourses on the gender/sport relationship: From women in sport to gender relations. Exercise and Sport Sciences Reviews, 16, 439-502.

Binrell, S. (1989). Racial relations theories and sport: Suggestions for a more critical analy- sis. Sociology of Sport Journal, 6(3), 212-227.

Birrell, S. (1992). Women of color, critical autobiography, and sport. In M. Messner & D. Sabo (Eds.), Sport, men, and the gender order (pp. 185-3 99). Champaign, IL: Hu- man Kinetics.

Birrell, S.. & Cole, C. (1990). Double fault: Renee Richards and the construction and natu- ralization of difference. Sociology of Span Journal, 7( I), ] -21.

Birrell, S., & Richter, D. (1987). Is a diamond forever?: Feminist transformation of sport. Women's Studies International Forum, 10(4), 395-409.

Bourdieu, P. (1978). Sport and social class. Social Science Information, 17, 819-840. Bourdieu, P. (1984). Distinction: A social critique of the judgement of taste. New York:

Routledge & Kegan Paul. Bourdieu, P., & Passeron, J.C. (1977). Reproduction in education, society and culture. Lon-

don: Sage. Bragg, R (1995, March 17). Blue collars around the greens. {Toronto] Globe & Mail, p. A l l . Brohm. J.M. (1978). Sporr- A prison of measured time. (1. Fraser, Trans.). London: Ink

Links. Cantelon, H., & Gruneau, R. (Eds.). (1982). Sport, culture and the modem state. Toronto:

University of Toronto Press. Coakley, J. (1993). Sport and socijUization. Exercise and Sport Sciences Reviews, 21,169-200. Coakley. J. (1994). Sport in society: Issues and controversies (5tb ed.). Toronto: Mosby. Coakley, I., & White, A. (1992). Making decisions: Gender and sport participation among

British adolescents. Sociology of Sport Journal, 9(1), 20-35. Curry, T., & Jiobu, R (1984). Sports: A social perspective. Englewood Clitfs, NJ: Prentice

Hall. Curtis, J., & White, P. (1992). Toward a better understanding of the sport practices of

francophone and anglophone Canadians. Sociology of Sport Journal, 9(4), 403-422. Davis, K., & Moore, W. (1945). Some principles of siratification. American Sociological

Review, 10, 242-297. Davis, L. (1990). The articulation of difference: White (n-eoccupaiion with the question of

racially linted genetic diff(M"ences among afhl^es. Sociology of Spon Journal, 7(2), 179-187.

Davis, L., & Delano, L. (1992). Fixing the boundaries of physical gender: Side effects of anti-drug campaigns in athletics. Sociology of Sport Journal, 9(1), 1-19.

Dewar, A. (1991). Incorporation or resistance: Towards an analysis of women's responses to sexual opfffcssion in sport. International Review for the Sociology of Sport, 26( 1), 15-23.

Donnelly, P (1991)- Spon for all: Concerns for the future. In P Qja & R. Telama (Eds.), Sport for All (pp. 425-438). Amsterdam: Elsevier.

Donnelly, P. (1993). Democratization revisited: Seven theses on the democratization of sport and ac^ve leisure. Loisir et soci^ti/Society dE Leisure, 16(2), 413-434.

APPROACHES TO SOCIAL INEQUALITY 239

Drcdian. M. (1995, Mardi 25). The Briti^ fi^ a new class war. [Toronto] Globe & Mail, p. D4. Dunleavy, A., Miracle, A., & Rees, R. (! 982). Studies in the sociology of sport. Fort Worth:

Texas Christian University Press. Fejgin, N. (1994). Participation in high school competitive sports: A subversion of school

mission or contribution to academic goals. Sociology of Sport Journal, 11(3), 211 -230. Figler, S. (1981). Sport and play in American life. Philadelphia: Saunders. Foley, D. (1990). The great American football ritual: Reproducing race, class, and gender

inequality. Sociology of Sport Journal, 7(2), 111-135. Gadd. J. (1995. August 12), Where women stand now. [Toronto] Globe & Mail, p. A5. Gebauer. G. (1986). Festordnung und geschmacksdistinktionen: Die i!lusion der integra-

tion im freizheirspon. In G, Hortleder & G. Gebauer (Eds.), Sport-Eros-Tod (pp. 113-143). Frankfurt: Suhrkamp.

Goodhart, P., & Chataway, C. (1968). War without weapons. London: W.H. Al!en. Gnineau, R. (1975). Sport, social differentiation, and social inequality. In D. Bali &. J. Loy

(Eds.), Sport and social order: Contributions to the sociology of sport (pp. 117-184). Reading, MA: Addison-Wesley.

Gruneau, R. (1976). Class or mass: Notes on the democratization of Canadian amateur sport. In R. Gruneau & J. Albinson (Eds.), Canadian sport: Sociological perspec- tives {pp. 108-140). Toronto. Addison-Wesley.

Gruneau, R. (1983). Class, sports, and social development. Amherst: University of Massa- chusetts Press.

Hall, A., Slack, T , Smith, G., & Whitson, D. (I99I). Sport in Canadian society. Toronto: McClelland & Stewart.

Hargreaves. J. (1982). Sport, culture and ideology. In J. Hargreaves (Ed.), Sport, culture and ideology (pp. 30-61). London: Routledge & Kegan Paul.

Hargreaves, J. (1986). Sport, power and culture. Cambridge: Polity Press. Harvey, J.,&CanteJon, H. (Eds.). {]9S%). Not just a game: Essays in Canadian sport soci-

ology. Ottawa: University of Ottawa Press. Harvey. J.,&Houle,F, (1994). Sport, world economy, global culture, and new social move-

ments. Sociology of Sport Journal, 11(4), 337-355. Hoberman, J. (1992). Mortal engines: The science of performance and the dehumanization

of sports. New York: The Free Press. Hoch, P. (1972). Rip off the big game. Garden City, NY: Anchor. In^iam, A. (1980). "Sport and society": Interpenetration and resistance. Unpublished manuscripL Jngham, A. (1987. May 22-25). Some comments on theorizing social relations. Unpub-

hshed conference paper. North American Society for Sport History, Columbus, OH. Ingham, A. (1995). Personal communication. fngham,A..&Broom, E. (Eds.). (1982). Career patterns and career contingencies in sport.

Vancouver: University of British Columbia. Ingham, A.. &Loy, J, (1980). Heroes, sport and society: Issues of transformation and re-

production. Unpublished conference paper presented at the Intemational Sport Sym- posium, Los Angeles, CA.

Jamieson. K., Bailey, I., &Ross, D. (] 993, November). Influence of ethnicity and socioeco- nomic status on sport participation among Mexican heritage and Anglo-American females. Paper presented at the NASSS Conference, Ottawa.

Kidd. B. (1991). From quarantine to cure: The new phase in the stmggle againsi apartheid sport. Sociology of Sport Journal, 8, 33-46.

Laberge, S. (1989). L'^galiti dans le sport amateur au Canada. Paper presented at the 7th Canadian Olympic Academy.

340 DONNELLY

Laberge. S., & Girardin. Y. (1992). Questioning the inference of ethnic differences in achieve- ment values from type of spon participation: A commentary on White and Curtis. Sociology of Sport Joumal, 9(3), 295-306.

Lamprecht, M., & Stamm, H. (in press). Age and gender patterns of sport involvement among the Swiss labor force. Sociology of Sport Joumal.

Lenskyj,H. (1984). "Akindof precipitate waddle": Early opposition to women running, in N. Theberge & P Donnelly (Eds.), Sport and the sociocultural imagination (pp. 153-161). Fort Worth: Texas Christian University Press.

Lorenz, K. (1967). On aggression. New York: Harcourt, Brace & World. Loy. J.. McPherson, B.. & Kenyon, G. (1978). Sport and social systems. TorODto:

Addison-Wesley. Uischen, G. (1967). The interdependence of spon and culture. International Review of Sport

Sociology,!, 127-139. Liischen, G. (1969). Social stratification and social mobility among young sponsmen. InJ.

Loy & G. Kenyon (Eds.), Sport, culture and society (pp. 258-276). New York: Macmillan.

Majors, R. (1992). Cool pose: Bl^ik masculinity and spons. In M. Messner & D. Sabo(Eds.), Sport, men, and the gender order (pp. 109-114). Champaign. IL: Human Kinetics.

Mandell, R. (1990). Modem criticism of spon. In D. Kyle & G. Stark (Eds.), Essays on sport history and sport mythology (pp. 118-138). College Stmion: Texas A&M Uni- versity Press.

McAll, C. (1992). English/French Canadian differences in spon panicipation: Comment on White & Curtis. Sociology of Sport Journal, 9(3). 307-313.

Messner, M., & Sabo, D. (Eds.). (1992). Spon, men. and the gender order. Champaign, IL: Human Kinetics.

Novak. M. (1976). The joy of sports. New York: Basic Books. Page, C. (1973). Pervasive sociological themes in the study of sport. In J. Talamini & C.

Page (Eds.). Sport and sociery: An anthology (pp. 14-37). Boston: Little Brown. Phillips, J. (1993). Sociology of sport, Boston: Allyn & Bacon. Rees, R.. & Miracle. A. (Eds.). (1986). Sport and social theory, Chatnpaign, IL: Human

Kinetics. Renson, R. (1976). Social status symbolism of sport stratification, Riper presented at the

Olympic Scientific Congress. Qu6bec City. Rigauer. B. (1981). Sport and work (A. Gutanann. Trans.). New Ycat Columtta University Press. Sage, G. (Ed.). (1974). Sport and American society: Selected readings (2nded.). Reading.

MA: Addison-Wesley. Simpson, L. (1987). Values, respect and recognition; On race and culture in the

neoconservative debate. Praxis International,!, 164-173. StangI, J.. & Kane, M.J. (1991). Structural variables that offer explanatory power for the

underrepresentation of women coaches since Title IX: The case of homologous re- production. Sociology of Sport Joumal, 8(1). 47-60.

Stephens, T . & Craig. C L . (1990). The well-being of Canadians: Highlights of the 1988 Campbelis survey. Ottawa: Canadian Fitness and Lifestyle Research Institute.

Theberge, N. (1985). Toward a feminist altemative to spon as a male preserve. Quest, 37. 193-202.

Theberge, N.. & Donnelly. P. (1984). Sport and the sociological imagination. Fon Wonh: Texas Christian University Press.

Vedlitz, A. (1988). A question of values: Conservatives and the culture of poverty. Social Justice Research, 2, 235-248.

APPROACHES TO SOCIAL INEQUALITY 241

Vertinsky, P. (1995). Body history for spon historians: The case of gender and race. In K. Wamsley (Ed.), Method and methodology in sport and cultural history (pp. 50-61). Dubuque, IA: Brown & Benchmark.

Vinnai, G. (1973). Football mania (D. Triesman, Trans.). London: Orbach & Chambers. Whannei, G. (1983). Blowing the whistle: The politics of sport. London. Pluto. Wherry, G. (1896). Alpine notes and the climbing foot. Cambridge: Macmillan & Bowes. White, P., & Curtis, J. (1990a). English/French Canadian differences in types of spon par-

ticipation: Testing the school socialization expianatron. Sociology of Sport Journal,

7(3), 347-368. White, P.. & Cunis, J. (1990b). Participation in competitive sport among Francophones and

Anglophones in Canada: Testing competing hypotheses. International Review for the Sociology of Sport, 25(2), ! 25- ] 41.

Williams, R. (1977). Marxism and literature. Oxford: Oxford University Press. Williams, R. (1983). Keywords: A vocabulary of culture and society. London: Fontana. Willis, P. (1978). Pmfane culture. London: Routledge & Kegan Paul. Willis, P. (1982). Women in sport in ideology. In J. Hargreaves (Ed.), Sport, culture and

ideology (pp. 117-135). London: Routiedge & Kegan Paul. Yiannakis, A. (1976). A theory of spon stratification. Sport Sociology Bulletin, 4, 22-32.

Notes

'This view is characterized in the traditional icka that it is not wttelher you win or lose but bow you play the game.

'1 am well aware thai a large body of literature exists on the cultural reproduction of inequality, particularly with regard to the media and sport. However, limitations of space permit only passing mention of ihis important body of work. Also, the analysis in this paper is primarily concerned with Canada, although relevant data have been drawn from a number of non-Canadian sources. And, as a final qualification, [ am using the term sport in this paper in a generic sense. This is for convenience and space-saving purposes, but it has the problematic etfect of disguising an enonnous amount of variation in form and level of organization of sports; this variety has various distributional and relational implications. In general, the type of spon primarily involved in the production and reproduction of social inequality would be organized competitive sport.

'The theses are also reiterated, to some extent, in Coakley's (1994) consideration of func- lionalism, conflict theory, and critical theory {chapter 2).

''Coakley refers to this view as the "socialization-as-imemalization" approach (1993, p. 171).

'The levels of analysis were (kveloped in a preliminary way by Gruneau (1975) and Ingham (1980) who identified the distributive and relational levels. The categorical level emerged in discussions between Ingham and Gruneau (Ingham, 1995) and was first ffeveJopcd in a confer- ence paper by Ingham (1987). Dewar (1991) has provided one of the most succinct descriptions of these levels of analysis, and is cited here.

*'Loy (1995) s u g ^ s t s there may be a closer fit between the two typt^ogies than is pro- posed here, with a clear correspondence between the reflection thesis and categorical analysis, and between the resistance thesis and relational analysis. However, the temporal progression of the levels of analysis is less clear than the theses; though there is less research now using a categorical level of analysis, all three levels may be found in current research.

1̂ am well aware that a statement such as this accepts men and women (and various racial/ ethnic divisions) as distinct categories, a notion that is increasingly in question. However, for the purposes of the point being made (i.e., popular beliefs about spon), the statement stands.

342 DONNELLY

"•Such arguments suffer from an important statistical flaw in that within categories differ- ences (e.g., between the fastest and the slowest) tend to be significantly greater than the mean differences between categories.

•̂ Veninsky (1995) has documented the development of scientific/medical searches for dif- ferences between men and women. Blacks and Whites, and Jews and non-Jews. Also, Lenskyj (1984) provides an account of biomedical research purportedly documenting the frailty of (middle-class) women.

'"Whannel (1983) has noted how mediated sport is a powerful tool for establishing differ- ence in that commentators/journalists tend to employ stereotypical references to the national character/playing style of. in his case. non-British athletes.

"Both Hall et al. (1991) and Lamprecht and Stamm (in press) note that it is in recreational and low intensity exercise activities that male and female involvement reaches similar levels.

'-As Curtis and White (1992) note, however, no value judgement should be made abouj lower levels of involvement in competitive sport and higher levels of involvement in recreational spon. Such value judgements are sometimes made, though, in an attempt to explain differences in social and economic terms (i.e., lower levels of competitiveness).

"See also Wi!]i5 (J978, p. 196) for the use of a "cluster of methods" in relational analysis (participant observation, observation, just being around, group discussions, recorded discussions, informal interviews, use of existing surveys), h is possible to add content analysis of suhcuitural literature and questionnaires to this list.

'̂ For the longest time, courses, books, and articles about gender were usually about women. The Messner and Sabo (1992) collection placed the issue of gender firmly within relational analysis by including a number of papers focusing on men in relation to women.

'̂ This recognition is now becoming established in distributive analyses (e.g., Fejgin, 1994; Jamieson, Bailey, & Ross. 1993), in which there is a growing recognition of the intersections of gender, race, and class. TTus is having interesting results—for example. Fejgin {1994) reports no racial difference.s in high school panicipation when class is controlled. Note also a number of carefully controlled studies by CurtiL and White.

"'For example, references to "gender" concem historically specific and socially constructed definitions of what consiitutes masculinity and femininity.

'̂ Similar questions could be generated with regard to the specifics of inequity in any of the sets of social relations under consideration.

'*For a short period of time io South Africa there was an attempi by Whiles to democratize sport without democratizing the larger society. Such a situation creates impossible tensions and contradictions that were only resolved with the ending of apartheid.

"Unfortunateiy the movement for children's rights does not yei seem to have had an im- pact on sport

Acknowledgment

I would like to acknowledge the assistance of Alan Ingham, John Loy, and Philip White for the fruitful discussion of some of these ideas. 1 would also hke to thank the three anonymous reviewers for their comments.