Introduction

Participants in sports tend to very actively interact with other participants while playing sports or when they are not exercising. This is closely related to social capital that is formed by interaction among individuals. Attempt in explaining the effect of social capital through sports has been made in the study by Hanifan (1916) that the term, social capital, was firstly emerged. In her study that emphasized the role of the local community for school education in suburbs, the role of the baseball league has been given priority as a core element to develop local social capital. The coach of the baseball team said, "They (these older boys) stayed in school not only to the end of the baseball season; they got a taste of books and have been regular in attendance to the end of the year. Some who had not been in school for over two years won their Free School Diplomas this year and are planning to go to high school next year?" (Hanifan, 1916: 137). In other words, she has explained how passion for sports activities was led to an interest in academic studies among adolescents.

“To build bridging social capital requires that we transcend our social and political and professional identities to connect with people unlike ourselves. This is why team sports provide good venues for social capital creation.” (Putnam, 2000: 411). As pointed out by Putnam, sports organizations are the carriers of sound social capital. In addition, Putnam (2002) has explained that the issue of how American society became 'atomized' was related to an increase in population who enjoyed bowling and also decrease in bowling club in the 『Bowling alone. Interaction among members in official group is a core element of social capital (Blackshaw & Long, 2005). The number of bowling clubs, generalized in American society, has been decreased. However, they pointed out that the number of participants has been increased while symbolically suggesting the dissolution of social interaction.

Social capital includes topics that social behaviors of individuals or groups, density of relationship, quantity, a sense of belonging, and bond. Therefore, it acts as the most fundamental basis for social integration (Jenson, 1998; Wooley, 1998). It is recognized as an important element for solving social inequality, crisis, conflict, and other important areas and components of social cohesion (Côté, 2005).

Social capital is relevant to investment about social relationship that expects the profit in the market (Lin, 2001). Bourdieu suggested that social capital is "the aggregate of the actual or potential resources which are linked to possession of a durable network of more or less institutionalized relationships of mutual acquaintance or recognition" (Bourdieu, 1985: 248). It also consists of "social obligations" ("connections") and, under certain conditions, may be converted into economic capital and institutionalized in the form of the title of the nobility." (Bourdieu, 1985: 243). Coleman defined social capital as a function. In his view, social capital has two kinds of characteristics in common: "They all consist of some aspect of social structure, and they facilitate certain actions of individuals who are within the structure" (Coleman, 1990: 302). Another scholar, Putnam said the social capital is "social organization such as networks, norms, and social trust that facilitate coordination and cooperation for mutual benefit" (Putnam, 1995: 67). There are different opinions on the function/component/level of capital depending on scholars. However, social capital is 'a form of capital that occurs in the group sharing specific goals.'

Social capital occurs in various levels including individuals, groups, and nations (Coleman, 1988; Forrest & Kerarns, 2001; Onyx & Bullen, 2000). Efforts in measuring and utilizing social capital have been made in various fields. However, there has not been an agreement due to multi-lateral and multi-faceted characteristics of definition and scale of social capital to be interpreted and applied by scholars in each field. Debate over the concept of social capital has still been repetitively suggested as much as the usability and utilization of it. The reason why discussion of social capital has been remained as to why it existed was because the attempt was made to identify the essence of social capital in terms of logistics of traditional capital, in other words; finance, or physical/human/environment factors.

An approach made on the essence of multi-lateral and multi-faceted social capital by finding appropriate components of the concept instead of directly clarifying the concept might be much better. However, different components are applied depending on the direction of usability for social capital. Therefore, social capital has been measured in various fields. However, it is still required to define conceptual components and expressive factors in the context of sports clubs. In addition, it is required to develop scale indices that reflect them.

Discussion about social capital is classified as individual and group level. In other words, it is classified into 'profit' and 'return.' First of all, functional perspective of regarding the result of social capital as 'profit' is a 'good will' that is represented by Coleman and Putnam (Adler & Kwon, 2002). They focus on functional aspects of social capital. However, scholars such as Bourdieu or Lin who focused on 'inequality' related to closed feature and exclusiveness of social relationships suggested how social capital might provide a negative role depending on the types of interactions among participants and also the types of group.

On the other hand, sports policies from the government before and after the 2000s have been promoting physical activities of citizens based on 'sports for all'. Government of the people started in the IMF economic crisis in 1998 has reduced the size of sports group for 'small government' but proceeded decentralization and private initiative of sports. 5-year national sports plans proceeded from 1998 to 2002 have established the environment for people to participate in sports for all, expanded sports facilities as a place for residents in communities to participate in physical activities, extended the program for 'sports for all' for non-participating population, and also diffused privately-led 'sports for all' (Sport white paper, 2010). Such a trend of policies has led to participatory government from the government of the people expanding resident-friendly sports space, systematically cultivating sports clubs, and operating various programs for expanding the participation in physical activities to improve quality of lives of citizens through promotion in sports for all.

The government led by former president Myung-bak Lee from 2008 has established/promoted local sports clubs, expanded facilities of 'sports for all', improved usability of them, and proceeded the expansion for leisure sports and space to enhance conditions of participation in physical activities. They have expanded the participation of sports for the handicapped, minor groups. In addition, they have developed volunteer activities. The government led by current president, Geun-hye Park, in 2013 has been supporting the improvement of environment including outstanding sports clubs and sports facilities in schoolyard to accomplish the vision, 'Happiness for Citizens, New Era of Hope.' They have been distributing the customized programs in each life-cycle, providing authentication policy for sports, physical strength of the citizens, and establishment of comprehensive sports clubs (Sport white paper, 2014)

Hereupon, policy keynote of government for the support of sports for all has been consistently maintained in spite of replacement of regime. Ministry of Culture, Sports, and Tourism has been operating classes of sports for all in each cities and dos as well as various programs of sports for all to 'provide an opportunity of participating in sports for all.' In addition, they have been promoting national sport event committee to promote and cultivate club and holding and operating sports club league policies. A wide range of policy projects are in progress. Participations in sports club provides a positive opportunity on exchange of information, cooperation, competition, community spirit, formation of communitarianism, and relationship based on the promotion of cooperative relationship with sport activities.

The purpose of this study is to review the concept of social capital and to construct social capital scale in sports club (below SCSSC) context. In the sports club context, the meaning, structure, and dimensions of social capital were examined to develop theoretically sound, statistically valid, and reliable scale.

Methods

An indicator is a scale instrument that is to measure a certain phenomenon. Composite indicator provides a way to understand the complex phenomenon (Fetscherin & Stephano, 2016). It provides a simple number for a complex phenomenon. The process of scale development denotes a careful and meticulous methodology for arriving at a scale. A scale must confirm reliability and validity.

Following the instructions by Churchill (1979) and other scale-development studies (Fetscherin & Stephano, 2016; Ridfler, Diamantopoulos, & Siguaw, 2012; Rossiter 2002), I conducted the qualitative research and the quantitative researches. To expend previous research on social capital measurement and to develop the social capital scale in sports club members, the premise for using scale development for this study was that, if social capital does exist in sports club, then it ought to be definable and measurable. After discussing the process of selecting the scale development for this study, this chapter describes the methodology of scale development and evaluation, as demonstrated in Figure 1.

View original

Figure 1.

Scale development process

Study 1 carries out the scale generation and initial purification. Conducted literature review to set domain definition out, and then focus group interview for the initial item generation. Study 2 conducts the first scale refinement and purification with exploratory factor analysis. Confirm the content and face validity. Also, check the reliability and internal consistency of the indicators. Study 3 is the scale validation and application. Confirmatory factor analysis was conducted to confirm the dimensional structure of the sports club's social capital scale and to establish the effectiveness of convergent and discriminant validity. Higher-order factor analysis was conducted to confirm the number of second-order factor. And then, assess predictive validity.

Results

Factor analysis

After implementing the analysis, items with high error variance in regards to individual items and for simplicity purposes, items in relation to networks, reciprocal norms, social sorting which had low factor loadings were deleted sequentially. The error variance of the 4th item under social sorting (I am from a former district or school) and item 5 (I work or am engaged in a special workplace or occupation) reflected to be 1.417 and 1.519 respectively and were deleted accordingly. The total purified 19 items consisted of 3 items under trust, 5 items under network, 5 items under reciprocal norms, 3 items under social sorting, and 3 items under selective exclusion.

A confirmatory factor analysis was implemented on the purified 19 items and results overall satisfied the goodness-of-fit model. Specifically, the CFI, TLI, and RMSEA resulted in .939, .926, and .074(90% CI .065-.083) respectively, satisfying the goodness-of-fit (Hooper, Coughlan & R. Mullen, 2008). The SRMR value of .056 also satisfied the standard presented by Hu and Bentler (1999) of .08 or less. The Cronbach's α value of each factor all resulted in .8 or above, thereby securing reliability. In line with Kline (2011)’s proposal, the non-standardized estimate, standard error, and standardized estimate are presented in Table 7 and Figure 2.

View original

Table 7.

The results of CFA

Contents	Est.	S.E.	S.Est
Trust (Cronbach's α = .887)
1. I believe our club members have a better athletic performance than other clubs	1.000	-	.904	.879	.711
2. Our clubs have better athletic skills (level) than other clubs	1.001	.044	.913
3. I grant my club members' athletic performance	.735	.049	.744
Network (Cronbach's α = .909)
4. I often make informal gathering with club members who are close to me	1.000	-	.828	.868	.569
5. I participate in official club meetings	.952	.056	.812
6. I am actively involved in the operation of our club	.971	.056	.830
7. I am particularly close to one of our club members	.853	.056	.765
8. I often communicate with my club members	.898	.051	.850
Reciprocal norms (Cronbach's α = .863)
9. Our club's management meetings systematically resolve members' comments or problems	1.000	-	.720	.892	.624
10. Our club members are familiar with the rules of the club.	.944	.082	.682
11. Our club members respect each other	.972	.083	.735
12. Our club's rules are well-stocked	1.056	.084	.753
13. Our club strives for the well-being of its members	1.100	.083	.838
Social sorting (Cronbach's α = .867)
14. Among our club members, low-level(of Exercise level) members are passive in interpersonal relationships	1.000	-	.838	.808	.584
15. Our club does not accept people with a high competitive spirit as members	.964	.062	.801
16. Our club excludes members who lack sports skills from the game	1.057	.065	.845
Selective exclusion (Cronbach's α = .902)
17. I exercise often with members with similar inclinations	1.000	-	.846	.885	.719
18. I am friendly with members with similar inclinations	1.100	.051	.923
19. I usually workout with friendly members	1.031	.057	.845

*p<0.05, ***p<0.001

View original

Figure 2.

CFA diagram tru-Trust, net-Network, rec-Reciprocal norms, sor-Social sorting, exc-Selective exclusion

In order to test the convergent validity and discriminant validity of each factor, a correlation coefficient with the factor estimate, concept reliability, and AVE was calculated. Kline (2011) stated the following: i) convergent validity would be secured if the factor estimate was .7 or above and ii) discriminant validity would be secured if the covariance between the factors was .9 or below. The standardized estimate of the 2nd item of reciprocal norms was slightly below the criteria at .682. However, the covariance between the factors all satisfied the criteria. Although the 2nd item of reciprocal norms fell slightly below Kline’s criteria, it exceeded the criteria presented by Fornell and Lacker(1981) of C.R. at 0.7 and above and AVE at 0.5 and above.

Higher-order factor analysis

In order to verify the higher-order factor, one secondary factor and at least three primary factor is needed (Kim, 2016). Accordingly, the higher-order factor in this study was verified because the sub factors of social capital in sports clubs consisted of 5 primary factors. Furthermore, the secondary factor model was satisfied as there was hardly any difference in the model fit of the CFA and higher-order factor. This suggests that the 5 factors extracted through the exploratory factor analysis can be well explained by the social capital in sports clubs. The detailed goodness-of-fit index for comparison purposes is presented in Table 8.

View original

Table 8.

Model fit of CFA & HFA

	χ²	df	RMSEA (90% C.I.)	CFI	TLI	SRMR
Confirmatory factor analysis	385.1	142	.074 (.065-.083)	.939	.926	.056
Hierarchical factor analysis(1)	457.1	147	.082 (.074-.091)	.922	.909	.083	accepted
Hierarchical factor analysis(2)	457.0	146	.082 (.074-.091)	.922	.903	.083	rejected

After verifying the higher-order factor model, we checked to assess whether sports club social capital could be classified into positive factors of trust, network, reciprocal norms, and negative factors of social sorting and selective exclusion. The left of Figure 3 reflects the model with one secondary factor and two secondary factors. Results of measuring the higher-order model which had two secondary factors showed that there were no differences in goodness-of-fit index. However, the standard loading dose of emotional factors reflected to be .299 while the correlation coefficient between the above factors reflected to be 1.030. Also, compared to each estimation of χ², there is no significance difference (the threshold value is df 1△, χ² = 3.84). Accordingly, results showed that social capital of sports clubs was a single factor which could not be divided into positivity and negativity.

View original

Figure 3.

Higher-order factor diagram scsc- SCSSC, tru-Trust, net-Network, rec-Reciprocal norms, sor-Social sorting, exc-Selective exclusion

Assess predictive validity

After retesting the social capital scale items in sports clubs, satisfaction of life which is a dependent variable of social capital in various theories and empirical studies was selected as the dependent variable to verify the predictive validity (Evans & Kelly, 2004; Helliwell, 2003; Konow & Earley, 1999; Shin & Choi, 2010). Accordingly, a factor analysis and hypothesis testing was implemented. Diener et al. (1985)’s SWLS 5 items were used for the dependent variable, satisfaction of life. The items which were in English were translated through two people who were familiar with both English and Korea to secure reliability. One person translated English into Korean whereas the other person translated Korean into English. The two translations were compared to the original items.

Results of the hypothesis testing for the predictive validity analysis showed goodness-of-fit at satisfactory levels as per Table 9. Furthermore, hypothesis testing showed i) there was significant influence on satisfaction of life of social capital in sports clubs with path coefficient at .579; and ii) it was able to explain 33% of the variant (Table 10, Figure 4). Such results suggest that there is a higher possibility that satisfaction of life of the participants will increase if the social capital in sports is high. Accordingly, it can be determined that the 5 composition factors related to social capital in sports clubs and scale items can sufficiently explain the parent variable of social capital in sports clubs.

View original

Table 9.

Model fit of predictive validity

	χ²	df	RMSEA (90% C.I.)	CFI	TLI	SRMR
Predictive model	607.9	246	.069 (.062-.075)	.929	.921	.075

View original

Table 10.

Predictive validity

	b	se	β	t	p
SCSSC ⟶ SWLS	.721	.091	.579	7.910	.000

View original

Figure 4.

Predictive validity diagram

Discussion

Like other social science theories, social capital is also abstract and includes various levels and sub factors which makes scale difficult. Paradoxically, the difficulty of scale leads to objective scale requirements. Álvarez and Romaní (2017) analyzed preceding researches related to measurement of social capital and proposed mechanism and measurement approaches according to the intended level of measuring social capital. The studies of Lochner, Kawachi, and Kennedy (1999) propose collective efficacy, psychological sense of community, neighbourhood cohesion and community competence as subordinate variables, based on social capital theory of Putnam, making a suggestion of a method of uniting the individual-level response to the community-level response. Fandiño, Marques, Menezes, and Bentes's (2015) study, which developed a social capital measure for Portuguese workers, consists of three dimensions (cognition, structure, and relationship) of Nahapiet and Ghosal. Takakura, Hamabata, Ueji, and Kurihara (2014) developed social capital scale for high school student.

The study of social capital scale in sports in Korea is Kim's (2006) research. He developed a social capital scale based on sports participants and called it SCSPS (social capital scale in participant sports). This research is significant in that it was one of the first papers to reflect the development of social capital scale on sports participants which had not been implemented in the sports participation. Despite of the distribution, his research contains the following limitations: First, the subject is unclear. The subject of study in his research is all sports participants. Social capital is established by the people who are affiliated in or form a group. Accordingly, the target should be subjects which can reflect the characteristics of a group. Second, it borrows the scale used in preceding national/international studies during the process of selecting initial items. The preceding studies on scale development propose that open-ended interviews be conducted to confirm the language and experience used by the targets subject to scale (Netemeyer et al., 1995; Rossiter, 2002). However, his research verifies and measures only the content validity of the scale of preceding studies. Third involves problems associated with the estimates due to the lack of cases (only 98 data were used for factor analysis). At least 150 cases are needed to implement the confirmatory factor analysis (Anderson & Gerbing, 1988; Holbert & Stephenson, 2002). Furthermore, studies by Chou and Bentler (2002) state that at least 200 data need to be secured to obtain accurate estimates. Fourth, the EFA and CFA were verified using a single research data. It is a factor analysis which needs to include more theoretical limitations as the same data was used to compare the CFA and EFA. In other words, it’s a process which verifies the scale based on a theoretical model. The scale developer must establish a theoretical model through the data collected from the EFA and secure CFA data in line with that model.

The differences between this research and Kim’s (2006) research are as follows: i) this study restricted the subject of scale development in a physical and spatial environment called sports clubs; and ii) the language and experience of the measuring subject was presented through in-depth interviews and target group interviews; and iii) from each of the EFA and CFA, data from at least 300 people was obtained and verified; and iv) statistically appropriate items were deleted to present results.

Conclusions

This research was aimed at developing social capital scale in sports club which reflects the characteristics of Korean sports clubs. This study was conducted by literature research, qualitative research, and quantitative research for the systematic scale development procedure. Specific development procedures and results are as follows.

First of all, seven factors constitute social capital of sports club as follows; trust, network, reciprocity, norms, sense of community, social sorting, and selective exclusion. Results derived from the literature review were employed to configure the initial coding framework of content analysis. The results of the qualitative research were used to select and refine the first item.

The exploratory factor analysis was performed on 7 factors with 78 items. In order to determine the number of factors, a scree plot, a parallel analysis, and a 3 to 5 factor model were compared and a 5 factor model was selected. They were named as trust, networks, reciprocity norms, social sorting, and selective exclusion. It was purified with a total of 35 questions.

Confirmatory factor analysis, high-order factor analysis, and predictive validity were conducted to confirm and apply SCSSC. As a result of confirmatory factor analysis, the final 19 items were selected, and a high - order factor analysis revealed one secondary factor structure. The result of predictive validity, the scale developed in this study proved to be statistically significant with SWLS.

The objective and systematic scale deduced from this research hold important significance in that is was able to overcome the limitations in existing scales and is applicable to sports club participants. The significance of this scale can be explained in more detail as the following.

First, methodological contribution is the development of scale from the perspective of sports club members. Renowned social capital theorists such as Hanifan (1916), Putnam (2000), and Broudieu (2011) state the importance of sports clubs in the formation and development of social capital. Such are reflected by the active studies which have been conducted on social capital in sports clubs, both nationally and internationally. Despite such importance, discovering a social capital scale specific to sports clubs is not easy.

The majority of social capital studies on sports clubs have been implemented based on qualitative research methods, and such qualitative studies hold great significance in being able to verify the usefulness of social capital formed in sports clubs and its diverse sub factors (Kim et al., 2012). However, compared to the number of qualitative studies on social capital in sports clubs, quantitative research is relatively difficult to find. From a sports club perspective, studies which have applied social capital as the cause variable or dependent variable have omitted to perform objective validity test procedures on tools and have therefore borrowed tools in other fields (Park, Seo, & Chae, 2014; Hong, Won, & Ko, 2011).

Social capital is a multi-level, multidimensional concept in which connections with the research field must be taken into consideration. (Hong et al., 2007; Hong, 2007). From a sports club perspective, data related to sports participation or sports level is a unique domain of social capital in sports clubs (Tonts, 2005). In this regards, through interviews with people who are participating in or have experience in operating sports clubs, this study developed items based on the language directly used by these participants. The scale developed in this research from the perspective of sports club participants will provide basic data in evaluating social capital of sports clubs.

Second, theoretically, this study attempted at a balanced scale of social capital by presenting social sorting and selective exclusion based on the exclusivity of social capital as scale factors. As a multiple concept, social capital incorporates both positive as well as negative factors. The existing social capital scales were measured by focusing on the productivity or effectiveness of social capital (Narayan & Cassidy, 2001; Onyx & Bullen, 2000). The contribution of these studies is of great importance from the perspective that they have presented social capital as the main domain of community research. However, social capital theorists who put Portes (1998) or Lin (2001), etc., as the leading figures state exclusivity, limitation of freedom following closed relationships and excessive requests regarding members as adverse effects of social capital. Woolcock (1999) classified social capital in groups into bonding, bridging, and linking based on the level of closure of the groups and also stated the necessity for management of the adverse effects (Woolcock & Narayan, 2000).

This study applied a cognitive and structural perspective on the exclusive factors arising in sports clubs. The cognitive perspective refers to the mutually exclusive factor which arises between members within sports clubs whereas structural perspective refers to the socioeconomic factors which act as barriers to entry in the selection of sports clubs. Accordingly, the study holds the second significance in that by exploring both positive and negative factors which have reflected the characteristics of sports clubs, it was able to establish a more balanced scale.

The practical contribution is as follow. It provides a tool to policy makers in the local government in determining sports clubs which provide high contribution to the local society. In relation to the public sports club business initiated from 2006, the government has been proceeding business which unite and develop member clubs in local governments. Its attempt was to enable public sports clubs to perform a central role in the local society as a method to modernize sports for all (Ministry of Culture, Sports, and Tourism, 2014). Such requires the input of massive financial, physical, and human resources. Verifying the developmental directions and effectiveness of the public sports club under such circumstances is an essential stage in determining the form and range of resource input. From this perspective, the scale of social capital in sports clubs will contribute to enhancing the effectiveness of resource input.