Studying the Roles of Nonprofits,Government,and Business in Providing Activities and Services to Youth in the Phoenix Metropolitan Area

Abstract

The article addresses the questions, What do children in urban areas do on Saturdays? What types of organizational resources do they have access to? Does this vary by social class? Using diary data on children’s activities on Saturdays in the Phoenix-Mesa-Scottsdale metropolitan area, the authors describe the different types of venues (households, businesses, public space, associations, charities, congregations, and government/tribal agencies) that served different types of children. They find that the likelihood of using a charity or business rather than a government or tribal provider increased with family income. Also, the likelihood of using a congregation or a government facility rather than a business, charity, or household increased with being Hispanic. The authors discuss the implications for the urban division of labor on Saturdays and offer research questions that need further investigation.

Keywords

urban communities stratification organizations consumption quality of life

This article examines what children in a metropolitan community do on Saturdays and which venues they use for these activities. The goal is to describe the interface between the household and organizational sectors, by examining the affiliation of different types of households and children with different types of venues. The research builds on the work of McPherson (1983); Lareau (2002, 2003); Marwell (2007); Small (2009); Allard, Tolman, and Rosen (2003); and Watkins-Hayes (2011), who examined urban residents’ ties to organizations and the impacts of these ties on their lives. There is a considerable literature on what children do in their leisure time (e.g., how much they play or watch sports or shop or engage in cultural activities); however, we need to understand better why and how families utilize different venues or providers that provide these activities in their communities. Experiences vary in quality by venue. Kids can shoot hoops at the park or they can play in a Catholic Youth Organization (CYO) basketball league. Because not everyone has access to the same experiences and venues, not everyone will have the same life chances later on. To even the playing field, we need to understand the barriers and opportunities that different families face to access the organizations that provide valuable activities for their children.

After reviewing the literature on what children do in their leisure time, we review the organizational literature on the different goals and incentives of businesses, nonprofits, and government service providers. We then offer a theory that says that families will select venues based on their preferences and what they can afford. At the same time, organizational establishments are following their own strategies, striving to achieve organizational objectives. We assert that the end result, to borrow from McPherson (1983), is an ecology of affiliation where different segments of the community patronize different types of establishments.

This article contributes to public policy as well as the urban ecology and stratification literatures. It offers an explanation for why different population subgroups should connect to different types of organizational venues based on organizational incentives and families’ resources. The results show that interventions must operate at two levels—the household and the organizational. Simply empowering households with vouchers and information may not be enough to enable them to access providers that others use. One must also incentivize providers to serve families that are marginalized because of income, race, or ethnicity. Realigning families’ affiliations with organizations and venues is a complicated task.

The data for the article came from a random-digit-dial phone survey of 1,036 households with children between the ages of 5 and 12 in the Phoenix-Mesa-Scottsdale, Arizona metropolitan area. The Institute for Social Science Research at Arizona State University collected the data in fall 2003 and winter 2004 (excluding holiday weekends). They asked for diary data on children’s activities on the previous Saturday. In the course of reporting diary data on how their children spent their time, respondents identified the establishments that provided them with different kinds of services and activities and the auspices of these establishments.

What Children Do in Their Leisure Time?

There is a great deal of research on children’s activities outside the home after school and on weekends. Hofferth and Sandberg (2001), looking at children’s time diaries for 1997, found that children between 6 and 8 and between 9 and 12 years of age spent the bulk of their time sleeping and at school. In their free time, they mostly watched television (12.8 and 13.6 hours per week, respectively), played (11.9 and 8.8 hours, respectively), engaged in sports (5.3 and 6.4 hours, respectively), and visited friends and relatives (3.5 and 3.5 hours, respectively). Looking at time diary data for 2002, Bianchi, Robinson, and Milkie (2006) found that children between 5 and 11 years of age had, on average, 38.3 hours of “free time” a week. Again, most time was spent watching television (13.9 hours), playing (10.5 hours), visiting friends and family (2.3 hours), and playing sports (2.2 hours).

Research has also found that different categories of children do different things. Boys were more involved in group sports (Timmer, Eccles, and O’Brien 1985), while girls were more likely to take dance or cooking classes (Medrich et al. 1982). Younger children spent more time going to church than older children (Timmer, Eccles, and O’Brien 1985; Hofferth and Sandberg 2001), and whites spent more time in sports than minorities (Medrich et al. 1982; Hofferth and Sandberg 2001). Family structure also seemed to affect what children did in their leisure time. Bianchi, Robinson, and Milkie (2006) found that children between 5 and 18 from single-mother households spent less time on home computer activities, religion, sports, art, and reading and more time watching television than children from two-parent homes. Hofferth and Sandberg (2001) found that children from working female–headed families tend to spend less time in church, less time in play, and more time at daycare centers than children from a male breadwinner/female homemaker family.

The literature has also shown that social class is important in explaining what children do in their leisure time. Lareau’s (2002, 2003) in-depth study of fifteen U.S. families emphasized the importance of social class rather than other factors, such as race or ethnicity, in explaining families’ access to institutional resources. Drawing on Bourdieu (1977), Lareau argued that those with more financial capital could pay for better schools, ballet classes, movies, tutors, and coaching. She also found that those with cultural capital, which often accompanies financial capital, were better able to know what will and will not benefit their children. Cultural capital also enabled families to negotiate more effectively with those who administered and provided services. These cultural frames were rooted in the lived experiences of those who were upper, middle, and lower class and influenced the tactics or strategies of parents as they sought to find worthwhile experiences for their children (see Galaskiewicz et al. 2012).

At the aggregate level, research has shown that race and class affect who has access to amenities. Small and McDermott (2006) found that race matters more than class in explaining where organizations are located. Poorer areas had more hardware stores, grocery stores, convenience stores, pharmacies, banks, credit unions, restaurants, and laundries than less poor areas; however, as the proportion black increased there were fewer of these facilities (although more childcare establishments). The poverty effects seemed to be contingent on cities’ economic conditions and location, being stronger in the South and West.

Research on “healthy communities” has found that recreational facilities are not as available to poor and minority children as they are to middle-class and white children, although parks are (L. Moore et al. 2008; Sister 2007). Research also found that not all residents have access to healthy food. Moore and Diez Roux (2006) found that, in the three states they studied, grocery stores were more common in poorer areas and in minority and mixed areas than in white communities, but wealthier and white communities had many more supermarkets (see also Morland et al. 2002; Powell et al. 2007). Other research looked at the location of full service and fast food restaurants and found that the latter clustered around schools (Austin et al. 2005; Kipke et al. 2006); however, research linking socioeconomic characteristics of neighborhoods and restaurant types has been mixed (see, for example, Powell, Chaloupka, Bao 2007).

Numerous researchers have examined the consequences of participating in different leisure-time activities. While leisure time may not be as important as time spent at home or in school in explaining future gains or losses, we suspect that this “free time” exacerbates whatever advantages or disadvantages a child realizes from home and school (see McFarland and Thomas 2006). In this respect, we agree with Medrich et al. (1982) that out-of-school/out-of-home life involves far more than leisure. Studies have found that after-school activities increased academic achievement (Valentine et al. 2002; Marsh and Kleitman 2002; Cooper et al. 1999), reinforced parental ideologies (Dunn, Kinney, and Hofferth 2003), and contributed to less delinquency in minority schools (Hoffmann and Xu 2002). Extracurricular activities have also been shown to help with adolescent social integration, which is a strong predictor of adult social integration (Spady 1970; Otto 1976), and to increase future civic involvement (Hanks and Eckland 1978; Verba, Schlozman, and Brady 1995; McFarland and Thomas 2006). Also, the relationships formed with the staff of after-school programs promoted positive youth development (Kahne and Bailey 1999).

The Three-Sector Society

While inequities among population subgroups are apparent, there has been little research on why they persist. In addition to family preferences and resources, another possible explanation is that the incentives that govern organizational decision-making drive providers to serve some population segments more than others (Weisbrod 1998). In other words, the auspices of organizations could affect which segments of the community are served. Unfortunately, only a few researchers have looked at this (see Medrich et al. 1982; Schlesinger and Dorwart 1984; Olfson and Mechanic 1996; Blustein and Hoy 2000; Kushman 1979), and the studies have been highly descriptive.¹

Building on Hansmann’s (1996) seminal work, our analysis distinguishes between for-profit enterprises, nonprofit organizations, and governmental organizations. Hansmann (1996) argued that the most important difference among these types of organizations is the nondistribution constraint (NDC). The NDC means that nonprofit and governmental organizations are prohibited from distributing residual earnings to private persons, while for-profits are legally obligated to earn reasonable returns for owners (Kahn 1997). This builds in an incentive for for-profits to maximize revenues and minimize costs, which suggests that for-profits are likely more efficient than their nonprofit and government counterparts (Weisbrod 1988). It also follows that for-profits are more likely to charge premium prices, externalize costs, skimp on quality, and exploit workers as long as these practices contribute to the bottom line. Social costs (or benefits) are not their concern (Weisbrod 1988).

Government organizations and agencies are bound by the NDC, mandated to be public regarding and sensitive to equity issues, and expected to be nondiscriminatory. Legislative, executive, and judicial bodies prescribe public policies that these organizations are obligated to act upon. Legal mandates dictate what public organizations can do, the methods they can use, the prices they can charge, and whom they should serve (Wamsley and Zald 1973; Bozeman 1987; Rainey 1997). Because they are not autonomous, public sector managers are preoccupied with enforcing and following rules, not with making choices. Although there have been calls for more managerial discretion (M. Moore 1995), most public sector managers have less autonomy and authority than their private sector counterparts. At the same time, public sector managers must be aware of the political implications of their actions, public scrutiny, and their reliance on governmental appropriations (Rainey 1997, 73–74). In contrast to for-profits, which are institutionally obligated to serve private interests, government organizations are institutionally obligated to serve the public interest.

In the United States, all nonprofit organizations are bound by the NDC, and they enjoy tax-exempt status; however, there are important distinctions among them. Charitable organizations are legally bound to pursue a public purpose, while many associations are not. Associative organizations are bound by the NDC, but the organization is tightly controlled by its membership. These organizations are characterized by their exclusivity and the fellowship benefits they provide to members (Hansmann 1996). Obvious examples of this type of organization are country clubs and social clubs, but labor unions, homeowners associations, business leagues, professional associations, mutual insurance companies, cooperatives, credit unions, and veterans’ organizations are also nonprofit organizations that primarily benefit members. Social welfare organizations are also membership organizations, but typically they are oriented toward some public purpose rather than the benefit of a select group of individuals. Many are involved in advocacy and lobbying. Examples include the Sierra Club, the National Rifle Association, and the American Association of Retired Persons, and these often are involved in political advocacy.

Charitable organizations are distinct from other types of organizations, because they are allowed to receive tax-deductible contributions from individuals and corporations. The two most common types of charitable organizations are private foundations and public charities. The former typically funds activities from an estate or company; the latter typically provides services and relies on public support.² Charitable status imposes a number of constraints on these organizations in addition to those that the NDC imposes. First, they must be organized exclusively for a charitable purpose as described in IRS Section 501(c)(3). The purpose can be “charitable, religious, educational, scientific, literary, testing for public safety, fostering national or international amateur sports competition, and preventing of cruelty to children or animals” (IRS 2012a). According to the IRS (2012a), “The term charitable is used in its generally accepted legal sense and includes relief of the poor, the distressed, or the underprivileged; advancement of religion; advancement of education or science; erecting or maintaining public buildings, monuments, or works; lessening the burdens of government; lessening neighborhood tensions; eliminating prejudice and discrimination; defending human and civil rights secured by law; and combating community deterioration and juvenile delinquency.”

Second, “A section 501(c)(3) organization must not be organized or operated for the benefit of private interests, such as the creator or the creator’s family, shareholders of the organization, other designated individuals, or persons controlled directly or indirectly by such private interests. No part of the net earnings of a section 501(c)(3) organization may inure to the benefit of any private shareholder or individual. A private shareholder or individual is a person having a personal and private interest in the activities of the organization” (IRS 2012c). The deduction requires that the revenues used by the charity do not benefit the donor directly but produce goods and services that yield external benefits to the larger society. The deduction is both a subsidy for a socially valuable good and recognition that gifts have a public, not private, purpose. Thus, there is an expectation that donor-supported charities will be public, not private, regarding. This distinguishes them from associative organizations.³

Third, “Section 501(c)(3) organizations are restricted in how much political and legislative (lobbying) activities they may conduct” (IRS 2012b). They may not attempt to influence legislation as a substantial part of their activities (i.e., lobby) and they may not participate in campaign activity for or against political candidates. Nonetheless, foundations and public charities can conduct educational campaigns, sponsor get-out-the-vote efforts, meet with public officials and legislators, and testify before executive and legislative bodies. Of all the constraints on charitable organizations, the restriction on their political and legislative activities is the most nebulous.

Despite their label and the public benefit requirement, charitable organizations have no legal obligation to be charitable, that is, provide relief to the poor, distressed, or disadvantaged (Simon, Dale, and Chisolm 2012). IRS rulings and court decisions have suggested that redistribution is a criterion for charity status in certain circumstances; however, it is difficult to pin down a specific legal requirement for redistribution.⁴

Are charities redistributive in practice? Wolpert (1993, 1996) found that very little money raised by nonprofits in cities and suburbs crosses city/suburban borders and that the bulk of nonprofit organizations’ services were amenities, not redistributive goods. Clotfelter (1992) showed that only a small number of nonprofits serve the poor as primary clientele. Schlesinger, Mitchell, and Gray (2003) found that nonprofit HMOs were somewhat more likely than for-profits to open access to services, give money to local community organizations, and make facilities available to charities. Schlesinger and Dorwart (1984) and Olfson and Mechanic (1996) found that it is government-run facilities (rather than nonprofit or for-profits) that are more open to indigent or nonpaying mentally ill patients. In contrast, Blustein and Hoy (2000) found that enrollees in nonprofit Medicare health plans tended to be of a higher socioeconomic status, while those enrolled in for-profits tended to be of a lower status. Finally, Kushman (1979) studied daycare centers in North Carolina and found a three-tier system: government provides services to poor, one-parent families; for-profits serve families with two parents working; and nonprofits are in between.

Figure 1 provides a summary of our discussion. Government organizations are incentivized to ensure social welfare. They are strictly accountable, forbidden to distribute earnings, and must have a public purpose. In addition, they are expected to be redistributive and nondiscriminatory, and they are bound by legal/political mandates and rules. For-profits are incentivized to serve primarily the private interests of owners and have only tax-reporting requirements. Nonprofits are more complicated. Charities, like government organizations, are incentivized to provide public benefit, while associative nonprofits are not and serve mostly members’ needs.

A Theory of Organizational Affiliations

To explain why children and parents use one type of establishment or another for their Saturday activities, let us start with a simple choice model. Marketing practice assumes that consumers, as they evaluate the products of different providers, calculate the costs and benefits that they might realize. There are functional and emotional benefits and monetary, time, energy, and psychic costs. The consumer examines the value ratios for two competitors, V₁ = Benefits₁ / Costs₁ and V₂ = Benefits₂ / Costs₂. If the ratio V₁ / V₂ > 1, then the consumer will select provider one; if the ratio V₁ / V₂ < 1, then the consumer will select provider two; if the ratio V₁ / V₂ = 1, the consumer will be indifferent and numerous other factors may then influence the decision.

Assuming that families know what benefits their child the most and they have good information on all providers, the key variable in their decision is the family’s financial resources.⁵ Obviously, families that have more financial capital have access to a broader array of options to choose from, that is, since they can absorb more costs, they can base their choice more on benefits rather than on cost. In contrast, families with less financial capital have limited choices. As family income decreases, costs become more salient. This, in turn, shrinks the choice set.

If businesses, which in our case are mostly stores, recreational facilities, and restaurants, seek out the middle to high end of the market, middle- and upper-income families will be able to afford their services. Thus, very low-income families are less likely to spend their leisure time patronizing businesses and will be drawn to venues that they can afford. Government providers are the obvious providers of choice. As described above, because they are public regarding, barriers to consumption should be minimal.⁶ However, because government services need to be accessible to all, quality may suffer, for example, classes may be larger, services are standardized, equipment may be more worn, and so on. Thus a reasonable expectation is that poorer families will patronize government agencies (because they can afford them) rather than businesses (which they cannot), while wealthier families will patronize businesses (because they can afford them) rather than government agencies (which provide an inferior experience). This hypothesis is consistent with the research by Schlesinger and Dorwart (1984), Olfson and Mechanic (1996), and Kushman (1979), but is not consistent with Blustein and Hoy (2000).

It is more difficult to make a prediction about who nonprofits serve. On one hand, associative nonprofits, which in our case mean neighborhood associations, are more like for-profits and may serve members primarily. Because they benefit members and rely on dues, they are more sustainable if they recruit middle- and upper-income families. In this respect, they are similar to businesses, even though they provide collective goods, for example, an attractive landscape and security, for the neighbors. Given the importance of fees, we expect that associations are not as common among the poor.

Charitable nonprofits are more like government organizations. In our case, we looked at mostly sports clubs/leagues, Ys, Boys’ and Girls’ clubs, congregations, museums/theatres/zoos, and scouting. Because of the legal requirement that they have a public purpose, they may choose to fulfill this mandate by being redistributive and affordable for the poor, similar to government organizations. However, unlike government organizations, they also have to support themselves with fees, contracts, gifts, and grants (Frumkin and Galaskiewicz 2004). One option to support themselves in this way is to provide services at premium prices to middle- and upper-income households and then use the profits to provide services to lower income families at reduced rates or for free. In other words, the surplus they raise from commercially successful ventures will be used to cross-subsidize less profitable, but socially beneficial, enterprises (James 1986). This means that charities serve everyone, but unlike government organization, affiliations with charities, in the aggregate, should be unrelated to family income. Of course, our expectations are premised on the assumption that there is redistribution either among or within charities (e.g., wealthier Y branches subsidize Ys in poorer areas, and wealthier parishes subsidize poorer parishes).

Data and Methods

To answer the research question and test the relationship between children’s class background and their activities and their providers, we examined survey data that contain unique information on the activities of a random sample of children. We subcontracted the data collection to the Institute for Social Science Research at Arizona State University. They conducted a phone survey of 1,036 parents, guardians, and caretakers of children ages 5 to 12 in the Phoenix-Mesa-Scottsdale metropolitan area. The phone survey asked for a diary of a child’s activities on the previous Saturday (see, for example, Bianchi and Robinson 1997; Harvey 1999; Robinson 1999).⁷ The survey was conducted between September 2003 and February 2004 using random-digit dialing of landlines. Each of the 27,788 dialed numbers received at least ten calls on varying days of the week and times of the day and up to twenty calls if an answering machine reply gave a reason to believe that it was a residential number and not a business number. The cooperation rate was 55 percent (Edwards 2004).⁸

Upon reaching a residential household, interviewers continued with the survey only if at least one child between 5 and 12 years of age had lived in the household for at least five days a week. Only 6.8 percent of all reached numbers met this criterion, but the estimated response rate among eligible households (i.e., those with children of the given age range) was 22.7 percent.⁹ Because a significant proportion of the population in the Phoenix-Mesa-Scottsdale metropolitan area is Spanish-speaking, we translated the survey, and Spanish-language interviews accounted for about 19.4 percent of all 1,036 completed surveys. No data collection took place during the weeks of Thanksgiving, Christmas, and New Years. If there were two or more children 5 to 12, interviewers requested a list of children and randomly selected one of them. In the next step, interviewers asked to speak to an adult with the best knowledge of what the selected child did on the previous Saturday. In 93 percent of the interviews, the child’s parent gave the answers; for the remainder, the answers came from legal guardians, grandparents, aunts/uncles, or older siblings.

Once the adult with knowledge of the child’s activities was on the phone, interviewers asked for a diary of the selected child’s activities for the time between midnight on Friday and midnight on Saturday of the previous week.¹⁰ Of course, a child could participate in more than one activity outside the home. Figure 2 shows that 840 households (out of the 1,036) said that their child had activities outside the home; 196 said they did not. We then asked if the activity benefited the child, the child was simply accompanying the adult, or the activity benefited both the child and the adult.¹¹ Of the 840 households that had activities outside the home, 658 responded that the activity either benefited the child or benefited both the child and the adult.¹²

Altogether there were 1,256 activities that were outside the home and that benefited the child or the child and the adult. Table 1 presents the range of activities in which the children participated outside the home and the proportion of times each activity was mentioned. The activities mentioned most were socializing, shopping, and personal care (e.g., doctors’ visits, haircuts, etc.), eating, playing, and team sports.

Table 1

Descriptive Statistics: Types of Activities, Venues, and Children/Household Variables

	Nonmissing Observations	Mean	SD	Min.	Max.
Types of activities (N = 1,256)
Socializing (e.g., visiting, party, sleepovers)	1,253	.169	.375	0	1
Shopping/personal care (e.g., doctors’ visits, haircuts)	1,253	.143	.350	0	1
Eating	1,253	.133	.340	0	1
Play	1,253	.125	.331	0	1
Team sports (e.g., soccer, baseball)	1,253	.123	.328	0	1
Individual sports/physical activity (e.g., hiking, biking, go-karting)	1,253	.109	.312	0	1
Spectator events (e.g., movies, festivals, sporting events)	1,253	.070	.256	0	1
Miscellaneous (e.g., animal care, travel, helping parent at work)	1,253	.053	.223	0	1
Educational (e.g., tutoring, museum trips)	1,253	.021	.143	0	1
Religious (e.g., church attendance, religious education)	1,253	.021	.143	0	1
Dual sports (e.g., tennis, boxing)	1,253	.011	.105	0	1
Arts/performances	1,253	.010	.101	0	1
Hobbies/games	1,253	.006	.080	0	1
Cheerleading	1,253	.005	.069	0	1
Types of venues (N = 1,256)
Business	1,234	.369	.483	0	1
Household (other than respondent’s)	1,234	.233	.423	0	1
Government/tribe	1,234	.159	.366	0	1
Charity	1,234	.113	.316	0	1
Public spaces (e.g., street, desert, mall)	1,234	.050	.219	0	1
Out-of-town provider	1,234	.036	.188	0	1
Congregation	1,234	.032	.177	0	1
Neighborhood association	1,234	.008	.090	0	1
Child variables (N = 1,036)
Child female	1,032	.464	.499	0	1
Child’s age	1,036	8.507	2.285	5	12
Child Hispanic	1,018	.392	.488	0	1
Child non-Hispanic, non-white	1,018	.095	.294	0	1
Household variables (N = 1,036)
Number of children	1,036	1.689	.873	1	7
Years living in Phoenix metro area	1,035	15.694	12.414	0	75
Caregiver not married	1,028	.193	.395	0	1
Family income in $1,000s	1,036	64.570	49.233	−.292	250
Caregiver worked last Saturday	1,017	.165	.372	0	1

NOTE: The figures for family income were computed after we imputed values where there were missing data.

The interviewer then asked whether the activity was organized by an organization.¹³ If the activity was, we asked if the provider was a business, a voluntary association, a church or congregation, a nonprofit, a government agency, a school, or something else and for the exact location. We then asked several questions about the provider and the parent/guardian/child’s satisfaction with the provider (see Galaskiewicz et al. 2012). If the activity was at a household, we asked questions about who was there and the relationship between this household and the respondent/child’s household. We did not get the other household’s address. If the activity was at “some other location,” we asked for the cross streets or address of the venue. Often respondents mentioned public parks or “the street.”

Two issues arose. First, many of the organizational providers (e.g., a youth soccer league, a corporate sponsor, a girl scout troop) were not located or headquartered at the site where the service was delivered (e.g., a city park, a parade route, a history museum). This led us to verify the names, addresses, and auspices of the providers that were mentioned and to make sure that we had both the name of the site and the provider’s name if the two were different.¹⁴

Second, we noticed that respondents misidentified the auspices of many organizations. For example, different respondents identified the zoo as a public agency, a nonprofit, and a business. To correct this, we searched for this information ourselves by doing online research, calling organizations, and going to the Phoenix area. This resulted in coding venues into new categories. Table 2 presents a cross-classification of how respondents described the venue and our recoding. There was not a problem for neighborhood associations, churches, and businesses. Some activities that charities provided were reported as provided by voluntary associations or schools instead of nonprofits, but that was not necessarily inaccurate; however, 16 of the 132 activities provided by charities were reported as provided by businesses. Respondents also erroneously coded twenty-two activities that government agencies provided as being provided by voluntary associations or nonprofit organizations. As a result, we changed our categories somewhat. We kept codes for business, churches/congregations, and government agencies as they were, but we recoded voluntary associations and nonprofit organizations into neighborhood associations and public charities, depending on whether they had 501(c)(3) status (which many respondents would not know). In almost every case, the organizations identified as voluntary associations by respondents were indeed public charities, so this was a necessary change. We also checked the auspices of schools and assigned them to the appropriate category.

Table 2

Respondent Coding of 650 Venues That Provided Organized Activities for Children, Cross-Tabulated by Researcher Coding of Venue Auspices

	Project’s Recoding
Respondent’s Coding	Public Spaces	Charity	Neighborhood Association	Government/Tribe	Household	Congregation	Business	Travel	Missing	Total
Business	0	16	0	6	0	0	385	6	1	414
Voluntary association	0	27	2	10	0	2	0	1	0	42
Church	0	1	0	0	0	27	0	0	0	28
Nonprofit organization	0	66	0	12	0	1	2	1	0	82
Government agency	0	2	0	28	0	1	0	0	0	31
School	0	5	0	13	0	1	1	0	0	20
Don’t know	0	15	0	7	0	0	7	1	2	32
Refused	0	0	0	0	0	0	1	0	0	1
Total organized activities	0	132	2	76	0	32	396	9	3	650

We also decided to reexamine the “unorganized” activities that were provided at “other locations” and to assign a provider name based on the activities and site the respondent described. This meant going to maps and sometimes physically visiting sites to see what was located at a particular address. These were unorganized activities that took place at various locations in the family’s neighborhood or around the city but not at another household or as part of an organized program. The child may have been riding his or her bike down the street or in the desert, going for a walk or playing a pick-up basketball game in the park, or swimming or playing at an apartment complex pool. We coded the venues as either businesses, churches/congregations, government agencies, neighborhood associations, or charities depending on who owned the property. At times, however, none of these categories applied (e.g., riding a bike in the street or desert), and so these were coded as public spaces.

Table 1 also presents the proportion of all activities that were provided by different venues. The three most popular were businesses, households (other than the respondent’s), and governmental/tribal facilities. Charities ranked fourth, followed by public spaces, out-of-town venues, congregations, and neighborhood associations. Surprisingly, the latter two accounted for only 3.2 percent and 0.8 percent, respectively, of all activities.

Finally, parents or guardians provided data on the household and child during the phone interviews. The key variable in our analysis was household income, and we imputed values where we had missing data.¹⁵ Guided by the literature on children’s leisure time activities, we collected data on the child’s age, gender, and race/ethnicity (Hispanic, non-Hispanic white, non-Hispanic other), the number of children in the household, the number of years the family lived in the Phoenix-Mesa-Scottsdale metro area, whether it was a single-parent household,¹⁶ and whether the respondent worked at all the previous Saturday. The descriptive statistics are in Table 1.

The data have several strengths and weaknesses. They are cross-sectional, limited to one metropolitan area, and dated. We can draw generalizations for the Phoenix-Mesa-Scottsdale metropolitan area about what children did on Saturdays, and the demographics for our sample are comparable to family household characteristics in 2005.¹⁷ However, researchers know that urban realities in the western United States are different from those in places such as Chicago or in East Coast cities (Myers 2002), so we certainly cannot say anything about what children do outside the West. Future research will also need to determine if findings in Phoenix apply to other southwestern cities, such as Los Angeles, San Diego, Las Vegas, Tucson, and Albuquerque. Finally, the Phoenix area itself may be very different today than in 2003–2004.

Analysis

Table 3 is a cross-tabulation that shows a statistically significant relationship between family income and children’s activities (χ² = 55.3; df = 21; p < .001). By looking at a cell’s contribution to the χ² statistic, we can see which income groups were affiliated with different activities. We focus on cells where the contribution to χ² is 1.0 or greater. Children from upper-income families ($110,001–$300,000) were more likely to engage in sports-related and miscellaneous activities. Children from upper-middle-income families ($60,001–$110,000) were more likely to engage in sports-related activities as well and were less likely to shop/engage in personal care, play, or just socialize. In contrast, children from families that earned $20,001 to $60,000 (lower-middle-income) were more likely to shop/engage in personal care, play, or socialize and less likely to play sports. Similarly, children from families earning under $20,001 (low income) were more likely to play and socialize and less likely to engage in sports-related activities, attend spectator events, or engage in miscellaneous activities. Children at all income levels engaged in educational/art/performance/religious activities and eating at comparable rates.

Table 3

Association between Family Income and Activities

		Activities
Family Income		Eating	Shopping/Personal Care	Play	Social	Spectator Events	Education/Arts/Performance/Religious	Individual/Dual/Team Sports/Cheer	Miscellaneous	Total
≤$20,000	Frequency	14	14	27	24	4	8	18	3	112
	Expected	14.9	16	14.7	18.9	7.9	5.8	27.8	5.9
	Contribution to χ²	0.1	0.3	10.2	1.3	1.9	0.8	3.5	1.4	19.4
$20,001–$60,000	Frequency	63	83	73	92	30	23	93	24	481
	Expected	64.1	68.7	63.3	81.4	33.8	25	119.4	25.3
	Contribution to χ²	0	3	1.5	1.4	0.4	0.2	5.8	0.1	12.3
$60,001–$110,000	Frequency	69	58	44	65	37	25	144	24	466
	Expected	62.1	66.6	61.4	78.8	32.7	24.2	115.7	24.5
	Contribution to χ²	0.8	1.1	4.9	2.4	0.6	0	6.9	0	16.8
$110,001–300,000	Frequency	21	24	21	31	17	9	56	15	194
	Expected	25.9	27.7	25.5	32.8	13.6	10.1	48.2	10.2
	Contribution to χ²	0.9	0.5	0.8	0.1	0.8	0.1	1.3	2.2	6.8
Total		167	179	165	212	88	65	311	66	1,253
		1.8	4.8	17.4	5.3	3.7	1.1	17.5	3.7	55.3

NOTE: Pearson χ² (21) = 55.300; p < .001.

Figure 3 presents a multiple correspondence analysis (Clausen 1998) in which the child’s gender (male/female) and race/ethnicity (non-Hispanic white/non-Hispanic other/Hispanic) and family income (four categories as in Table 3) are row (or population) categories, and the seven activities (excluding miscellaneous and missing activities) are the column (or activity) categories. The entries in the input matrix are simple counts from the cross-tabulation of each row variable and the column variable. Correspondence analysis is a descriptive technique for detailed exploration of relationships between variables in a contingency table.¹⁸ The relationships between variable categories (profiles) are interpreted by looking at the distances between them on the map. The closer the row profiles (in our case male, female, Hispanic, white, other, and the four income categories) are to one another on the map, the more similar their distributions across column profiles (the seven activities in our case) and vice versa. The association between rows and columns can be inferred from the map by examining the angle between two lines, one going from the point of origin to a row profile (e.g., Hispanic child) and the other going from the point of origin to a column profile (e.g., government provider). The angle between the lines corresponds to a cosine that is a visualization of the correlation coefficient (Hsung and Breiger 2009). Smaller angles indicate positive correlations, a right angle indicates zero correlation, and large angles correspond to negative correlations. Points toward the center (zero-zero coordinates) represent categories that contribute very little to the inertia; that is, values in the cells simply reflect row and column marginals.¹⁹

The model in Figure 3 fits the data reasonably well. The total inertia is 0.341, and the eigenvalue associated with the first dimension is 0.137, which is 40.2 percent of the total inertia (variance). This dimension captures differences in activities across family income, with lowest and highest income categories at the opposite ends of the axis. For example, upper-middle-income and, to a lesser extent, upper-income families and non-Hispanic white children engage in sports, while lower-income and Hispanic children engage in play. The eigenvalue associated with the second dimension is 0.093 and explains 27.4 percent of the variance. This dimension captures differences along gender lines. The map shows that males fall somewhere between sports and play, while females engage in developmental activities, such as educational, art, performance, and religious activities, and shopping and personal care. Eating was in the center, because eating was something that everyone was equally likely to do.

Table 4 presents the cross-tabulation of family income and different venues: public spaces, charities, neighborhood associations, government/tribal facilities, other households, congregations, businesses, and out-of-town providers. There is a statistically significant association between family income and venue (χ²= 69.8; df = 21; p < .001), and the table again shows the relative contribution of each cell to the chi-square statistic. Again, we focus on cells where the contribution to χ² is 1.0 or greater. We see that low-income families are unlikely to use public spaces, businesses, charities, or out-of-town providers and more likely to use government providers or households. Like poor families, lower-middle-income families are less likely to use charities and more likely to use households. Upper-middle-income families are more likely to use public spaces or charities and less likely to use households or government. Upper-income families are more likely to use out-of-town providers and charities and less likely to use government or neighborhood associations. All income strata were equally likely to use congregations.

Table 4

Association between Family Income and Venue

		Venues:
Family Income:		Public Spaces	Charity	Neighborhood Association	Government/Tribe	Household	Congregation	Business	Out-of-Town Provider	Total
≤$20,000	Frequency	3	6	0	33	32	4	28	2	108
	Expected	5.4	12.2	0.9	17.2	25.1	3.5	39.8	3.9
	Contribution to χ²	1.1	3.1	0.9	14.6	1.9	0.1	3.5	1	26.1
$20,001–$60,000	Frequency	21	32	5	83	128	15	176	17	477
	Expected	24	53.7	3.9	75.8	110.9	15.5	175.9	17.4
	Contribution to χ²	0.4	8.8	0.3	0.7	2.6	0	0	0	12.8
$60,001–$110,000	Frequency	31	72	5	57	85	16	178	14	458
	Expected	23	51.6	3.7	72.7	106.5	14.8	168.9	16.7
	Contribution to χ²	2.8	8.1	0.4	3.4	4.3	0.1	0.5	0.4	20.1
$110,001–$300,000	Frequency	7	29	0	23	42	5	73	12	191
	Expected	9.6	21.5	1.5	30.3	44.4	6.2	70.4	7
	Contribution to χ²	0.7	2.6	1.5	1.8	0.1	0.2	0.1	3.6	10.7
Total		62	139	10	196	287	40	455	45	1,234
		4.9	22.6	3.2	20.5	9	0.4	4.1	5	69.8

NOTE: Pearson χ² (21) = 69.764; p < 0.001.

Figure 4 presents results from a correspondence analysis in which characteristics of the child and his or her family were again row variables and venue forms constituted the column variable in the input matrix. The total inertia for this model is 0.339, the eigenvalue associated with the first dimension is 0.142 and explains 41.8 percent of the total variance, and the eigenvalue associated with the second dimension is 0.073 and explains 21.6 percent of the variance. As in the analysis of activities, the first dimension in this figure captures differences across family income. The second dimension also seems to tap a gender dimension. Children from low-income families and Hispanic families are more likely to use government/tribal agencies, and Hispanics tend to affiliate with congregations. Children from lower-middle-income families and non-Hispanic other race children used household providers. Females tended to use businesses or neighborhood associations. Whites and upper-income families used out-of-town providers, public spaces, and businesses. Finally, upper-middle-class families tended to use charities and public spaces. The latter finding is a result of the open spaces on the edges of the metropolitan area and the popularity of off-road vehicles.

Given that the use of venues by different population segments is probably due to the activities that different venues provided, we examined which organizational forms provided which types of services/activities. Indeed there is a division of labor among charities, neighborhood associations, nonprofits, for-profits, households, public spaces, and government organizations. In Table 5 we present the percentage of different activities provided by different venues. Looking at cases where the venue provided at least 20 percent of the activities of a given type, we see that businesses are important providers of food services, shopping and personal care, and, to a lesser extent, spectator events. Congregations accounted for the bulk of educational/art/performance/religious activities; family households were the sites for social activities and play; government/tribal agencies provided sports activities, spectator events (e.g., Arizona State University sports), educational/art/performance/religious activities, and play opportunities; and charities were an important provider of sports. The correlation between activity and venue makes it imperative that we control for the former in our next analysis.

Table 5

Cross-Tabulation of Activity Types and Venue Types

	Venue:
Activity:	Public Spaces	Charity	Neighborhood Association	Government/Tribe	Household	Congregation	Business	Travel	Missing	Total
Eating	0	1	0	0	9	1	154	0	2	167
(percentage)	(0.0)	(0.6)	(0.0)	(0.0)	(5.4)	(0.6)	(92.2)	(0.0)	(1.2)	(100.0)
Shopping/personal care	0	4	0	2	6	0	167	0	0	179
(percentage)	(0.0)	(2.2)	(0.0)	(1.1)	(3.4)	(0.0)	(93.3)	(0.0)	(0.0)	(100.0)
Play	10	1	6	56	80	0	5	0	7	165
(percentage)	(6.1)	(0.6)	(3.6)	(33.9)	(48.5)	0	(3.0)	0	(4.3)	(100.0)
Socializing	0	10	1	12	167	5	13	0	4	212
(percentage)	(0.0)	(4.7)	(0.5)	(5.6)	(78.8)	(2.4)	(6.1)	0	(1.9)	(100.0)
Spectator events	0	11	1	19	2	0	55	0	0	88
(percentage)	(0.0)	(12.5)	(1.1)	(21.6)	(2.3)	(0.0)	(62.5)	(0.0)	(0.0)	(100.0)
Education/arts/performance/religious	0	10	0	18	4	27	5	0	1	65
(percentage)	(0.0)	(15.4)	(0.0)	(27.7)	(6.2)	(41.5)	(7.7)	0	(1.5)	(100.0)
Individual/team/dual sports/cheer	46	102	2	89	10	7	47	0	8	311
(percentage)	(14.8)	(32.8)	(0.6)	(28.6)	(3.2)	(2.3)	(15.1)	(0.0)	(2.6)	(100.0)
Miscellaneous	6	0	0	0	9	0	6	45	0	66
(percentage)	(9.1)	(0.0)	(0.0)	(0.0)	(13.6)	(0.0)	(9.1)	(68.2)	(0.0)	(100.0)
Missing	0	0	0	0	0	0	3	0	0	3
(percentage)	(0.0)	(0.0)	(0.0)	(0.0)	(0.0)	(0.0)	(100.0)	(0.0)	(0.0)	(100.0)
Total	62	139	10	196	287	40	455	45	22	1,256
(percentage)	(4.9)	(11.1)	(0.8)	(15.6)	(22.8)	(3.2)	(36.2)	(3.6)	(1.8)	(100.0)

Table 6 presents the results of a multinomial logistic regression. Multinomial regression is the most suitable statistical tool for testing hypotheses about choosing one provider rather than another, because it is designed for estimating models where dependent variables are nominal variables with three or more categories. The dependent variable in our case is the type of venue. The multinomial regression models the log odds of categorical outcomes as a linear combination of independent variables. The multinomial logistic function is an extension of a logit function. A set of coefficients is estimated for each category of the dependent variable, except the one that is chosen as a base category.²⁰

Table 6

Coefficients from the Multinomial Logistic Regressions

	Dependent Variable: Venue (Reference Category: Charity)
Independent Variables	Business: b (robust SE)	Congregation: b (robust SE)	Household: b (robust SE)	Government/Tribe: b (robust SE)
Child female	.417	−.406	.485*	−.071
	(.286)	(.484)	(.283)	(.262)
Child Hispanic	−.120	1.155**	−.012	.527*
	(.315)	(.556)	(.327)	(.291)
Child other race/ethnicity	−.573	.295	.243	.125
	(.567)	(.728)	(.493)	(.482)
Child age	−.033	−.007	.021	−.094
	(.060)	(.110)	(.065)	(.061)
Number of children	.083	−.148	−.016	−.030
	(.151)	(.264)	(.161)	(.153)
Years living in Phoenix metro area	.011	−.031	.010	−.008
	(.010)	(.019)	(.010)	(.010)
Family income in $1,000s	−.001	.000	−.004	−.007***
	(.003)	(.005)	(.003)	(.003)
Caregiver not married	.905**	.620	1.064***	.647*
	(.377)	(.647)	(.395)	(.350)
Caregiver worked last Saturday	−.343	−.234	−.198	−.926**
	(.432)	(.633)	(.376)	(.370)
Eating/shopping/personal care/spectator events^a	2.718****	−1.977	−3.065****	−1.562***
	(.465)	(1.236)	(.461)	(.499)
Educational/art/ performance/religious^a	−1.129	2.122***	−3.979****	−1.071**
	(.694)	(.710)	(.695)	(.534)
Individual/dual/team sports/cheerleading^a	−1.360***	−1.849**	−5.440****	−1.780****
	(.426)	(.796)	(.458)	(.358)
Constant	.338	−.451	2.695****	3.092****
	(.708)	(1.269)	(.699)	(.684)
Log-likelihood	−881.12
N of activities	1,067
Wald chi-squared	830.26****
Pseudo R²	.411

Reference category is “play” and “socializing.”

p < .1. **p < .05. ***p < .01. ****p < .001 (two-tailed tests).

The units of analysis were the activities that the children engaged in. The dependent variable is a multinomial response variable that took on seven values: public space, charity, neighborhood association, government/tribal, household, congregations, or business. Out-of-town providers (travel) and missing cases were excluded because we were unsure about the auspices of the providers. Informed by our literature review, the independent variables were household income, the race/ethnicity of the child (Hispanic/non-Hispanic other/non-Hispanic white), the child’s gender and age, the number of children in the household, the number of years the family lived in the Phoenix-Mesa-Scottsdale metro area, whether the caregiver was single, whether she or he worked that previous Saturday, and whether the activity was education/art/religious, play, socializing, sports, eating, personal care, or a spectator event. We excluded activities that were miscellaneous or had missing values.

A constraint on our model is that every venue (the categories of the multinomial dependent variable) had to provide every service to at least one child. However, from Table 5, we see that congregations did not provide spectator events, play activities, or personal care; neighborhood associations did not provide eating, personal care, or educational/art/religious activities; and there was no eating, personal care, socializing, spectator events, or educational/art/religious activities in public space. This means that we had to collapse either venues or activities, or eliminate some. We decided to drop two venues—public space and neighborhood association—because they provided so few activities (see Table 1). We also collapsed spectator events, personal care (e.g., haircuts), shopping, and eating, since, for the most part, they were commercial transactions and provided by businesses. Since we did not want to lose play, we folded socializing and play together because both were informal activities.

Because activities were not independent (the same children participated in multiple activities), we used the cluster option in Stata and estimated robust standard errors. Socializing/play was the reference category for the activity dummies, and charity was the reference category for the multinomial dependent variable. In columns 1 through 3, we see that charity providers were not any more or less likely to serve rich or poor families than were businesses, congregations, or households. Looking at the comparisons between government and charities (column 4), however, we see that as family income decreased, families were more likely to use government providers than charities. Looking at the controls, we see that single caregivers were more likely to use businesses or other households than charities, and caregivers who worked on the previous Saturday were more likely to use charities than government. We also see that Hispanics were more likely to use churches or government agencies than charities, although the latter was significant only at the .10 level (p < .070, two-tailed test).

Looking at activities, we see that businesses were more likely to provide eating, personal care, and spectator events than play and socializing activities in comparison to charities, but charities were more likely to provide sports than play or socializing opportunities in comparison to businesses. Congregations were more likely to provide educational/art/performance/religious activities than play or socializing in comparison to charities, but charities were more likely to provide sports than play or socializing opportunities in comparison to congregations. Finally, charities were more likely to provide eating, shopping, personal care, and spectator activities, developmental activities, and sports than play and socializing activities compared to both households and government/tribal agencies. The centrality of charities in providing sporting activities is fairly clear.

We subsequently redid the analysis in Table 6, substituting different venues for the reference category, which initially was charity. This enabled us to create an ordering among the venues for those with more or less income (detailed results available upon request). Table 7 summarizes the findings and speaks directly to our hypotheses. Focus on the upper right off-diagonal cells; each cell contains the logistic regression coefficient describing the increase in the log odds ratio of using venue i (e.g., a business) rather than j (e.g., a congregation) with a $1,000 increase in family income, controlling for family and child characteristics and the activity. There are only two statistically significant results. As family income increased, children were more likely to use businesses than government/tribal venues (b = .007), and less likely to use government/tribal venues than charities (b = –.007).

Table 7

Coefficients from the Multinomial Logistic Regressions with Family Income (Upper Right Off-Diagonal) and Hispanic Ethnicity (Lower Left Off-Diagonal) as the Independent Variables and Venue Type as the Dependent Variable (with Different Venues as the Reference Category)

	Reference Category
Choice	Business: b (SE)	Congregation: b (SE)	Household: b (SE)	Government/Tribe: b (SE)	Charity: b (SE)
Business		−.000 (.005)	.003 (.003)	.007** (.003)	−.001 (.003)
Congregation	1.275** (.558)		.003 (.005)	.007 (.005)	−.000 (.005)
Household	.108 (.288)	−1.167** (.559)		.004 (.003)	−.004 (.003)
Government/tribe	.667** (.269)	−.007 (.005)	.539** (.253)		−.007*** (.003)
Charity	.120 (.315)	−1.155** (.556)	.012 (.327)	−.527* (.291)

p < .1. **p < .05. ***p < .01 (two-tailed tests).

The second finding fit our expectations: wealthier families were more likely to use businesses than government providers, while poorer families were more likely to use government providers than businesses. The first finding was surprising, since we had expected that income would be unrelated to the use of charities—because of cross-subsidization, we thought both the rich and the poor would be served by charities. Instead, the wealthy chose charities over government, while the poor affiliated with government rather than charities. This is consistent with correspondence analysis results, and we discuss this further in the conclusion.

Further Analysis of Race/Ethnicity

Our theory speculated that residents’ affiliations with different providers would be driven by costs and perceived quality. In the analysis discussed above, we saw higher-income families’ aversion to government providers. In an exploratory analysis, we examined an alternative basis for differentiation: race and ethnicity. If income alone is dictating who uses what, controlling for income, and race and ethnicity should have little or no effect on choice of venue.

Using the same results that were shown in Table 6 and again alternating the reference categories, we come up with the results in the lower left hand off-diagonal cells in Table 7, which assess the impact of being Hispanic or non-Hispanic non-white on families’ choices. First, being non-Hispanic non-white had no effect on the choice of venue. There is no doubt that this is because this was a residual category that included blacks, Asians, Pacific Islanders, and Native Americans. Thus, we do not present these results. Second, when we compared Hispanics to non-Hispanic whites, we found that independent of income, child and household characteristics, and activity, Hispanics were more likely than non-Hispanic whites to use congregations or government/tribal agencies rather than businesses, households, or charities; while businesses, households, and charities remained the domain of non-Hispanic whites.²¹ Given that these results were independent of family income, they are provocative. These results are also consistent with the correspondence analysis results and are discussed in the conclusion.

Conclusion

This article describes the ecology of affiliation in the Phoenix-Mesa-Scottsdale metro area, looking at what children did on Saturdays during the school year and the organizations and venues they utilized. We drew on the recent work in economic sociology, institutional economics, and organizational ecology to describe in a single-case metropolitan community the affiliations of different population subgroups or market segments with different venues.

We found that the poor and Hispanics were more likely to use government/tribal venues than businesses or charities, while middle- and upper-income families and whites were more likely to use businesses or charities than government/tribal venues. Hispanics also preferred congregations over businesses, households, and charities; while non-Hispanic whites preferred businesses, households, or charities over congregations. Finally, Hispanics favored government venues over households, while whites favored the latter over the former. Both the regression analyses in Tables 6 and 7 and the correspondence analysis in Figure 4 support these conclusions.

We expected and found that lower-income residents were more likely to use government providers than businesses, and wealthier families were more likely to use businesses than government providers. Looking at Table 4, however, we see that this effect is mostly due to the fact that families earning $20,000 or less did not use business vendors and relied heavily on government/tribal providers, and upper-middle-income and upper-income families stayed away from government facilities. Thus, the effects are not strictly linear. We argued that businesses charge fees, which create barriers to entry for low-income families, and thus poorer residents are kept out of their facilities. At the same time, wealthier families avoid public venues where services are supposedly inferior because more people utilize government providers.

So far we have assumed that governments provide inferior services compared to businesses. That is, services provided by government are cheaper because they are inferior to those provided by businesses. Since the wealthy have more disposable income, they naturally avoid the low-end provider. In our analysis, we did not control for the quality of the service provided, but we can examine the effects of auspice on family members’ satisfaction with services their children received. With a measure of adults’ satisfaction with the child’s provider as the dependent variable, we can regress this measure on the various auspices of providers. Galaskiewicz et al. (2012, 257) examined the factors that contributed to parents/guardians’ satisfaction with the providers’ of their children’s organized activities. In Table 8, model 1, we reproduced the results in Galaskiewicz et al.’s (2012) Table 4.²² We then added provider auspice to model 1. Model 2 shows that none of the substantive effects changed from what Galaskiewicz et al. (2012) found. More importantly, parents who utilized a government provider were somewhat more satisfied than those that used a business, charity, or church, but none of the differences was statistically significant. Thus it does not appear that government provided an inferior product or service. If middle- and upper-income people did not use government facilities, it may be due to some other explanation.

Table 8

Ordinary Least Squares (OLS) Regression (N = 650 Activities Organized by Organizations)

Dependent Variable: Parents’ Satisfaction with Provider
Independent Variables	Model 1: b (Robust SE)	Model 2: b (Robust SE)
Control variables
Child female	.006 (.018)	.005 (.019)
Child’s age	.003 (.004)	.003 (.004)
Child Hispanic	−.018 (.023)	−.019 (.023)
Child non-Hispanic, non-white	−.066 (.027)**	−.067 (.026)**
Number of children	−.008 (.012)	−.007 (.013)
Years living in Phoenix	.003 (.001)***	.003 (.001)****
Single parent	.022 (.023)	.023 (.023)
Family attends church	.011 (.020)	.013 (.021)
Parent(s) work full-time	.069 (.030)**	.071 (.030)**
Financial and cultural capital variables
Caregiver’s years of education	.008 (.004)**	.008 (.004)**
Family income (in $1,000s)	−.000 (.000)	−.000 (.000)
Activity variables
Activity is free	.016 (.026)	.005 (.027)
Religious/art/performance/educational	.079 (.037)**	.068 (.042)
Team/dual/individual sports/cheerleading	.064 (.019)***	.044 (.025)*
Social capital variables
Family in metro area	−.027 (.019)	−.026 (.019)
Friends in metro area	.021 (.019)	.022 (.019)
Ties to organization	.089 (.025)****	.083 (.027)***
Auspice variables
Business provider		−.039 (.030)
Church provider		−.018 (.048)
Charity provider		−.011 (.031)
Constant	.515 (.067)****	.553 (.073)****
N of activities (max. = 650)	609	606
N of children (max. = 372)	347	345
F	6.41****	5.54****
R-squared	.111	.116

NOTE: Model 1 comes from Galaskiewicz et al. (2012, 257), Table 4.

p < .1. **p < .05. ***p < .01. ****p < .001 (two-tailed tests).

We expected that charities would serve a broad range of people, earning revenues from the wealthy that would cross-subsidize programs for the poor. However, we found that upper-income families preferred charities over government providers, just as they preferred businesses over government venues. In other words, charities’ clientele was the same as that of businesses—middle- to upper-income families. A probable reason for our results is that many of the charities in our sample were sports clubs, teams, and leagues. Table 5 showed that 102 of the 139 charities in our study were sports-related; we subsequently found that 82 of these 102 were sports clubs or associations.²³ From our discussions with informants, we discovered that many functioned more like associative nonprofits than charities, relying heavily on dues, fees, and volunteers.

Indeed, looking at revenue sources for the 196 public charities that provided recreation and sports services in the Phoenix-Mesa-Scottsdale metro area in 2004, on average, 26.7 percent of their revenue was from dues.²⁴ This is in contrast to revenues of 2.7 percent from dues for all reporting operating human services in 2005, which includes recreational and sports services (Wing, Pollak, and Blackwood 2008, 183). Therefore, this type of charity will be more accessible to middle- and upper-income parents who have the means to pay these dues, but it will discourage the participation of lower-income children. Given that other research has found that charities are often not redistributive, these results are not completely surprising.

We found that governments and congregations were the preferred provider of Hispanics and that businesses and charities were the preferred provider of whites. That the results were independent of income makes them all the more intriguing. Since our theory is premised on institutional incentives and family resources, it has difficulty explaining why race or ethnicity should impact affiliations. It could be that Hispanics prefer religious rather than secular venues, but it is difficult to explain why non-Hispanic whites prefer charities and business over congregations and why Hispanics are drawn to government providers. Thus, we may be required to look at other factors to explain our findings.

It may be necessary to introduce spatial factors into our analysis. The unspoken assumption in this article is that it does not matter where things are located. The model assumes that venues are distributed randomly across the metropolitan area, and transportation costs have little or no effect on what venues parents choose. Yet a core idea of Wilson’s (1987) discussion was that there are many neighborhoods in metropolitan areas that are “organizational deserts”; that is, organizations that would provide basic services to a community are simply not present. This was also the conclusion of Small and McDermott (2006). In the Phoenix-Mesa-Scottsdale metropolitan area, it may be that businesses or charities are not located in areas where Hispanics or poor people live, but congregations and government facilities are. It is true that minorities in these situations can drive to facilities outside their neighborhoods, but it adds a cost. Thus, to explain why residents utilize the organizational resources they do, researchers need to examine data on what organizational resources are located nearby.

Alternatively, prejudice and discrimination can play a role. Groups within the metropolitan area may have a preference for exclusivity. In other words, both wealthy and poor families, Hispanic and non-Hispanic white families may prefer not to associate with people who are different from themselves. This pattern is found often in network studies of friendship (McPherson, Smith-Lovin, and Cook 2001) and is labeled a homophily effect. This would mean that segments of the community may avoid venues that are “home territories” for other groups (Lofland 1973).²⁵ These avoidance strategies may be aimed at specific establishments, but they could be used in the decision-making process to eliminate types of organizations that are perceived to be somebody else’s.

Establishments may also contribute to this. If exclusivity is important to customers and users, vendors that seek the patronage of certain groups may actively discriminate to exclude others. While discriminating on the basis of income is commonplace among businesses, it seems distasteful if charities do so. However, discrimination on the basis of race or ethnicity is illegal. Yet Feagin’s (1991) classic description of the indignities suffered by middle-class blacks in department stores and restaurants illustrates how racial prejudice still plays out in public venues and creates barriers to entry that have nothing to do with ability to pay. Whether businesses or sports-related charities in the Phoenix area favor white over Hispanic users is certainly worthy of further investigation.

In summary, we, like previous researchers, found that different segments within the community affiliate with different types of providers. We also found that family income and organizational auspice can explain to some extent who matches up with what. However, ability to pay alone does not explain all our results, and it seems necessary to broaden our inquiry to take into account other factors that could explain the patterns we found in our case community.

Footnotes

NOTE:

This article was presented as a paper at the conference, “Organizational Intervention and Urban Poverty in the 21st Century,” sponsored by the University of Chicago, March 10–11, 2011. The National Science Foundation (SES-0241559) provided the funding for this research. We want to extend our special thanks to Shap Wolf and Bill Edwards, both formerly affiliated with The Institute for Social Science Research at Arizona State University, who collected these data; and to Mathew Green, Steve Corral, Joy Inouye, and Paola Molina, who helped with the data cleaning. Also thanks to Ronald Breiger for helping us appreciate the potential of affiliation data and correspondence analysis; George Hobor, who helped us find references on what children do in their leisure time; and Kendra Thompson-Dyck and Kathryn Anderson for references on supermarkets, recreational facilities, and residents’ health. We also wish to thank the Social and Behavioral Science Research Institute and the Udall Center at the University of Arizona for their support. Finally, this article was greatly influenced by the research of Edward O. Laumann, Franz Pappi, and J. Miller McPherson, who did pioneering research on the interface between community residents, elites, and organizations.

Notes

Joseph Galaskiewicz is a professor of sociology at the University of Arizona. He studies the Phoenix metro area, looking at the census of establishments that provide activities and services for children. He is studying which neighborhoods gained organizational resources between 2003 and 2007 and which lost resources during the Great Recession.

Olga V. Mayorova is an associate professor in the Department of Management at the Higher School of Economics-Perm in Russia. She is interested in stratification, social networks, and organizations. She has coauthored articles on stratification processes in Russia and reintegration processes in the Republic of Georgia.

Beth M. Duckles is an assistant professor of sociology at Bucknell University. Her current research is an ethnography of the U.S. Green Building Council’s development of the Leadership in Energy and Environment Design (LEED) 2012/v. 4 rating system. Her research interests include environmental sociology, nonprofits, and mixed methods.

References

Agresti

Alan

. 1996. An introduction to categorical data analysis. New York, NY: Wiley.

Allard

Scott W.

Tolman

Richard

Rosen

Daniel

. 2003. Proximity to service providers and service utilization among welfare recipients: The interaction of place and race. Journal of Policy Analysis and Management 22 (4): 599–613.

Austin

S. Bryn

Melley

Steven J.

Sanchez

Brisa N.

Patel

Aarti

Buka

Stephen

Gortmaker

Steven

. 2005. Clustering of fast-food restaurants around schools: A novel application of spatial statistics to the study of food environments. American Journal of Public Health 95 (9): 1575–81.

Bianchi

Suzanne M.

Robinson

John P.

1997. What did you do today? Children’s use of time, family composition, and the acquisition of social capital. Journal of Marriage and Family 59 (2): 332–44.

Bianchi

Suzanne M.

Robinson

John P.

Milkie

Melissa A.

2006. Changing rhythms of American family life. New York, NY: Russell Sage Foundation.

Blustein

Jan

Hoy

Emma C.

2000. Who is enrolled in for-profit vs. nonprofit Medicare HMOs? Health Affairs 19 (1): 210–20.

Bourdieu

Pierre

. 1977. Cultural reproduction and social reproduction. In Power and ideology in education, eds. Karabell

Jerome

Halsey

A. H.

, 481–551. Oxford: Oxford University Press.

Bozeman

Barry

. 1987. All organizations are public: Bridging public and private organizational theories. San Francisco, CA: Jossey-Bass.

Clarke

Lee

Estes

Carroll

. 1992. Sociological and economic theories of markets and nonprofits: Evidence from home health organizations. American Journal of Sociology 97 (4): 945–69.

10.

Clausen

Sten-Erik

. 1998. Applied correspondence analysis: An introduction. Thousand Oaks, CA: Sage.

11.

Clotfelter

Charles T.

, ed. 1992. Who benefits from the nonprofit sector? Chicago, IL: University of Chicago Press.

12.

Coleman

James S.

Hoffer

Thomas

Kilgore

Sally B.

1982. High school achievement: Public, Catholic and other schools compared. New York, NY: Basic Books.

13.

Cooper

Harris

Valentine

Jeffrey

Nye

Barbara

Lindsay

James

. 1999. Relationships between five after-school activities and academic achievement. Journal of Educational Psychology 91 (2): 369–78.

14.

DiMaggio

Paul

Anheier

Helmut

. 1990. The sociology of nonprofit organizations and sectors. Annual Review of Sociology 16 (1): 137–59.

15.

Dunn

Janet S.

Kinney

David A.

Hofferth

Sandra L.

2003. Parental ideologies and children’s after school activities. American Behavioral Scientist 46 (10): 1359–86.

16.

Edwards

Bill

. 12 March 2004. Valley youth activity provider survey. Methodological memo, Survey Research Laboratory, Arizona State University, Tempe.

17.

Feagin

Joe R.

1991. The continuing significance of race: Antiblack discrimination in public places. American Sociological Review 56 (1): 101–16.

18.

Ferris

James

Grady

Elizabeth

. 1989. Fading distinctions among the nonprofit, government and for-profit sectors. In The future of the nonprofit sector, eds. Hodgkinson

Virginia

Lyman

Richard

, 123–41. New York, NY: Jossey-Bass.

19.

Frumkin

Peter

Galaskiewicz

Joseph

. 2004. Institutional isomorphism and public sector organizations. Journal of Public Administration Research and Theory 14:283–307.

20.

Galaskiewicz

Joseph

Hobor

George

Duckles

Beth

Mayorova

Olga V.

2012. Caregivers’ social capital and satisfaction with their children’s service providers. Sociological Perspectives 55 (2): 241–66.

21.

Goodman

Leo A.

1996. A single general method for the analysis of cross-classified data: Reconciliation and synthesis of some methods of Pearson, Yule, and Fisher, and also some methods of correspondence analysis and association analysis. Journal of the American Statistical Association 91:408–28.

22.

Greenacre

Michael J.

1984. Theory and applications of correspondence analysis. London: Academic Press.

23.

Hanks

Michael

Eckland

Bruce K.

1978. Adult voluntary associations and adolescent socialization. Sociological Quarterly 19:481–90.

24.

Hannan

Michael T.

Freeman

John

. 1989. Organizational ecology. Cambridge, MA: Harvard University Press.

25.

Hansmann

Henry

. 1996. The ownership of enterprise. Cambridge, MA: Belknap.

26.

Harvey

Andrew S.

1999. Guidelines for time use data collection and analysis. In Time use research in the social sciences, eds. Pentland

Wendy E.

Harvey

Andrew S.

Lawton

M. Powell

McColl

Mary Ann

, 19–45. New York, NY: Kluwer.

27.

Hofferth

Sandra L.

Sandberg

John F.

2001. How American children spend their time. Journal of Marriage and Family 63 (May): 295–308.

28.

Hoffmann

John

Jiangmin

. 2002. School activities, community service, and delinquency. Crime and Delinquency 48 (4): 568–91.

29.

Hosmer

David W.

Lemeshow

Stanley

. 2000. Applied logistic regression. 2nd ed. New York, NY: Wiley.

30.

Hou Moynihan

Donald P.

2008. The case for countercyclical fiscal capacity. Journal of Public Administration Research and Theory 18 (1): 139–59.

31.

Hsung

Ray-May

Breiger

Ronald L.

2009. Position generators, affiliations, and the institutional logics of social capital: A study of Taiwan firms and individuals. In Contexts of social capital: Social networks in markets, communities, and families, eds. Hsung

Ray-May

Lin

Nan

Breiger

Ronald L.

, 3–27. New York, NY: Routledge.

32.

Internal Revenue Service (IRS). 2012a. Exempt purposes—Internal Revenue Code section 501(c)(3). U.S. Department of the Treasury. Available from http://www.irs.gov (accessed 15 September 2012).

33.

IRS. 2012b. Exemption requirements: Section 501(c)(3) organizations. U.S. Department of the Treasury. Available from http://www.irs.gov (accessed 15 September 2012).

34.

IRS. 2012c. Inurement/private benefit—Charitable organizations. U.S. Department of the Treasury. Available from http://www.irs.gov (accessed 15 September 2012).

35.

James

Estelle

. 1986. How nonprofits grow: A model. In The economics of nonprofit institutions: Studies in structure and policy, ed. Rose-Ackerman

Susan

, 185–95. New York, NY: Oxford University Press.

36.

Kahn

Faith

. 1997. Pandora’s box: Managerial discretion and the problem of corporate philanthropy. UCLA Law Review 44 (3): 579–676.

37.

Kahne

Joseph

Bailey

Kim

. 1999. The role of social capital and youth development: The case of “I have a dream” programs. Educational Evaluation and Policy Analysis 21 (3): 321–43.

38.

Kipke

Michele D.

Iverson

Ellen

Moore

Deborah

Booker

Cara

Ruelas

Valerie

Peters

Anne L.

Kaufman

Francine

. 2006. Food and park environments: Neighborhood-level risks for childhood obesity in East Los Angeles. Journal of Adolescent Health 40:325–33.

39.

Kushman

John E.

1979. A three-sector model of day care center services. Journal of Human Resources 14 (4): 543–62.

40.

Lareau

Annette

. 2002. Invisible inequality: Social class and childrearing in black families and white families. American Sociological Review 67 (5): 747–76.

41.

Lareau

Annette

. 2003. Unequal childhoods: Class, race and family. Berkeley, CA: University of California Press.

42.

Little

Roderick J. A.

Rubin

Donald B.

1987. Statistical analysis with missing data. New York, NY: Wiley.

43.

Lofland

Lyn H.

1973. A world of strangers: Order and action in urban public space. Prospect Heights, IL: Waveland.

44.

Marsh

Herbert

Kleitman

Sabina

. 2002. Extracurricular school activities: The good, the bad, and the nonlinear. Harvard Educational Review 72 (4): 464–515.

45.

Marwell

Nicole P.

2007. Bargaining for Brooklyn: Community organizations in the entrepreneurial city. Chicago, IL: University of Chicago Press.

46.

McFarland

Daniel

Thomas

Reuben

. 2006. Bowling young: How youth voluntary associations influence adult political participation. American Sociological Review 71 (3): 401–25.

47.

McPherson

J. Miller

. 1983. An ecology of affiliation. American Sociological Review 48 (4): 519–35.

48.

McPherson

J. Miller

Smith-Lovin

Lynne

Cook

Karen

. 2001. Birds of a feather: Homophily in social networks. Annual Review of Sociology 27:415–44.

49.

Medrich

Elliott A.

Roizen

Judith A.

Rubin

Victor

Buckley

Stuart

. 1982. The serious business of growing up: A study of children’s lives outside school. Berkeley, CA: University of California Press.

50.

Moore

Latetia V.

Roux

Ana V. Diez

. 2006. Associations of neighborhood characteristics with the location and type of food stores. American Journal of Public Health 96:325–31.

51.

Moore

Latetia V.

Roux

Ana V. Diez

Evenson

Kelly R.

McGinn

Aileen P.

Brines

Shannon J.

2008. Availability of recreational resources in minority and low socioeconomic status areas. American Journal of Preventive Medicine 34 (1): 16–22.

52.

Moore

Mark H.

1995. Creating public value: Strategic management in government. Cambridge, MA: Harvard University Press.

53.

Morland

Kimberly

Wing

Steve

Roux

Ana Diez

Poole

Charles

. 2002. Neighborhood characteristics associated with the location of food stores and food service places. American Journal of Preventive Medicine 22 (1): 23–29.

54.

Myers

Dowell

. 2002. Demographic dynamism in Los Angeles, Chicago, New York, and Washington, D.C. In From Chicago to L.A.: Making sense of urban theory, ed. Dear

Michael J.

, 17–54. Thousand Oaks, CA: Sage.

55.

Okada

Sosuke.

2011. The relational structure of musical taste. Master’s of Arts Paper, Department of Sociology, University of Arizona, Tucson.

56.

Olfson

Mark

Mechanic

David

. 1996. Mental disorders in public, private nonprofit, and proprietary general hospitals. American Journal of Psychiatry 153:1613–19.

57.

Otto

Luther B.

1976. Social integration and the status-attainment process. American Journal of Sociology 81:1360–83.

58.

Powell

Lisa M.

Chaloupka

Frank J.

Bao

Yanjun

. 2007. The availability of fast-food and full-service restaurants in the United States: Associations with neighborhood characteristics. American Journal of Preventive Medicine 33 (4S): S240–45.

59.

Powell

Lisa M.

Slater

Sandy

Mirtcheva

Donka

Bao

Yanjun

Chaloupka

Frank J.

2007. Food store availability and neighborhood characteristics in the United States. Preventive Medicine 44:189–95.

60.

Rainey

Hal G.

1997. Understanding and managing public organizations. 2nd ed. San Francisco, CA: Jossey-Bass.

61.

Robinson

John R.

1999. Time diary method: Structure and uses. In Time use research in the social sciences, eds. Pentland

Pentland Wendy E.

Harvey

Andrew S.

Lawton

M. Powell

McColl

Mary Ann

, 47–89. New York, NY: Kluwer.

62.

Schlesinger

Mark

. 1998. Mismeasuring the consequences of ownership: External influences and the comparative performance of public, for-profit, and private nonprofit organizations. In Private action and the public good, eds. Powell

Walter W.

Clemens

Elisabeth

, 85–113. New Haven, CT: Yale University Press.

63.

Schlesinger

Mark

Dorwart

Robert

. 1984. Ownership and mental-health services. New England Journal of Medicine 311 (15): 959–65.

64.

Schlesinger

Mark

Mitchell

Shannon

Gray

Bradford

. 2003. Measuring community benefits provided by nonprofit and for-profit HMOs. Inquiry 40 (2): 114–32.

65.

Simon

John

Dale

Harvey

Chisolm

Laura

. 2012. The federal tax treatment of charitable organizations. In The nonprofit sector: A research handbook, 2nd edition, eds. Powell

Walter W.

Steinberg

Richard

, 267–306. New Haven, CT: Yale University Press.

66.

Sister

Maria Chona E.

2007. Do blacks and browns have less green? Examining the distribution of park and open space resources in the greater Los Angeles metropolitan region. PhD diss., University of Southern California, Los Angeles.

67.

Small

Mario Luis

. 2009. Unanticipated gains: Origins of network inequality in everyday life. New York, NY: Oxford University Press.

68.

Small

Mario Luis

McDermott

Monica

. 2006. The presence of organizational resources in poor urban neighborhoods: An analysis of average and contextual effects. Social Forces 84 (3): 1697–1724.

69.

Spady

William G.

1970. Lament for the letterman: Effects of peer status and extracurricular activities on goals and achievement. American Journal of Sociology 75:680–702.

70.

Steinberg

Richard

Gray

Bradford H.

1993. The role of nonprofit enterprise in 1993: Hansmann revisited. Nonprofit and Voluntary Sector Quarterly 22 (4): 297–316.

71.

Timmer

Susan Goff

Eccles

Jacquelynne

O’Brien

Kerth

. 1985. How children use time. In Time, goods, and well-being, eds. Juster

F. Thomas

Stafford

Frank P.

, 353–82. Ann Arbor, MI: Survey Research Center, Institute for Social Research, University of Michigan.

72.

Valentine

Jeffrey

Cooper

Harris

Bettencourt

B. Ann

DuBois

David

. 2002. Out-of-school activities and academic achievement: The mediating role of self-beliefs. Educational Psychologist 37 (4): 245–56.

73.

Verba Schlozman

Kay L.

Brady

Henry E.

1995. Voice and equality: Civic volunteerism in American politics. Cambridge, MA: Harvard University Press.

74.

Wamsley

Gary L.

Zald

Mayer N.

1973. The political economy of public organizations: A critique and approach to the study of public administration. Lexington, MA: Lexington Books.

75.

Watkins-Hayes

Celeste

. 2011. Peace, respect, and red tape: Inside the black box of racially representative bureaucracies. Journal of Public Administration Research and Theory 21 (Supplement): 233–51.

76.

Weisbrod

Burton A.

1988. The nonprofit economy. Cambridge, MA: Harvard University Press.

77.

Weisbrod

Burton A.

1998. Institutional form and organizational behavior. In Private action and the public good, eds. Powell

Walter W.

Clemens

Elisabeth

, 69–84. New Haven, CT: Yale University Press.

78.

Wilson

William Julius

. 1987. The truly disadvantaged: The inner city, the underclass, and public policy. Chicago, IL: University of Chicago Press.

79.

Wing

Kennard T.

Pollak

Thomas H.

Blackwood

Amy

. 2008. The nonprofit almanac 2008. Washington, DC: Urban Institute Press.

80.

Wolpert

Julian

. 1993. Decentralization and equity in public and nonprofit sectors. Nonprofit and Voluntary Sector Quarterly 22 (4): 281–96.

81.

Wolpert

Julian

. 1996. Half a loaf: The limits of charity in the nineties. New York, NY: Twentieth Century Fund.