Measuring and indigenizing social capital in relation to children's street work in Mexico: the role of culture in shaping social capital indicators.

Drawing from social capital theory, this study assessed the relevance qf existing conceptions of social capital--largely from the United States and Canada--in the Mexican context, in an effort to contribute novel variables to the street-children literature. Using a cross-sectional survey design, 204 mothers of street-working and non-working children were interviewed within one community in Monterrey, Nuevo Leon, Mexico. Factor analysis was used to corroborate the internal construct validity of two dimensions of social capital: family social capital and community social capital. Findings reveal that culture can play an influential role in how social capital indicators are defined and measured.

Key words: children's street work, family social capital, community social capital, factor analysis

**********

Latin America faces a critical challenge as explosive urbanization, poverty, overcrowded cities, unequal distribution of wealth, and the effects of globalizing the market-oriented economy have contributed to an increase in the number of children who migrate to the streets to supplement their family's income as well as to survive. In many countries throughout the region, common catalysts like rapid urban population growth and urban poverty have prompted the numbers of street-working and street-living children to soar (Connolly, 1990; Peralta, 1995; De la Barra, 1998). While the United Nations Children's Fund (UNICEF) estimates that there are more than 100 million children who live and work on the streets in the developing world, Latin America is home to 40 million of the total street-children population (Covenant House, 1999).

To gain a deeper understanding of the movement of children into the streets to meet their basic human needs, researchers across multiple disciplines, including social work, psychology, sociology, public health, medicine and law, have sought to identify and measure the associated risk factors. Individual and familial precipitating factors, or microfactors, include such influences as school dropout and family poverty (Martinez & Silva, 1998; Raffaelli, 1996; Wittig, 1994). Structural influences, or macrofactors, include poverty, urbanization, external debt and inconsistencies between macroeconomic and social policies (Connolly, 1990; De la Barra, 1998; Fallon & Tzannotos, 1998). Community influences, or mezzofactors, are less clear within the literature, as the intrafamilial and family-community influences related to children's street work have largely been overlooked in prior studies. Rather, the traditional focal points with this population have been the intrapersonal and familial demographic risk factors, as well as the structural risk factors. To gain a more holistic understanding of this social phenomenon, it is vital for researchers to consider the mezzosocial influences as well, such as the nature of the relationships that occur between and among families, and how these may influence children's street work. Thus, this study focuses on the mezzosocial environment and on defining and measuring the relationships and interactions that transpire there.

In an effort to understand the effects of family- and community-based social relationships on an array of outcomes, the social capital framework has frequently been adopted as a means to further explore the intricacies of social interactions. According to Coleman (1988), family social capital refers to the relationships between parents and their children, which encompass the time, efforts, resources and energy that parents invest in their children. In contrast, exterior social capital--or community social capital--represents the family's interactions and relationships with the surrounding community, both with residents as well as with local institutions of socialization, such as schools (Putnam, 2000). Although at present there are no empirical precedents exploring the effects of social capital on the migration of children into the streets to work, considerable research does exist indicating the influence of social capital on children's well-being. Many of the oft-cited indicators of children's well-being are also correlated with children's street work.

The present study aims to conceptualize and operationalize the notion of social capital in Mexico for the purpose of introducing a new series of variables into the street-children literature to consider in future research. It constitutes part of a larger initiative to explore the relationship between levels of family and community social capital and children's street work in the Mexican informal economy. Given that Mexico is the geographical context for this study, and that existing social capital indicators have been largely developed in the United States and Canada, it was necessary to first consider the effect that a country's cultural context might have on the conceptual and operational definitions of constructs.

Empirical Review of Social Capital Literature: Systematic Review Method

To determine the scope of empirical literature related to social capital and children's well-being, the systematic review method (SR) was adopted (Larson, Pastro, Lyons, & Anthony, 1992). A variety of bibliographic databases (e.g., FirstSearch, OVID, Social Work Abstracts, Sociological Abstracts and Wilson) were searched from 1980 to the present, given that the majority of the social capital literature has been developed over the past two decades (Bourdieu, 1985; Coleman, 1988, 1990; Putnam, 1993, 1995, 2000). A manual search was also conducted in both dissertations and academic journals related to social capital over the past decade (1990 to the present). Finally, several annotated bibliographies and working-paper series related to social capital were consulted, which were compiled and produced by the Social Capital Initiative under the auspices of the World Bank, available online at: http://www.worldbank.org.

The methodology adopted to discern relevant and nonrelevant empirical literature concerning social capital consisted of four selection criteria. The study was included in the review cohort if: 1) it examined either family social capital and/or community social capital and the effects on individual and/or collective well-being; 2) it utilized quantitative, qualitative and/ or triangulation of research methods to assess levels of social capital; 3) it identified indicators of social capital at the family and/or community levels; and 4) it produced findings relevant to social work and/or social welfare policy. The SR method produced 22 peer-reviewed studies that complied with these criteria.

Review of Empirical Findings

For the review cohort, the findings are grouped within two general categories: 1) indicators of family social capital and 2) indicators of community social capital. The common variables across studies have been identified within each category.

Indicators of Family Social Capital

Eight of the 22 studies examined the effects of family social capital on outcomes related to children. Using Coleman and Hoffer's (1987) High School and Beyond study of 4,000 high school students as an empirical precedent, numerous subsequent studies have followed Coleman's initial operationalization of family social capital into five main components, namely: family structure, quality of parent-child relations, adult's interest in the child, parents' monitoring of the child's activities, and extended family exchange and support.

Family structure. All eight studies used family structure as a predictor of outcomes for children and youth. Across studies, high levels of uniformity exist among select indicators: single-parent vs. two-parent household, absence vs. presence of a paternal figure--either biological or stepfather, and both parents vs. one parent work(s) outside the home. Two-parent households were found to be consistently related to positive outcomes in successful social development among at-risk youth (Furstenberg & Hughes, 1995) and in successful physical and behavioral development among preschool children reared in unfavorable environments (Runyan et al., 1998). Three studies found two-parent households to be a buffer against youth at risk for dropping out of high school (Coleman & Hoffer, 1987; Teachman et al., 1996, 1997). Similarly, one study found two-parent households to be associated with lower levels of violent acts in youth (Johnson, 1999).

Quality of parent-child relations. Second, six of the eight studies sought to examine the quality of intrafamilial parent-child relations. As originally proposed by Coleman and Hoffer (1987), measuring the strength of parent-child relations reflects the quality of intrafamilial relationships in a given family. Common indicators of this construct include: number of times the parent helps the child with homework per week, number of sharing activities the parent and child participate in together per week, number of times per week the parent verbally encourages the child, and number of siblings in the household, which Coleman (1988) purports can dilute adults' attention to children. Three studies found that a higher frequency of social interactions between parents and children decreased the children's likelihood of dropping out of school (Coleman & Hoffer, 1987; Teachman et al., 1996, 1997), while one study found that higher levels of social interactions between parents and children were related to a lower likelihood that children fared negatively in future outcomes (Furstenberg & Hughes, 1995). Two studies found a significant relationship between a fewer number of siblings in the household and positive outcomes for children in their educational attainment (Coleman & Hoffer, 1987) as well as in their physical and behavioral development (Runyan et al., 1998).

Adult's interest in the child. Third, six of the eight studies assessed the adult's interest in the child. Common indicators for this component of family social capital were: the mother's academic aspirations for the child, the parents' levels of empathy for the child's needs, and the parents' involvement in the child's school-related activities. Parents' high expectations for children's school performance were found to be associated with positive outcomes for children in school as well as in social and behavioral development (Coleman & Hoffer, 1987; Furstenberg & Hughes, 1995; Runyan et al., 1998; Teachman et al., 1996). Further, high parental empathy towards children's needs were found to positively influence children's future outcomes (Furstenberg & Hughes, 1995; Runyan et al., 1998).

Parent's monitoring of the child. Five of the eight studies operationalized this component via the following measures: number of school meetings the parents attend; number of child's friends whom the parents know by sight, and number of child's friends' parents whom the parents know by sight. In three studies, high levels of parental monitoring of children's activities were consistently associated with positive outcomes in the educational attainment of children (Coleman & Hoffer, 1987; Teachman et al., 1996, 1997) and in the socioeconomic achievement of youth (Furstenberg & Hughes, 1995). Two additional indicators from the literature consist of knowing what the child is doing as well as with whom the child is when not at home (National Commission on Children, 1990). In a study examining how certain parental and peer-related risk and protective factors influence adolescents' school and emotional outcomes, Voydanoff and Donnelly (1999) found that both of these measures were related to positive outcomes, specifically, to better academic performance and to higher levels of psychological adjustment.

Extended family exchange and support. Finally, three of the eight studies explored the degree of extended family exchange and support. Coleman and Hoffer (1987), Furstenberg and Hughes (1995) and Stevenson (1998) adopted the following three indicators to measure this component: number of extended family members who lives in the home, number of interactions the child has with extended family members living in the home, and number of times the child visits extended family members living outside of the home. Two of the three studies found that high levels of social support from extended family members reduced the likelihood that children would drop out of school (Coleman & Hoffer, 1987) or experience depressive symptoms (Stevenson, 1998).

Indicators of Community Social Capital

Thirteen of the 22 studies examined community social capital and its effects on outcomes related to child welfare. Coleman and Hoffer (1987), in their seminal study, propose four general components of community social capital: 1) social support networks, 2) civic engagement in local institutions, 3) trust and safety, and 4) degree of religiosity. Each of the 13 studies assesses one or several of these components in relation to children's well-being, in addition to other elements of community social capital.

Social support networks. Findings from multiple studies reveal that increased parental social support can have positive effects on children's outcomes. For instance, two studies found that parents' increased relationships with schools and other parents decreased the likelihood that their children dropped out of school (Teachman et al., 1996, 1997). Putman (2000) found that the children of parents who were embedded in rich social networks were less likely to pursue gang membership, while Maccoby and colleagues (1958) discovered that the children of parents who had strong relationships with other parents were less likely to commit delinquent acts. Furstenberg and Hughes (1995) suggest that strong help networks for parents are related to favorable outcomes among youth in finishing school and attaining gainful employment. Likewise, high levels of social support for the primary maternal caregiver were associated with both positive behavioral outcomes for atrisk preschool children (Runyan et al., 1998) and lower levels of depression in at-risk teens (Stevenson, 1998). Common indicators of social support networks across studies include: number of mother's close friends and number of visits to these close friends per week.

Civic engagement in local institutions. Considerable empirical evidence indicates a positive relationship between parents' levels of civic participation in local community organizations and their children's overall well-being. Several studies found that there were more exchanges of resources and sharing of child-rearing responsibilities among families in neighborhoods that had higher levels of participation and activism (Garbarino & Sherman, 1980; Sampson et al., 1999). Also, Putnam (2000) cites several findings that in communities rated with high civic engagement, teachers reported higher levels of parental involvement in school-related activities and lower levels of student misconduct. The common measures of civic engagement consist of volunteering in a local group, serving as an active member of a local organization or club, participating in community meetings to solve local problems, organizing with neighbors to address local problems or to improve the neighborhood, and speaking with local politicians regarding neighborhood problems.

Trust and safety. Various studies explored parents' levels of neighborhood trust and safety in relation to children's well-being. Garbarino and Sherman (1980) discovered that mothers who felt safe in their surrounding environment were more likely to report a higher quality of life as well as to rate their neighborhoods as a more positive place in which to rear their children. Similarly, Sampson and colleagues (1999) found that parents' perceptions of vulnerability were lower in high-trust neighborhoods, while in these same neighborhoods, parents' willingness to assist their neighbors was higher. Putnam's (2000) centennial analysis of social capital trends also reveals that high social trust in neighborhoods can effectively break the link between social and economic impoverishment in neighborhoods and the delinquent activity by youths residing there. Across studies, the general measure of trust and safety was a single-item indicator, assessing the extent to which parents perceived that most people in the neighborhood can be trusted.

Degree of religiosity. As demonstrated in Coleman and Hoffer's (1987) original study, the frequency of attendance at religious services by families was found to be a strong predictor of the dropout rate among high school students. A decade later, Teachman and colleagues (1996, 1997) found that the child's attendance at a Catholic school--a related indicator of social capital that was originally proposed by Coleman (1988)--had significant and strong effects on reducing the likelihood of school dropout. Runyan and colleagues (1998) also found mother's regular church attendance to be a significant predictor of positive behavioral outcomes for at-risk preschool children.

Quality of school. Although not included as an indicator in Coleman and Hoffer's (1987) original study, several subsequent studies have used school quality as a component of community social capital. Furstenberg and Hughes (1995) found that high ratings, as perceived by adolescents, were strongly related to positive socioeconomic outcomes among the youths, such as graduating from high school, enrolling in college, attaining gainful employment, and remaining mentally and emotionally healthy. Similarly, Voydanoff and Donnelly (1999) discovered that parents' perceptions of high quality of the child's school were associated with positive outcomes in their children's educational achievement.

Quality of neighborhood. A final indicator used in 9 of the 13 studies in the review cohort, although also not included in Coleman and Hoffer's (1987) original study, consists of parents' perceptions of the quality of the neighborhood. Furstenberg and Hughes (1995) found that high neighborhood quality was a significant predictor of youths' future enrollment in college. Stevenson (1998) and Johnson (1999), on the other hand, discovered that poor neighborhood quality was associated with high levels of depression in youth and high rates of violent acts by youth, respectively. Finally, findings from multiple studies indicate strong support for neighborhood quality as a correlate of positive outcomes for children, including lower levels of child maltreatment (Garbarino & Sherman, 1980; Swanson Ernst, 2001); lower levels of youth delinquency (Maccoby et al., 1958); higher levels of children's physical and mental health (Morrow, 2000); and higher levels of educational attainment for children (Putnam, 2000). The most common indicators of neighborhood quality include parents' perceptions of the following: the neighborhood as a safe place to raise their children, the presence of any safe places in the neighborhood for children to play, and the extent of signs of underlying social disorder (e.g., litter, graffiti, abandoned buildings, gangs, etc.).

Method

Measures

Drawing from the empirical review, family social capital was expected to be comprised of five sub-factors: family structure, quality of parent-child relationships, adult's interest in child, parents' monitoring of child's activities and degree of extended family exchange and support. Similarly, community social capital was expected to consist of the following six sub-factors: quality of school, quality of neighborhood, social support networks, civic engagement, trust and safety, and degree of religiosity. All items selected to measure the latent constructs were derived from and defined by existing indicators of family and community social capital.

Participants

Two hundred and four families residing in the community of Genaro Vazquez, Monterrey, Nuevo Lepn, Mexico, participated in this study. Half of the sample (N=102) had at least one child between the ages of 6 and 16 years who worked in the streets and 50% (N=102) had children who did not work. Close to 90% of the families were dual-parent households and 80% of the families had between 5 and 16 members living in the home (mean = 6.2 members). Families, on average, earned $4300.90 pesos per month (US $430.09 in 2002). The primary source of employment for families in Genaro Vazquez was commercial and ambulatory sales. Families grew, packaged and sold a variety of food items, including: fruits, semillas (seeds), chile piquin (hot chili peppers), and tunas (fruits from the nopal cactus). Twenty two percent of the families were of indigenous origin, comprised of the Otomi, Mixteco and Nahuatal groups. Fathers in the sample were on average two years older than mothers, at 37.6 and 35.5 years, respectively. Mothers, on average, had 4.8 years of formal education while fathers had slightly more education, with 5.5 years.

Results

Factor analysis was performed to test the internal construct validity of the proposed factors by verifying which of the oft-cited indicators within the social capital literature were relevant to the Mexican culture and context. In the event that factor analysis produced results that were inconsistent with the theoretical definitions of family and community social capital (e.g., loadings were under the 0.40 cut-off for theoretically important variables), theory was given preference over statistical findings. This was the case given the study's aim to introduce novel, community-level variables into the street-children literature base for future study.

Family Social Capital

In an effort to ascertain whether all proposed indicators of family social capital were interrelated to one general construct, an initial factor analysis requesting one underlying factor was carried out on the original 22 indicators of family social capital. The extreme values for several variables as well as low-loading, individual scale items were eliminated prior to the analysis to achieve a more normal distribution among the data. Only 4 of the 22 variables, which proposed to measure family social capital, had loadings over .40. The one extracted factor accounted for 15.13% of the variability in the original variables, which is less than the amount that the four individual variables together would have explained. In order to achieve a more substantially explicative factor structure, the eigenvalues from this analysis were reviewed and subsequent factor analyses were conducted using the Direct Oblimin rotation method to account for inter-correlations among factors. Of these analyses, the three-factor structure produced the clearest explanation of the interrelationship among factors and indicators (see Table 1 below).

Together, Family Social Capital accounts for 35.59% of the total variance in the original variables and proposes to measure the time, efforts and resources that parents invest in their children. Figure 1 below illustrates the obtained three-factor structure. The construct, Family Structure, is comprised of 5 manifest variables; Adult's Interest in Child includes 12 indicators; and Monitoring of Child consists of 2 variables.

[FIGURE 1 OMITTED]

Regarding the correlations among the three factors, the coefficient between Family Structure and Adult's Interest in Child was -.12; between Family Structure and Monitoring of Child was -.18; and between Adult's Interest in Child and Monitoring of Child was -.13.

Community Social Capital

Factor analysis was then performed on 15 variables, which were proposed to measure Community Social Capital. Individual scale items with low factor loadings as well as the extreme values for several variables were deleted prior to the analysis in order to reach a more symmetrical distribution among the data. Upon selecting one general factor, only 3 of the 15 indicators had loadings over .40. The extracted factor explained 16.67% of the total variance in the original variables. In pursuit of a factor structure that explained more of the total variation, the eigenvalues from the initial analysis were examined and subsequent factor analyses were carried out using Direct Oblimin rotation. The two-factor structure presented in Table 2 resulted in the best depiction of the interrelationship among factors and variables.

Community Social Capital accounts for 32.10% of the data's variation and purports to measure the family's interactions and relationships with the surrounding community, both with residents as well as with local institutions of socialization. Figure 2 presents the obtained two-factor structure. Neighborhood Connections includes five indicators, whereas Neighborhood Perceptions is comprised of four variables. The two factors had a correlation of less than 0.1, indicating that they are virtually orthogonal in nature.

[FIGURE 2 OMITTED]

Discussion

Drawing from social capital theory, this study sought to test the relevance of existing conceptions of social capital, largely from the United States and Canada, in the Mexican context, in an effort to contribute novel variables to the street-children literature. To date, prior research suggests that street-working children are more likely to come from impoverished families, who reside in disadvantaged neighborhoods. However, it remains unclear what, specifically, about such families and communities can precipitate children's street work. This gap in the existing knowledge presents researchers with an opportunity to explore additional aspects of families and communities that may also be important determinants of children's informal street labor.

Findings from the factor analysis indicate that Family Social Capital, in the Mexican context, was comprised of three intercorrelated factors: Family Structure, Adult's Interest in Child and Parents' Monitoring of Child. With regards to family structure in Mexico, one notable difference between the social capital literature review and the obtained factor structure concerns the variable: number of extended family members residing in the home. Findings from studies conducted largely in the United States and Canada suggest that the number of extended family members is one of several indicators of a separate factor outside of the immediate family structure (i.e., extended family exchange and support). However, given that a third of the families in the sample (33%) had between one and seven relatives living in the home, it is understandable why this indicator loaded onto the factor Family Structure in this study. Although the factor loading was considerably low (.32) in comparison to other indicators measuring Family Structure, it challenges traditional conceptions of immediate family structure as restricted solely to members of the biological family. Further, such a finding serves as a caveat to utilizing existing conceptualizations of immediate and extended family as separate factors and highlights the accompanying need to develop a more inclusive and holistic definition of "family structure"--one that is grounded in cultural relevance.

Similarly, the clear dividing line between the residence of the "immediate family" and that of the "extended family," often found in the United States, was less apparent in many Mexican families in this study. In 97% of the families, close relatives (i.e., grandparents, aunts, uncles and cousins) lived next door, on the same block or close by. Thus, interactions with extended family were often a regular occurrence, which likely explains why the variables, number of activities with extended family living in the home and number of visits to extended family members living outside of the home, loaded moderately onto the factor, Adult's Interest in Child. Again, extant findings from studies performed in the U.S. and Canada suggest that such variables form a separate factor related to extended family, perhaps due to less geographical proximity between the immediate and extended family. In the Mexican context, however, children's activities and visits to kith and kin residing within and outside of the home often occurred daily. Future studies in the Mexican context would benefit from exploring the strength of the extended family as an indicator of family structure, rather than as a separate factor outside of the immediate family.

For the factor, Adult's Interest in Child, results reveal that indicators that were originally speculated on the basis of the social capital literature to measure multiple factors in essence loaded onto one factor only. In this study, this seems reasonable, given that the indicators selected all manifest ways in which adults demonstrate interest in their children. Two interesting cultural observations can be made with respect to these findings. First, the variable, place of mother's work, resulted as an indicator of Adult's Interest in Child rather than as an indicator of Family Structure, as anticipated. Given that 57% of the mothers in the sample worked inside of the home, often preparing food items for sale throughout the neighborhood, it is evident how being present in the home could be positively associated with the quantity of time mothers had to spend with their children. Secondly, knowing whom the child was with as well as what the child was doing when not at home loaded moderately onto the factor Adult's Interest in Child, rather than onto Monitoring of Child. Parents who were home with their children, or who had in-home relatives (or close-by relatives) interacting daily with their children, or who participated in daily activities (i.e., homework, play, errands, etc.) with their children, were more likely to be aware of where--and with whom--their children were when not at home.

With reference to the third construct, Parents' Monitoring of Child, two variables demonstrated high loadings on this factor: number of child's friends and number of child's friends' parents whom the mother knows. The negative loadings of both indictors are contrary to the anticipated relationship among indicators and the latent factor. This suggests that either the indicators used in this study were not valid measures of the construct, or rather, a more suitable factor label was necessary to explicate the relationship among variables and the factor. One possible explanation that may elucidate this unanticipated finding concerns the lack of a clear division between immediate and extended family, as discussed above. Because extended family members frequently lived within the home and/or nearby, children often played with siblings and relatives (i.e., cousins, nieces and nephews). Of the families in the study, 92% had between two and nine children under the age of 18 living in the home, including both biological and non-biological children. During the interviews, when mothers were asked how many of their child's friends they knew by sight, it was common to hear responses such as: "My child's friends are his/her siblings and cousins" or "I know all of my child's friends. They are all family." This may help clarify why such a high percentage of mothers knew both their children's friends and their children's friends' parents: 94% of mothers knew some or all of their children's friends, and 87% knew some or all of the parents of their children's friends.

Further, the number of activities per month that children participated in with extended family members residing in the home was significantly and moderately correlated with both the number of the child's close friends (r = .27, p< 0.05) and the number of the child's close friends' parents (r = .24, p< 0.05) whom the mother knew. It is likely that mothers in this study were interpreting their children's "friends" as those people outside of the immediate and extended family, and "family" as those individuals who form part of the immediate and extended clan. This, in turn, may help explain why the variables, number of child's friends and number of child's friends' parents whom the mother knows, failed to load onto the factor Adult's Interest in Child, and instead, loaded strongly and negatively onto a separate factor, Monitoring of Child. Parents' unfamiliarity with their children's friends and their children's friends' parents outside of the family may not, then, have been an indication of their lack of family social capital (i.e., ability to monitor their children), but rather, a reflection of underlying ties among immediate and extended family, which may have gone undetected using indicators developed from U.S. and Canadian realities.

For the second dimension of capital, Community Social Capital, two factors emerged, rather than the six variables that were originally speculated to exist on the basis of the social capital literature. The factor, Neighborhood Connections, was similar to the construct, social support networks, from the social capital literature, with one exception. Civic engagement, in the Mexican context, was an indicator of Neighborhood Connections, rather than a separate entity. A brief explanation of neighborhood politics in Genaro Vazquez may help clarify this finding. Within the community, residents who are involved in neighborhood associations and local politics are voted into their positions by other residents. As such, being embedded in rich networks of other community members is likely an advantage to securing an elected position in neighborhood associations. The significant, positive relationship found in this study between involvement in social support networks and participation in civic associations (r = .30, p<0.0001) is also consistent with findings from the existing social capital literature (Onyx & Bullen, 2000).

The second obtained factor, Neighborhood Perceptions, was comprised of three variables from the social capital literature, which measured quality of the neighborhood, as well as a fourth variable, trust and safety. Previous studies on social capital suggest that trust and safety is a separate factor; however, in this study, this composite-score variable from Onyx and Bullen's (2000) Social Capital Scale, loaded strongly onto Neighborhood Perceptions. One explanation for the high inter-correlations among trust and safety and the other neighborhood indicators found here is that many families residing in Genaro Vazquez worked as commercial and ambulatory salespeople throughout the community. It is possible that mothers who spent a considerable amount of time selling products in the neighborhood may likely be more aware of and comfortable with their surrounding environment as well as interact regularly with their neighbors. High levels of trust in neighbors would thus likely be associated with more positive perceptions of the neighborhood as a safe place to raise children and lower ratings of neighborhood problems. Development of social capital theory could certainly benefit from further exploration of the relationship among people's daily presence in the neighborhood, levels of trust, and perceptions of neighborhood quality, especially in regions where ambulatory labor within the informal economy is prevalent.

Finally, school quality and degree of religiosity were also anticipated to form part of Community Social Capital, on the basis of extant literature. Conversely, in this study, neither indicator loaded onto a general factor, nor formed a separate factor. Church attendance was high among families in the study altogether, yet displayed a very low loading on Neighborhood Perceptions. Only 8% of mothers proclaimed no religious preference at all (71% were Catholic and 20% were Protestant). Seventy percent of families attended religious services at least once per month, while 44% of these families attended services at least once a week. One cultural observation that may help explain the discrepancy between degree of religiosity as an indicator of social capital in the literature and its low loading in this study concerns the ubiquitous presence of churches in Genaro Vazquez, along with the accompanying opportunities for socialization provided by places of worship. In the absence of public and private secular social services in the community, many families likely partake in church activities and services (e.g., marriage counseling, clothing drives, food pantries, youth groups, sports teams and field trips) as a means to meet their needs. In this case, both mothers who spent considerable time in the neighborhood as well as those who were not as present outside of their homes, likely attended church services and activities, although higher trust in neighbors was positively associated with more frequent attendance at religious services (r = .13, p < 0.05). Both types of mothers were thus able to benefit from the social support and access to resources available to them. Future studies in the Mexican context could elucidate this finding by focusing on the role of religious institutions, not only as a place of worship (i.e., to measure religiosity), but also as an institution that meets the comprehensive needs of residents (i.e., to measure service provision).

Precise interpretation of the results depends upon prior analysis of the study's limitations. The reliability of the results from factor analysis is contingent upon the use of a large sample size, as well as the presence of moderate-to-high factor loadings for each manifest indicator (Mertler & Vannatta, 2001). Although the findings presented here clearly deviate from both of these criteria, social capital theory was given preference in selecting the indicators included in the obtained models of family and community social capital. By continuing to strengthen the conceptual and operational definitions of social capital, future studies can explore the relationship between social capital and children's street work in an effort to uncover new family- and community-level correlates of child street labor.

To date, there is a dearth of knowledge regarding the individual and collective effects of various dimensions of social capital on children's street work. Similarly, there are few empirical precedents identifying specific family- and community-based influences of child street labor within the mezzosystem. Rather, most studies have focused on the characteristics of the children, themselves, as child laborers. This, in turn, has created an unrealistic and unidimensional perception of these children as individuals--disconnected from their families, schools and communities--who work and reside in the streets (Ennew & Milne, n.d.). Adopting a social capital theoretical framework, future studies can remove the street-working children phenomenon from the street environment and instead, focus primarily upon the other dimensions of the children's lives, that is: home, school, and community life, in addition to the children's work lives. By re-inserting child street workers within the context of their families and communities, future studies will provide a more holistic and realistic account of these children's lives and of the lives of their families and communities. In the event that such studies can empirically demonstrate the strength of family and community influences on children's street work, a useful measure will exist to guide street-children organizations around the world in moving from a palliative approach to a preventive one to address the root causes of children's street work--in the family and the community.

Table 1

Family Social Capital

                                                 Factor *

                                     Family      Interest    Monitoring
Variable                            Structure    in Child     of Child

Mother's relationship to child         .711
Years living with child (mother)       .729
Father/partner's relationship to
  child                                .873
Years living with child (father)       .870
Place of work: mom **                             -.232
Place of work: father                 -.201
# siblings in home                                -.324
Help child with homework                           .258
Verbally encourage child                           .276
Sum sharing activities with
  child                                            .393
Sum school-related interactions                    .212
Mother's academic aspirations                      .301
Parents' empathy                                   .373
# school meetings attended            -.113
# child's friends                                              -.757
# child's friends' parents                                     -.941
Know whom child with                               .541
Know what child doing                              .465
# extended family in home              .316
# activities w/ in-home extended
  family                                           .455
# visits to extended family
  out-of-home                                      .322

* In the interest of parsimony, only the highest loading is
displayed.

** The variables, place of mother's and father's work, were
coded as follows: 0=in home; 1=out-of-home.

Table 2

Community Social Capital

                                            Factor *

                                  Neighborhood    Neighborhood
                                   Connections     Perceptions

Sum quality of school                 .142
Neighborhood grade                                    .449
Safe places                                           .411
Sum neighborhood problems                            -.216
Sum mom's social networks             .245
# mom's close friends                 .297
# mom's visits to friends             .268
Sum neighborhood connections          .968
Sum civic engagement                  .234
Sum trust and safety                                  .841
# times attend church                                 .205

* Only the highest loading for each indicator is displayed.

References

Bourdieu, P. (1985). The forms of capital. In J. G. Richardson (Ed.), Handbook of theory and research for the sociology qf education (pp. 241-258). New York: Greenwood.

Coleman, J. (1988). Social capital in the creation of human capital. In P. Dasgupta & I. Serageldin (Eds.), Social capital: A multifaceted perspective (pp.13-39). Washington, D.C: World Bank.

Coleman, J. (1990). The foundations of social theory. Cambridge: Harvard University Press.

Coleman, J. S., & Hoffer, T. B. (1987). Public and private schools: The impact of communities. New York: Basic.

Connolly, M. (1990). Adrift in the city: A comparative study of street children in Bogota, Colombia, and Guatemala City. In N.A. Boxhill (Ed.), Homeless children: The watchers and the waiters (pp. 129-149). Binghampton, NY: Haworth.

Covenant House (Casa Alianza). (1999). Street children: Living in the streets. Retrieved on September 29, 2000, from http://www.casa-alianza.org.

De la Barra, X. (1998). Poverty: The main cause of ill health in urban children, Health Education and Behavior, 25(1), 46-59.

Ennew, J., & Milne, B. (n.d.). Methods of research with street and working children: An annotated bibliography. Retrieved September 22, 2000, from http://www.rb.se/childwork/index.htm.

Fallon, P. & Tzannotos, Z. (1998). Child labor: Issues and direction for the World Bank. Washington, D.C: World Bank.

Furstenberg, F. F., & Hughes, M. E. (1995). Social capital and successful development among at-risk youth. Journal of Marriage and the Family, 57, 580-592.

Garbarino, J., & Sherman, D. (1980). High-risk neighborhoods and high-risk families: The human ecology of child maltreatment. Child Development, 51, 188-198.

Johnson, S. D. (1999). The social context of youth violence: A study of African American youth. International Journal of Adolescent Medicine and Health, 11(3-4), 159-175.

Larson, D., Pastro, L., Lyons, J., & Anthony, E. (1992). The systematic review: An innovative approach to reviewing research. Washington, D.C.: Department of Health and Human Services.

Maccoby, E., Johnson, J., & Church, R. (1958). Community integration and the social control of juvenile delinquency, Journal of Social Issues, 14, 38-51. Martinez, V., & Silva R. (1998). Familias desintegradas, ninos de la calle. Trabajo Social, 20, 47-52.

Mertler, C. A., & Vannatta, R. A. (2002). Advanced and multivariate statistical methods. Practical application and interpretation (2nd ed.). Los Angeles, CA: Pyrczak Publishing.

Morrow, V. (2000). 'Dirty looks' and 'trampy places' in young people's accounts of community and neighbourhood: Implications for health inequalities. Critical Public Health, 10(2), 141-152.

National Commission on Children. (1990). Survey of parents and children, United States [Computer file]. [CPSR version. Washington, DC: Child Trends Inc./ Princeton, N J: Princeton Survey Research/Ann Arbor, MI: DataStat[producers], 1990. Ann Arbor, MI: Inter-University Consortium for Political and Social Research [distributor], 1999.

Onyx, J., & Bullen, P. (2000). Measuring social capital in five communities. The Journal of Applied Behavioral Science, 36(1), 23-42.

Peralta, F. (1995). Status report on street children in Mexico City after the 1994 financial crisis. In N. Tello (Ed.), Redisenando el futuro: Retos que exigen nuevas respuestas (pp.75-87). Mexico: D. R. Plaza & Valdes, UNAM.

Putnam, R. D. (1993). The prosperous community. Social capital and public life. The American Prospect, Spring, 35-42.

Putnam, R. D. (1995). Tuning in, tuning out: The strange disappearance of social capital in America. The 1995 Ithiel de Sola Pool Lecture. Political Science and Politics, 664-683.

Putnam, R. D. (2000). Bowling alone: The collapse and revival of American community. New York: Simon & Schuster.

Raffaelli, M. (1996). The family situation of street youth in Latin America: A cross national review. Covenant House/Casa Alianza working paper. Florida, USA: Casa Alianza.

Runyan, D. K., Hunter, W. M., Socolar, R. S., Amaya-Jackson, L., English, D., Landsverk, J., Dubowitz, H., Browne, D. H., Bangdiwala, S. I., & Mathew, R. M. (1998). Children who prosper in unfavorable environments: The relationship to social capital. Pediatrics, 101(1), 12-18.

Sampson, R. J., Morenoff, J. D., & Earls, E (1999). Beyond social capital: Spatial dynamics of collective efficacy for children. American Sociological Review, 64(October), 633-660.

Stevenson, H. C. (1998). Raising safe villages: Cultural-ecological factors that influence the emotional adjustment of adolescents. Journal Of Black Psychology, 24(1), 44-59.

Swanson Ernst, J. (2001). Community-level factors and child maltreatment in a suburban county. Social Work Research, 25(3), 133-142,

Teachman, J. D., Paasch, K., & Carver, K. (1996). Social capital and dropping out of school early. Journal of Marriage and the Family, 58, 773-783.

Teachman, J. D., Paasch, K., & Carver, K. (1997). Social capital and the generation of human capital. Social Forces, 75(4), 1343-1359.

Voydanoff, P., & Donnelly, B.W. (1999). Risk and protective factors for psychological adjustment and grades among adolescents. Journal of Family Issues, 20(3), 328-349.

Wittig, M. C. (1994). Culture of poverty or ghetto underclass?: Women and children on the streets of Honduras. Unpublished doctoral dissertation, Tulane University, New Orleans, LA.

KRISTIN M. FERGUSON

University of Southern California

School of Social Work