Introduction

A robust literature identifies social support as a protective factor that promotes health and well-being among individuals across the life-span. For parents, social support shapes expectations, imparts knowledge and diminishes stress related to parenting [5] to ultimately promote reliance on positive and effective parenting practices while decreasing reliance on punitive parenting practices [9, 10, 59, 81]). But despite wide recognition of social support as a protective factor, few empirical studies have attended to the relations between social support and parenting in Latino families [3]. As an immigrant population and one disproportionately impacted by social disadvantage, social support may be especially important for Latino parents who must navigate a new and often unfamiliar culture in the absence of adequate socioeconomic resources [18, 29, 40, 58]. The aim of the present study was to examine social support as a protective factor for the early childhood functioning of Mexican- and Dominican-origin young children.

Social Support

Social support is a complex and multidimensional concept that may be understood in terms of a social network. Research identifies several important aspects of social support, including the size of the social network, members of the network, and the amount and types of support (emotional, instrumental) received through the network [3, 26, 66]. Of these, individual perceptions of the assistance one can rely on from family and friends (i.e., emotional and instrumental social support) appear to be the most beneficial to maternal physical and mental health [18, 25, 29, 54, 64, 66, 83].

Social Support and Child Development

The benefits of social support appear to extend beyond maternal well-being to children. Several studies have found direct associations between mothers’ social support and the emotional and behavioral functioning of their children [14, 53, 64, 69]. For example, children of mothers with limited social support appear to be at higher risk for social withdrawal, depression and hyperactivity [10, 80]. Moreover, the adverse effects of social isolation seem long-lasting. Studies have documented lower cognitive abilities and more conduct problems in children of mothers who received limited social support during pregnancy [1, 68]. These effects are believed to be mediated by parenting practices [18, 40, 58] in that mothers who experience support are expected to be more engaged and responsive in their parenting, and these positive parenting practices would promote better child development. The present study examines this model of social support and child development (see Fig. 1), in which Latino mothers with high levels of perceived emotional and instrumental social support are more effective in their parenting, with positive effects for the mental health functioning of their children [5]. In examining these hypothetical links with Latina mothers, we consider the ways in which experiences of social support are shaped by cultural norms and values.

Fig. 1
figure 1

Conceptual model of social support and child functioning, through parenting practices

Full size image

Familial and Extra-Familial Social Networks

In the Latino population, reliance on the family unit arises from the cultural value of familismo, specified as familial interconnectedness, belief that family comes before the individual, belief in family reciprocity, and belief in familial honor [48]. Latinos engage in high levels of behaviors that reflect these beliefs, including the formation of large family networks with whom they share housing, daily living activities and childrearing [4, 13, 32, 74, 83]. Given the strength of the interpersonal ties and mutual exchange between family members, familismo generally leads to high levels of support [15, 20, 61, 63, 76].

Support outside of the family context may be less common among immigrant Latino populations. For Latina mothers of young students, the school community offers unique access to a naturally occurring network of parents. But while the formation of social networks in schools is normative among middle class families [35], barriers related to language and documentation status, along with a lack of familiarity with the US school system, may make it unlikely for Latina mothers to create or join networks at the school level. In addition, the strong ties between extended family members dictated by familismo may interfere with the formation of non-kin networks. That is, Latinos may hesitate to form networks with non-family members either because their needs can be met within the family support system or because of a reluctance to form attachments with non-family members [16]. Still, no study to date has simultaneously examined support received by Latina mothers from family and parent-to-parent networks in schools. To address this gap, an aim of the present study was to describe the level of perceived emotional and instrumental support Latina mothers receive from the school community, relative to levels of support received from family networks, both of which are expected to be linked to parenting.

Parenting in Latino Populations

Research across cultures confirms the universal importance of parenting for children’s optimal development such as better self-esteem and lower aggression [42]. In the Latino population, research suggests that parenting is characterized by high levels of warmth [11, 12, 21] but also a greater reliance on harsh parenting compared with non-Latino white parents [24, 43]. The effects of these harsh parenting practices may be culture-specific. Specifically, it appears that while positive parenting practices that reflect responsiveness, acceptance and warmth contribute to healthy child outcomes across cultures [42], the effects of parental discipline on child development may vary across cultures (e.g., [45]). Consistent with the cross-cultural literature, studies with Latino families in the US show promoting effects of positive parenting, but mixed effects of harsh parenting, on children’s internalizing and externalizing behavioral functioning [12, 34, 39, 40, 46, 65]. That is, some studies have found an attenuated or null association between harsh parenting and child behavior problems, aggression and cognitive development in Latino families [6, 7, 52, 67]). These findings have been interpreted according to the cultural normativeness hypothesis, which argues that when physical discipline practices (e.g., spanking) are perceived as normative, they serve as less robust predictors of child functioning [45]. Among Latinos, the use of physical punishment in childrearing has been described as consistent with cultural values such as respeto (i.e., respect) and as common practice [12, 13].

The Link Between Social Support and Parenting

A number of studies have linked mothers’ social support to parenting but more research is needed to understand how multiple domains of social support relate to parenting practices. On the one hand, social support in general has been found to diminish the use of harsh discipline practices and to increase the use of positive parenting practices of responsiveness, acceptance and warmth in parents of all backgrounds, including low-income and Latina mothers [18, 40]. On the other hand, to the extent that social networks influence parenting practices by giving advice [46], it is possible that associations vary based on the source of support. For example, support from family may reinforce the use of parenting practices that are congruent with Latino culture (e.g., physical punishment), whereas support from a parent network in a U.S., public school may reinforce the use of practices (e.g., non-physical punishment such as time out) that are not necessarily rooted in Latino culture [40, 50]. In other words, to the extent that networks in schools include non-immigrant parents (whether Latino or not), exposure to parenting norms, parenting advice and “policing” of parenting practices may be driven by more mainstream norms and beliefs, such as disapproval of spanking [50]. Thus, in the present study, we considered how support from family versus school networks may be uniquely associated with positive parenting practices as well as harsh parenting practices.

School-based social networks may also facilitate connections between mothers and the school itself, making it more likely for mothers to get involved in their child’s schooling. Parent involvement in education [51, 73, 77], or the resources that parents dedicate to their child’s learning experience [30], has been consistently linked to children’s behavioral, socioemotional and academic functioning [22, 37, 70]. However, while Latino parents appear to be highly invested in their children’s education [19, 71], they do not often attend school events, volunteer in the classroom or communicate directly with school staff [47]. Barriers such as low parental education, conflicting work schedules, and different home languages help explain these lower levels of school-based involvement [49]. Additionally, school-based parent involvement, which depends in large part on the collaborative relationship between parents and school staff, is lower when there is a cultural gap between school staff and parents [38, 41]. When parents perceive a cultural mismatch, particularly racism, within the school setting [51, 73, 77], they are less likely to participate in school-based activities. Support from other parents within the school may help to close the cultural gap between Latina mothers and the school community, reducing barriers to school-based involvement practices. We considered this hypothesized link between school-based social support and school-based parent involvement in the present study.

The Latino Population in the US

As an immigrant and ethnic minority group in the US, Latinos face a number of considerable challenges that may undermine parental well-being and healthy child development [17]. Nearly one in five Latinos in the US lives in poverty, and rates are even higher among children [72]. Experiences of discrimination and acculturative stress can further disadvantage Latino families. While all Latino ethnic groups experience considerable social and economic disadvantages, there are important differences between groups. In New York City (NYC; where the present study took place), the Dominican-origin population is well-established, having served as a source of migration to the city for multiple generations. Recent Dominican immigrants typically settle into ethnic enclaves, facilitating the acculturative process and granting access to large and multigenerational family networks within the Dominican community [82]. In contrast, the Mexican-origin population is relatively new to NYC, and there are fewer ethnic enclaves available to new immigrants, potentially exacerbating the sense of isolation experienced by newcomers. Demographically, the Mexican community in NYC has lower levels of acculturation (i.e., English language competence, US cultural knowledge), formal education and employment as well as higher levels of poverty [82]. These distinct social and demographic profiles—which are likely to have implications for the social support each group receives—underscore the critical need for studying parenting and child development with attention to ethnic subgroup differences.

The Present Study

As reviewed above, social support, as experienced by mothers, is associated with better child developmental outcomes, making the study of social support of great potential utility in understanding protective processes in Latino families, who often experience considerable risk related to poverty and other social stressors (e.g., discrimination) [46]. With this overarching goal, the present study focused on the emotional and instrumental social support perceived by Latina mothers and its role in early childhood developmental functioning. Our first aim was to describe the level and source (family, school) of emotional and instrumental social support experienced by mothers from two Latino subgroups (Mexican, Dominican). Given the ecological context of each group, such as the rich enclaves Dominican American (DA) mothers in NYC are likely to live in and the higher levels of acculturation expected for DAs, we expected social support levels from both types of networks to be higher for DA mothers relative to Mexican American (MA) mothers. Our second aim was to test a path model of social support and child outcomes in MA and DA families of young (i.e., 4–5 year old) children. We focused on child functioning outcomes at home and school because these are the two settings where young children spend most of their time, and we focused on adaptive behavior and problem behavior outcomes to provide a balanced depiction of early childhood functioning that acknowledges children’s strengths and deficits. In our model, we examined how social support from family and school networks was associated with child outcomes indirectly, through parenting practices and parent involvement in education. We hypothesized that social support from school networks would be associated with greater school-based parent involvement practices, more positive parenting and less harsh discipline. We made no directional hypothesis about the association between support from family networks and parenting because while the literature shows less frequent use of harsh discipline among mothers who receive support, family members within Latino culture may reinforce the culturally-sanctioned use of harsh discipline. Also because of the culturally-sanctioned use of harsh discipline [45], we made no hypothesis about its association with child functioning. We did however expect positive parenting and parent involvement to be associated with better child functioning (i.e., more adaptive and less problem behavior).

Methods

Participants

Participants were drawn from a longitudinal study to examine the early childhood development of MA and DA children (N = 750). Mothers who self-identified as MA or DA and had a child in pre-kindergarten (pre-k) or kindergarten in one of 24 public elementary schools in NYC were eligible to participate. The analytical sample used in the present study included participants who had complete study data at Time 1 (N = 610; 81 %). There were no differences on demographic or study variables between the families that were included and those that were not.

The children in the study were on average 4 years (SD = 0.58) and were evenly distributed across gender (49 % boys) and grade (43 % in pre-k). The MA (n = 344) and DA (n = 266) samples differed, however, on most demographic characteristics, as shown in Table 1. Compared to DA mothers, MA mothers were younger, more likely to be poor, less likely to have graduated from high school, and less likely to be working for pay. MA children were more likely to live in a two-parent home and in a Spanish-speaking home environment.

Table 1 Sample characteristics by ethnic group
Full size table

Measures

Demographic Characteristics

Mothers provided information about their family’s demographic characteristics including age (mother and child), country of birth (mother and child), educational and occupational status, marital status, household income, length of residence in the US, and language used in the home.

Social Support

The Multidimensional Scale of Perceived Social Support (MSPSS) [84] measures perceived social support from distinct social networks including family and friends. For the purposes of the present study, we changed the reference from “friends” to “other parents in your school community” to assess support drawn from a parent network specifically from the child’s school. The family support scale had eight items of emotional (e.g., “I get the emotional help and support I need from my family”) and instrumental (“I can count on relatives when things go wrong”) support and an alpha of 0.88 for MA and 0.91 for DA mothers. The school community support scale had four items of emotional (e.g., “There are parents at this school who are real source of comfort to me”) and instrumental (e.g., “There are parents at this school who are around when I am in need”) support and an alpha of 0.87 for MA and 0.88 for DA mothers. Items were rated on a seven-point scale and averaged to obtain a family support and a school network support score.

Parenting Practices

To assess harsh parenting, we drew from the Parenting Practices Interview (PPI; Webster-Stratton, [78]) and the Parenting Styles and Dimensions Questionnaire (PSD) [62]. Fourteen items that measured the use of physical punishment (e.g., spanking) and harsh verbal punishment (e.g., yelling, criticizing, threatening) were included. Across both measures, items were rated on a five-point scale, with higher scores reflecting more harsh discipline. The alpha coefficient for the scale was 0.75 for both MA and DA mothers.

To assess positive parenting, we again drew from the PPI and the PSD. Twenty-five items that measured positive maternal behaviors including warmth (e.g., “I give comfort and understanding when my child is upset”) and the use of positive reinforcement (e.g., “I compliment my child when he behaves well”) were included. Across both measures, items were rated on a five-point scale, with higher scores reflecting more positive parenting. The alpha coefficients were 0.84–0.85 for MA and DA mothers, respectively.

The Involve-T [79] was used as a measure of mothers’ parent involvement. Teachers rated mothers’ school-based involvement activities such as frequency of contact with the school, including informal conversations on the child’s schooling (e.g., “Has this child’s parent stopped by to talk to you in the past 2 months?”) and attendance at school meetings and events (e.g., “How often has this child’s parent been to school meetings in the past 2 months?”). These ten items were rated on a five-point scale and averaged to obtain a total score, with higher levels suggesting more involvement. The alpha coefficients were 0.7–0.86 for MA and DA mothers, respectively.

Child Functioning

The Behavior Assessment System for Children-2 (BASC-2) [60] is a widely–used standardized measure of childhood externalizing problems (e.g., aggression, hyperactivity), internalizing problems (e.g., anxiety, depression, somatization), and adaptive behaviors (e.g., adaptability, social skills, functional communication). The BASC-2 has both a parent report form (PRS) and a teacher report form (TRS) and is available in Spanish as well as English. The Spanish form demonstrated adequate psychometric properties with the subsample of 311 Latino children and adolescents (including 82 preschoolers) who participated in the standardization study of the BASC-2. Past studies of the BASC also suggest that its factors are cross-culturally robust [8]. In the present study, we used the Total Behavior Problem composite scale, which includes externalizing and internalizing symptoms, and the Adaptive Behavior composite scale of child functioning at home (based on the PRS) and at school (based on the TRS). BASC-2 scales have a mean of 50 and a standard deviation of ten. Scores above 65 are considered clinically significant. The scales showed adequate internal consistency with both ethnic group samples in the present study (0.83–0.95).

Procedure

Recruitment took place in 24 New York City public schools that had pre-k and kindergarten classrooms serving MA or DA children. Families were recruited at the beginning of the school year, when bilingual research staff were present at school events and daily drop-off and pick-up to inform parents of the study. Parents who enrolled (79 % of eligible participants) were asked to consent to a parent interview, teacher ratings of family and child functioning and child assessments (not considered in the present study). Parents participated in an in-person interview in their language of choice (i.e., Spanish or English), and majority (88 %) were interviewed in Spanish. Interviews lasted approximately 90 min and included measures of parenting and child functioning. Mothers were paid $35, for their participation.

Teachers of study children were asked to complete an assessment packet that included measures of child functioning. As an incentive to participate, teachers were offered help in the classroom (e.g., preparing materials for bulletin boards or classroom activities) by research staff. The vast majority (92 %) of teachers agreed to participate, and there were no significant differences on any study variables between children with and without teacher data. All data used in the present study came from the first time point (i.e., in the fall of pre-k or kindergarten) in this longitudinal study.

Analytic Approach

Before conducting analyses, we examined clustering effects because 63 % of the teachers (126 out of 199 teachers) provided ratings on multiple students [the average number of students rated by each teacher was 2.80 (SD = 2.27)]. We calculated design effects [1 + (average group size − 1) × intraclass correlation coefficient] and followed guidelines suggested by Muthén and Satorra [56] to determine whether traditional statistical techniques could be employed without concern for bias from the clustered nature of the sampling design. In our sample, the design effects for teacher rated variables were all <2.0, suggesting that traditional statistical techniques could be used. Next, we confirmed that all endogenous variables in the model were normally distributed.

To test the conceptual model (Fig. 1), we conducted structural equation modeling (SEM) using MPLUS 6 [55]. To judge the closeness of fit for the hypothesized model, three indices were used as recommend by Muthen and Muthen: chi square (χ² > 0.05), root mean square error of approximation (RMSEA <0.05), and comparative fit index (CFI >0.95). We tested the conceptual model using maximum likelihood estimation method (ML). To examine mediation paths, we examined indirect effects using MPLUS. To consider ethnic group differences (MA and DA), we first conducted multigroup SEM analyses to determine whether there were statistically significant subgroup differences in model fit. In multigroup SEM, the first step is to test the nonrestricted model in the two groups by allowing all path values, means, variances, and covariances to be freely estimated. If there is evidence of fit, the next step is to examine a less restricted model by constraining path estimates to be equal in all groups, but allowing means, variances, and covariances to be free. More constraints (i.e., on means, variances, and covariances) can be imposed in subsequent steps if the restricted model does not cause a significant decrement in model fit. If there is insufficient evidence of fit in the least restricted multigroup SEM model (in which path values, means, variances, and covariances are all freely estimated), then the SEM model is tested separately for each group (i.e., MA and DA). In all analyses, we adjusted for potential confounders, including family poverty, marital status and maternal educational status.

Results

Preliminary Analyses

Table 2 presents the means, standard deviations, and correlations for the study variables for each ethnic group. All mothers reported high levels of family support (M = 5.64–6.08 on a scale of 1–7) relative to school support (M = 3.55–3.68). In addition, relative to MA mothers, DA mothers reported higher levels of family support, but the groups did not differ in perceived level of support from the school community. DA mothers reported the use of more positive parenting practices, less harsh discipline and more parent involvement practices; there were no group differences in school-based parent involvement practices as rated by teachers. MA children were rated by mothers and by teachers as lower in adaptive behavior than DA children, and DA children were rated by teachers as showing more problem behaviors.

Table 2 Descriptive statistics and correlations among study variables by ethnic group
Full size table

Table 3 shows the correlations between demographic variables and social support by ethnic group. There was only one significant association, seen only among MA mothers; maternal education was negatively associated with perceptions of support from the school community.

Table 3 Correlations among social support and demographic characteristics by ethnic group
Full size table

Model Testing

As per the analytic plan, we first tested a nonrestricted model using multigroup SEM analyses. The overall χ² statistics showed a good fit of the nonrestricted model, χ² (16) = 26.19, p = .05, RMSEA = 0.05 and CFI = 0.98. We then tested the model that constrained all parameter estimates (or path values) to be equal across groups, but that allowed free estimate for means, variances, and covariances. The model yielded a reasonable fit, χ² (45) = 88.15, p < .001, RMSEA = 0.06 and CFI = 0.93, but still led to a significant decrement in model fit (χ²∆ (29) = 61.96, p < .001). This suggested that the paths may be different for MAs and DAs, so we conducted SEM analyses separately for the MA and DA groups.

Using SEM analyses separately for each group, we found a fit of the hypothetical model for both MAs and DAs [MA: χ² (8) = 6.52, p = .56, RMSEA = 0.00 and CFI = 1.00 and DA: χ² (8) = 19.81, p = .01, RMSEA = 0.07 and CFI = 0.97]. Figure 2a (for MA) and 2b (for DA) present the standardized path coefficients for the significant paths and the R² values for the mediating and dependent variables (i.e., parenting and child functioning). For MAs, one significant mediational path was found. Specifically, familial support was associated with more positive parenting, which was then associated with higher levels of adaptive behavior in children (parent report). The standardized indirect effect (SIE) was 0.10, p < .001, indicating that one full standard deviation increase in familial support via its effect on positive parenting would result in 0.10 standard deviation increase in parent-rated adaptive behaviors. No path linking social support and child functioning via harsh parenting was found, though we did find harsh parenting was associated with less adaptive behavior and more problem behavior as rated by mothers and teachers. No path linking social support and child functioning via school-based parent involvement practices was found either, and school-based parent involvement practices was not related to child functioning.

Fig. 2
figure 2

a SEM Model for MA. b SEM Model for DA

Full size image

For DAs, we found that familial support was associated with more positive parenting, which was associated with higher levels of child adaptive behaviors as rated by mothers (SIE = 0.08, p = .02). We also found that receiving less support from the school network was associated with more use of harsh parenting, which was related to more problem behaviors and less adaptive behavior among children as reported by mothers, but this indirect effect did not reach significance. No link between social support and school-based parent involvement practices was found, but parent involvement was related to higher levels of adaptive behavior as reported by teachers and lower levels of adaptive behavior as reported by mothers.

Discussion

The present study aimed to describe social support as perceived by Latina mothers and to test its association with child functioning, through parenting practices. While there is ample evidence that social support has positive effects on parenting and child development among non-Latino White samples [40], less is known about the effects of social support among immigrant Latinos, an important gap in the literature considering the generally high levels of social disadvantage experienced by this population of parents. Attending to cultural context, we focused on two specific ethnic groups, Mexican- and Dominican-origin, and distinguished between social support from family versus extrafamilial networks.

Perceived Social Support from Family and School Networks

The first aim of our study was to describe the levels of perceived support from two distinct sources: family and school community. As expected, support was high from family as experienced by both DA and MA mothers, supporting the notion of centrality of family for Latino populations [13]. However, DA mothers reported more support from family than MA mothers, perhaps because DA mothers in NYC had access to larger, multigenerational networks of family members who resided nearby and could provide both emotional and instrumental support [82]. In contrast, MA mothers in NYC are more likely to represent the first generation of Mexican immigrants to the city [23, 82]. As such, they may have received high levels of emotional support from family in their country of origin via phone or social media [2] but given the physical distance between family members, emotional support may have been less salient and instrumental support may have been altogether limited. Research shows that during the beginning of a migratory flow (e.g., Mexicans to NYC), immigrants tend to live with other migrants who are not part of their family network. In contrast, once a destination has become an established receiving community (i.e., an ethnic enclave), immigrants tend to live with extended family members [23].

Surprisingly, there were no group differences in perceived social support from the school community. Although DA mothers were more highly acculturated (e.g., more proficient in English) and, as residents in an ethnic enclave, were more likely to know other parents of children zoned to the same public school as their child, both MA and DA mothers reported similar and modest levels of support from school. Perceived support was not correlated with demographic variables, including maternal work status or years in the US, but future studies should aim to identify factors associated with support outside of the family, given that support from extrafamilial networks is protective [63, 75].

Perceived Social Support and Child Outcomes

The second aim of the present study was to test a path model of social support in relation to child outcomes, indirectly through parenting. Specifically, we examined how social support from family and school networks was linked with parent and teacher ratings of children’s problem and adaptive behaviors, indirectly through positive parenting, harsh parenting, and parent involvement in education. Our model was partially supported, and only some of the hypothesized associations were observed. Our main finding, which emerged for both MA and DA families, showed a path linking support from family networks with child adaptive behavior in the home, through positive parenting practices. Although untested in the present study, it may be that support from family members increased parental self-efficacy [40] and/or helped mothers to manage their stress, thereby promoting the use of positive parenting practices [33, 50, 58].

Though the paths linking support from the school community to child outcomes were not significant, we did find that more support from school networks was associated with lower self-reported use of harsh practices among DA mothers. We expected that mothers who received support from the school community would have more exposure to mainstream childrearing values and norms, including the school-sanctioned endorsement of non-physical discipline strategies such as time out. While this appears to have been the case with DA mothers, it is not clear why the association was not observed among MA mothers. Qualitative studies with Spanish-speaking parents have shown that language barriers limit the depth and meaningfulness of social connections within school communities [28]. It is possible that in the present study, the lower levels of English proficiency among MA mothers shaped the types of support they received from other parents at the school (e.g., shared recreational time versus advice or access to resources), limiting its effects on parenting. More work is needed to identify which components of support help immigrant Latina mothers rely less on harsh practices, especially considering the robust associations found between harsh parenting and child functioning.

Based on our conceptual model, we also expected, but did not find, that higher levels of social support within the school setting (by other parents, not by school staff specifically) would be associated with more parent involvement, conceptualized as teacher ratings of mother involvement in school-based activities. It may be that parent–parent relationships are not sufficient to offset cultural gaps between Latina mothers and US American schools, and instead, parent-teacher relationships are needed to promote school-based parent involvement. Not surprisingly, given the cultural gap between Spanish-speaking immigrant mothers and non-Spanish-speaking US American teachers, on average mothers were rated as only “somewhat” involved. Previous research with schools serving MA populations has highlighted significant barriers to parent involvement, showing that even with communication skills training, Spanish-speaking parents struggle with school-based involvement. Gonzales and Dumka [27] recommend that beyond language issues, scholars should attend to how parent involvement may be promoted by creating a receptive school climate that parents understand how to navigate.

Parenting and Child Outcomes

In considering parenting and child outcomes, we made no hypothesis about the association between harsh practices and child functioning. We found that harsh parenting was associated with less adaptive behavior and more problem behavior as rated by both mothers (for DA and MA children) and teachers (for MA children only). It is important to note that all mothers, whether MA or DA, reported high levels of positive parenting and low levels of harsh parenting (with DA mothers reporting modestly but significantly higher levels of positive parenting). Much scholarly discussion has centered around the use of harsh practices among Latino parents, as past studies have indicated that physical punishment is culturally-normative [45] and possibly unrelated to Latino child outcomes [6, 67]. The present study findings suggest that while harsh practices such as spanking may be acceptable, they nonetheless appear to be used infrequently with 4–5 year old children. Our findings further suggest that, despite their infrequent use, harsh practices are associated with poor outcomes among young Latino children. More research using longitudinal study designs is needed to test the direction of these associations.

In addition, our hypothesis that school-based parent involvement would be associated with positive child functioning was only partially supported. Parent involvement was not related to MA child outcomes at all and was differentially related to DA child outcomes depending on context. Specifically, school-based parent involvement was associated with more adaptive behavior at school but with less adaptive behavior at home. A positive association between school-based parent involvement and child functioning has been found in past studies, as mothers who are more engaged with their child’s teacher may be best able to support their child’s success in the classroom [36]. The negative association between school-based parent involvement and child functioning at home was unexpected, but it may be that mothers who are concerned about their child (e.g., in terms of low adaptability, difficulty making friends) are more likely to engage and communicate with school staff in an attempt to address these concerns. Again, longitudinal data are needed to examine how child characteristics influence parenting and vice versa.

Limitations and Future Directions

The reliance on cross-sectional data in the present study was one major limitation. Longitudinal data is needed to examine causality and also to examine change over time in social support and parenting. For example, it may be that the effects of social support from the school community are not immediate but develop over time. Studies are needed to examine how relationships between immigrant Latina mothers and a community of parents from a mainstream school develop, are maintained, and influence the parenting values and practices of mothers. The ethnic makeup of the school population may be important to consider as well. For example, school communities that are made up of other Spanish-speaking parents may facilitate the integration of new Latino immigrant parents into school networks, providing invaluable social support and influencing the ways in which new parents engage in school-based parent involvement practices [44]. In the present study, families were drawn from schools that ranged from approximately 20 % Latino to nearly 100 % Latino, and it is likely that the ethnicity of parents in the school network is important in the study of social support [44]. It will also be important for future studies to examine a more comprehensive model of social support that includes correlates of social support from different sources and the mechanisms by which support influences parenting. Finally, the present study findings may be specific to Mexican- and Dominican-origin families during the first years of schooling (i.e., pre-k and kindergarten) and should not be generalized to other ethnic groups or child ages without further study.

Despite these notable limitations, the present study used a large sample, attended to ethnic group differences, and included both mother and teacher report to examine social support among mothers as a protective factor for the development of young children. Support, specifically from family networks, does appear to be related to positive parenting and positive child outcomes. Importantly, this association was seen even for MA mothers whose support may have come primarily from family members who remained in Mexico (i.e., via social media). Though causality cannot be inferred, these findings underscore the importance of maintaining family ties and suggests that interventions would be strengthened by recognizing the role of extended family and perhaps even by encouraging the participation of extended family [31, 57]. Though attention to the potential costs that are associated with high levels of family involvement and obligation is warranted [13], familial support appears to be an important factor for Latino families and one that can potentially mitigate the risks associated with socioeconomic disadvantage (e.g., living in poverty; minority status).

Summary

The role of social support in diminishing the stress related to parenting may be especially important for immigrant families who must deal with the stress of adapting to a new culture without the adequate amount of economic resources [18, 40]. Despite evidence showing that social support is an important protective factor, little is known about the role of social support in parenting and child development in Latino families [3]. Considering the importance that Latinos give to the family unit [48], the first aim of our study was to describe the level of perceived support that Latina mothers receive from the school community and from family networks. In addition, we focused on examining social support as a protective factor for the early childhood functioning of children, mediated by parenting practices. In examining parenting practices, we consider the respective roles of positive relative to harsh parenting on Latino child functioning, an issue that is debated in the literature [34, 39]. Six hundred and ten Mexican and Dominican American families participated in our study. Results showed that familial support was higher for both ethnic groups than support received from school community. Dominican mothers reported higher levels of family support than Mexican mothers, but the groups were not different in the level of perceived support from the school community. We also found some evidence of the protective effect of social support on children´s functioning, indirectly through parenting. For both ethnic groups, familial support was associated with child functioning, indirectly through positive parenting practices. Although support from the school community was associated with less frequent use of harsh practices, it did not show significant association with child behavior. Notably, mothers reported high levels of positive parenting and low levels of harsh parenting practices. In addition, harsh parenting was related to less adaptive behavior and more problem behavior for both ethnic groups. Finally, mothers showed a modest level of involvement in school-based activities, which was associated with outcomes among Dominican but not Mexican children. Overall, our study provides some evidence that familial support is an important factor for Latino families that may promote early childhood development.