Friend Support and the Parenting of Latina Adolescent Mothers: The Moderating Role of Maternal Age

Abstract

Objectives

This study examined the role of maternal age in the relation between social support from friends and parenting adjustment in a sample of young Latina mothers and their 18-month-old children (N=168).

Methods

Hierarchical multiple regression analyses tested friend social support types (emotional, socializing, child care) as differential predictors of maternal behavior (sensitivity, cognitive growth-fostering, detachment) displayed during mother-child play interactions. To consider maternal development, the moderating role of maternal age on these associations was tested.

Results

The relations between friend emotional and child care support and parenting were moderated by maternal age. Emotional support was related to the use of more growth-fostering parenting behaviors for older (≥ 19.5 yrs.), but not for younger Latina mothers. Child care support from friends was related to the display of more detachment and less cognitive growth-fostering behaviors among the younger (≤ 18.7 yrs.) mothers only. Immigrant mothers reported significantly less overall friend support and emotional support than mothers born in the mainland U.S.

Conclusions

The findings emphasize the importance of assessing the types of friend support as separate measures in an ecological context that takes into account mothers’ generational and developmental level.

The literature extensively documents the risk factors related to adolescent parenting. Younger mothers are more likely to be single, live in impoverished neighborhoods, have lower educational attainment, and report higher depressive symptoms (Hoffman & Maynard, 2008; Mollborn & Morningstar, 2009). Adolescent mothers of Latina origin face additional challenges such as immigration and discrimination (Zeiders, Umaña-Taylor, & Derlan, 2013) and have the highest teen birthrate in the U.S. (Hamilton, Martin, Osterman, Curtin, & Matthews, 2015). Not surprisingly, adolescent mothers exhibit less sensitive and more intrusive parenting behavior than do adult mothers (Berlin, Brady-Smith, & Brooks-Gunn, 2002). Moreover, findings link these parenting difficulties to compromised child development starting in the second year of life (Grau et al., 2015; Jahromi, Guimond, Umaña-Taylor, Updegraff, & Toomey, 2014). Despite these risks, the unique set of factors that facilitates Latina adolescent mothers’ adjustment remains largely unknown.

Ecological parenting models (Belsky, 1984; Contreras, Narang, Ikhlas, & Teichman, 2002) highlight the impact of social support on parenting adjustment, especially for mothers who are parenting in disadvantaged conditions. Consistently, research on protective factors for adolescent mothers centered on the role of social support (i.e., exchange of emotional, social, and tangible resources; Grau, Wilson, Weller, Castellanos, & Duran, 2012). This research focused on support provided by either the adolescent’s own mother (grandmothers) or by her social network (Diniz, DeSousa, Koller, & Volling, 2016; Grau et al., 2012). Studies focusing on overall network support found positive associations with parenting (Contreras, Mangelsdorf, Rhodes, Diener, & Brunson, 1999; Voight, Hans, & Bernstein, 1996). However, despite grandmothers being a primary source of support for adolescent mothers (Crase, Hockaday, and McCarville, 2007; Devereux, Weigel, Ballard-Reisch, Leigh, & Cahoon, 2009), the effects of their support on maternal parenting adjustment are mixed. Whereas some studies found positive associations with some aspects of parenting adjustment (e.g., parenting knowledge, self-efficacy; Jahromi et al., 2014; Umaña-Taylor, Guimond, Updegraff, & Jahromi, 2013), several other studies, especially those assessing observed parenting behaviors, found that greater involvement and support from grandmothers related to less optimal parenting (Contreras, Lόpez, Rivera-Mosquera, Raymond-Smith, & Rothstein, 1999; Driscoll & Easterbrooks, 2007). These negative effects were more likely to occur in certain contexts, such as when the adolescent resided with her mother or when they experienced relationship strain (Contreras, Lόpez, et al., 1999; Sellers, Black, Boris, Oberlander, & Myers, 2011; Spieker & Bensley, 1994). The type of support also influenced the relation to parenting, with child care showing more negative effects than other types of support (Contreras, 2004). Given these negative findings, it is important to identify additional sources of support that can be protective for young mothers’ parenting adjustment.

Guided by Contreras, Narang, Ikhlas, and Teichman’s (2002) model of Latina adolescent parenting which integrates the developmental, cultural, and socio-demographic factors contributing to the ecological context of Latina adolescent mothers, we focused on the adolescents’ friends as an additional source of support. In this model, maternal developmental factors such as age, and cultural factors such as generational status are proposed to relate to maternal support expectations and moderate mothers’ ability to benefit from this support (Contreras et al., 2002). Maternal age likely influences the role of social support on parenting in at least two important ways. First, the developmental age of the mothers affects the relative importance of the different sources of support that are available to or sought out by mothers. The period of adolescence to young adulthood is a period of identity exploration, when individuals venture out of the family sphere to socialize with friends (Steinberg & Morris, 2001). The increasingly important valuation of friendships as adolescents mature, as well as the meaningful social and emotional developmental opportunities presented within those friendships, have been well documented in the literature (Stanton-Salazar & Spina, 2005; Way & Chen, 2000). Second, age affects how the mother utilizes support in meeting her developmental goals. Young mothers are faced with the divergent developmental goals of adolescence and parenthood (Contreras et al., 2002). The adolescent developmental goals of autonomy-seeking and identity exploration are partially inconsistent with parenting developmental goals (Contreras et al., 2002; Nadeem & Romo, 2008). For example, supports that facilitate identity exploration (e.g., socializing, child care support) may be detrimental to the young mothers’ ability to develop parenting skills, whereas supports that facilitate parenting may limit identity exploration and foster social isolation (Contreras et al., 2002; Easterbrooks, Chaudhuri, & Gestsdottir, 2005). Thus, although overall support can have positive effects on the well-being of adolescents (Steinberg & Morris, 2001; Way & Chen, 2000), in the case of parenting adolescents, the extent that friend support promotes maternal behavior may vary by the specific types of support and the developmental goals of the mother.

From this developmental perspective, friends constitute a developmentally salient source of support for adolescents, and research indicates that adolescent mothers in fact do perceive their friends as valuable sources of support (Richardson, Barbour, & Bubenzer, 1991; Voight et al., 1996). Emotionally strong friendships provide a context for adolescents to practice reciprocal behavior that may extend to their parenting practices (Thompson et al., 2006). Moreover, friends may be uniquely positioned to provide positive support without the conflict that can be present in familial relationships (Nath, Borkowski, Whitman, & Schellenbach, 1991). Consistent with these developmental considerations, there has been an increased focus on friend support in the adolescent parenting literature (Bravo, Derlan, Umaña-Taylor, Updegraff, & Jahromi, 2018; Huang, Roberts, Costeines, and Kaufman, 2019). Overall, this literature suggests positive effects on psychological adjustment and perceived parenting stress among Latina and African American adolescent mothers. Although a few studies found that higher levels of friend support related to greater psychological distress (Thompson & Peebles-Wilkins, 1992), other studies found a relation to less emotional and life stress (Leadbeater & Linares, 1992), higher life satisfaction (Crnic & Greenberg, 1987; Unger & Wandersman, 1985), and lower parenting-role stress (Huang et al., 2019; Richardson, Barbour, & Bubenzer (1995).

However, only three studies examined the effects of friend support on observed parenting behavior in adolescent mothers and they yielded mixed results. In two studies, composites of friend support were unrelated to maternal behaviors (e.g., sensitive, affectively positive, growth-promoting) in samples of European American and African American mothers (Crnic, Greenberg, Ragozin, Robinson, & Basham’s, 1983; Voight et al., 1996). However, in one of these studies (Voight et al., 1996), the number of friends in the network related positively to desirable parenting behavior. Finally, research with a sample of British mothers found that higher levels of friend emotional support related to more positive and less negative parenting behaviors (responsiveness/positive affect and negative control/affect composites; Ensor & Hughes, 2010). Thus, further research is needed to clarify these findings and determine the extent to which friend support can be protective for young mothers’ parenting behavior.

Following developmental considerations outlined by the model of Latina adolescent parenting (Contreras et al., 2002), it is also important to examine whether maternal age influences how friend support relates to young mothers’ parenting behaviors. Most adolescent parenting studies treat maternal age solely as a control variable (Bravo et al., 2018; Umaña-Taylor et al., 2013). This approach assumes the predictors and consequences of parenting operate similarly at all ages. However, there are developmental differences among middle to late adolescence, and the early adulthood years, which can influence how protective factors operate (Bornstein, Putnick, Suwalsky, & Gini, 2006; Thompson et al., 2006). In fact, research suggests that the protective influence of social support may vary according to the mothers’ developmental age (Bornstein et al., 2006; Ensor and Hughes, 2010). For example, Davis’ (2002) research with African American adolescent mothers found that for mothers with lower family support, higher levels of support from friends related to less depressive symptomatology, but only for the older mothers in the sample. For the younger mothers, the same support context related to higher levels of depression. Similarly, Ensor and Hughes (2010) found that friend support related to more optimal parenting behaviors for adolescent mothers but not for adult mothers. Although scholars have called for the consideration of how maternal age may influence both predictors and consequences of parenting across all ages (Bornstein & Putnick, 2007; Contreras, et al., 2002), few studies have actually empirically tested age as a moderator in samples of adolescent mothers.

In addition to the limited consideration of maternal age in the literature, prior studies relied on global measures of support that do not adequately differentiate among specific support types (e.g., emotional, socializing, child care). Emotional and socializing are the most common types of support reported by adolescent mothers (Crase et al., 2007; Richardson et al., 1991; 1995; Voight et al., 1996). Child care support, although less commonly provided by friends, is a vital tangible support need for adolescents who are also mothers (Contreras, Lόpez, et al., 1999; Driscoll & Easterbrooks, 2007), yet the influence of friend child care support remains unexamined. Young mothers may use different types of support based on the needs created by the stressors they face and the developmental goals they pursue (Contreras et al., 2002; Nath et al., 1991). The closeness and intimacy inherent in emotional support from friends may help mothers cope with parenting demands and, thus, translate into more optimal parenting. Socializing and child care received from friends are two types of support which may meet identity development or tangible needs but may not provide mothers, especially those with less developmental maturity, the opportunities to learn and develop parenting skills.

Finally, in addition to developmental considerations, it is also important to consider cultural factors that influence social support processes in Latina families (Contreras et al., 2002). Although family is the predominant source of support for Latino adolescents, research with non-parenting young Latinos indicates that they also value friendships, especially as a source of emotional support (Stanton-Salazar & Spina, 2005). However, generational status, a characteristic emphasized in the diverse social support networks of Latinos, may shape how many friends the adolescents have and how often friends are utilized for support (Contreras et al., 2002; López & Cooper, 2011). First generation adolescents born outside the U.S. report smaller networks and are more likely to prioritize family support than do adolescents of later generations who are born in the U.S. (de Guzman, Jung, & Anh Do, 2012; Stanton-Salazar & Spina, 2005). The utilization of more family than friend support aligns with the Latino cultural value of familismo (i.e., interdependence, loyalty, and support in family relationships; Carlo, Koller, Rafaelli, & de Guzman, 2007), which is endorsed more strongly by first generation immigrant adolescents than those of later generations (Sabogal, Marín, Otero-Sabogal, Marín, & Perez-Stable, 1987). Additionally, individuals who were born and raised in their communities have also had more time and opportunities to form extended friend and community networks (Leonardo, 2016). Given the limited research on the friendships of young Latina mothers, and the importance of generational processes in understanding these social networks, studies are needed that consider differences in friend support by generational status with Latina adolescent mothers.

The current study examined the protective role of different types of friend support for the parenting of Latina adolescent mothers, as well as how maternal age may moderate these effects. For our primary aim, we tested whether specific types of friend support differentially related to parenting behaviors displayed by mothers. Friend emotional and socializing support were included due to literature suggesting their relevance to the adolescent mother. Friend child care assistance was selected to test a type of support that may be more central to the parenting role. Since the effects of these social support types may change during this important developmental period for adolescent mothers, we examined whether maternal age moderated the associations between support and parenting. We tested these associations controlling for the support received from all other providers in the mothers’ network to better examine the relative contribution of friend support. Following the parenting literature, we coded observed parenting behaviors with demonstrated relations to child development and created composites reflecting key parenting constructs that have been commonly studied in the literature (i.e., sensitive/affectively positive behavior, detachment: Berlin et al., 2002; cognitive stimulation: Voight et al., 1996). We examined amount of support utilized in the last month, rather than perceived support, because research indicates that enacted support is most relevant for individuals parenting in adverse, stressful conditions, and it informs intervention efforts more directly (Collins, Dunkel-Schetter, Lobel, & Scrimshaw, 1993; Rini, Dunkel-Schetter, Hobel, Glynn, & Sandman, 2006). We expected that friend emotional support would be associated with more optimal parenting regardless of maternal age. We also predicted that, for younger adolescents, friend socializing and child care assistance would be associated with less competent parenting, whereas these types of support would relate to more competent behaviors for mothers in early adulthood. In light of the scarcity of studies that explored the support networks of young Latina mothers, a complementary goal was to provide descriptive information on the mothers’ friend networks within their cultural context. Thus, we tested differences in support between US-born and immigrant mothers. We expected that US-born mothers would report having more friends and receiving more support from them than immigrant mothers.

Method

Participants

The participants of this study were 168 Latina adolescent mothers and their toddlers who participated when the children were 18 months old. The mean age of the mothers was 17.94 years (SD = 1.32; range: 14.25 – 19.92) at the time of the child’s birth; the mean age at the time of data collection was 19.5 years (SD = 1.33; range: 15.76 – 21.55). Mothers were predominantly of Puerto Rican origin (81.2%; 7.1% Mexican; 10.8% Central and South American origins). The majority of mothers (88.1%) reported receiving some form of government assistance and 45.24% were born outside the mainland. At the time of data collection, 32.1% of mothers had completed high school or some type of additional education (e.g., vocational training), whereas 9.5% had started but did not complete 12^th grade, 15% completed 11^th grade, 15.5% completed 10^th grade, 18.5% completed 9^th grade, and 9.5% had not completed 9^th grade. However, 26.2% of mothers were still attending school full-time or part-time, and among mothers old enough to have completed high school (18.5 years and older), the high school graduation rate was slightly higher than the one for the whole sample (37.9% vs. 32.1%). In addition, 41.7% of mothers reported being employed full-time or part time. The mean age of toddlers (54.8% male) was 18.21 months (SD = .96; range: 15.94 – 20.79) at the time of data collection. The majority of children were the only or first child (84.5%). Over 92% of them were born in the mainland U.S.

Procedure

The study followed procedures approved by two Institutional Review Boards. Participants were recruited through face-to-face contact in the county hospital (51.8%) and satellite community clinic (26.5%) serving low-income Latino neighborhoods in a large Midwestern city. Other participants were recruited through advertisement (8.2%) and referrals from friends or relatives (7.1%) or from professionals (6.4%). ANOVA and chi-square analyses comparing across three recruitment groups (hospital [n = 88], satellite clinic [n = 45], and all other methods [n = 37]) indicated that there were no significant differences (all p’s > .05) in main study variables (i.e., friend support, maternal behavior, maternal age) or demographic characteristics (i.e., maternal education, school enrollment, employment, generational status, and child age and gender). The participation rate was 70.5%. Two female researchers, at least one of whom was bilingual, conducted home visits in the participants’ preferred language (70.6% English; 29.4% Spanish). The researchers obtained informed consent from the participant (and a parent if she was under 18), videotaped mother-child interactions, and conducted computer-assisted interviews. On average, the home visits lasted two and a half hours.

Measures

Social support.

The Social Support Network Questionnaire (SSNQ; Gee & Rhodes, 2007), developed to assess support resources among minority adolescent mothers, was used to assess amount of support received from the mothers’ network and friends across three types of support. Mothers nominated persons they perceived as available to provide emotional (talking about something personal or private), socializing (getting together to have fun or relax), and child care (helping take care of child) support. For each person nominated, the mother was asked to categorize the type of relationship (e.g., friend, mother, partner) and report the amount of support received from that person in the last month, on a 4-point scale (0 = “Never,” 1 = “Once or twice this month,” 2 = “About once a week,” and 3 = “More than once a week”). The SSNQ has been used to assess overall network support and support from different network members in samples of African American and Latina adolescent mothers, and it demonstrated conceptually meaningful relations to the young mothers’ psychological and parenting adjustment. For example, greater overall network support related to higher maternal expressivity, and child care support from the adolescents’ mothers related to lower maternal sensitivity (Contreras, 2004; Contreras, Mangelsdorf, et al., 1999; Gee & Rhodes, 2007).

Four friend support variables were derived: Total friend support (total amount of support received from all friends), and the amount of emotional, socializing, and child care support received from friends during the last month. Given that socializing interactions with friends likely include a component of emotional support, to better index the amount of purely socializing interactions, we derived the final socializing support variable by subtracting the frequency of emotional support from that of socializing support received from friends. For mothers who reported more emotional than socializing support, and therefore had negative numbers on this new variable, we substituted the negative scores with zeros. The resulting variable indicated that 66.7% of the mothers did not receive purely socializing support (i.e., they report also receiving emotional support during their socializing interactions). The remaining 33.3% of the mothers reported some frequency of interactions where they engaged in socializing even after subtracting emotional support. This variable was kurtoted and was dichotomized (no versus some socializing support) for the analyses. The majority of mothers (82%) did not report help with child care from friends in the past month, and the amount of child care assistance variable was skewed and kurtoted. Thus, this variable was dichotomized to reflect whether mothers received any child care assistance.

Four parallel support variables reflected support received from all network members, excluding that received from friends. The total network support variable was calculated by summing total amount of support received by the mother. The network emotional, socializing, and child care support variables were calculated by summing the amount of each type of support. We created a final variable of purely socializing network support following the same procedures used for friend socializing support. Approximately half of the mothers (50.6%) did not receive purely socializing support from their network. This variable was kurtoted and was dichotomized (no versus some support) for the analyses.

In sum, four support variables were computed for friends and for the rest of the support network, respectively: total support, emotional support, socializing support, and childcare assistance. Prior studies, including those with Latino samples, have utilized similar measures for frequency of support (Bámaca-Colbert, Tilghman-Osborne, Calderón-López, & Moore, 2017; Devereux et al., 2009; for a review, see Uchino, 2009).

Maternal behavior.

We assessed maternal behaviors during a five-minute, semi-structured, videotaped mother-toddler interaction. Mothers were asked to play with their child as they normally would when playing without toys. We coded the episode for maternal sensitivity, positive affect, detachment, cognitive stimulation, and repertoire using five global 9-point scales (from 1 = “Not at all characteristic of the behavior” to 9 = “Very characteristic of the behavior”). The sensitivity, positive affect, and repertoire scales were adapted from scales used by Isabella (1993) to better asses behavior among young Latina mothers (Contreras et al 1999). Detachment and cognitive stimulation were assessed with scales from the NICHD Study of Early Child Care, which were originally used to code a 15-minute mother-child interaction (NICHD Early Child Care Research Network, 2001). These two scales were transformed from 4-point to 9-point scales by adding intermediate anchor points in order to better capture nuances in behavior within the shorter, 5-minute task used in this study and to match the number of points in the other three scales. Sensitivity assessed the timing and appropriateness of the mothers’ responses to their children’s cues. Positive affect was the frequency and intensity of maternal warm (e.g. smiling, kissing, positive vocal tone). Detachment reflected the extent of emotional disengagement, under-involvement, and lack of monitoring of the child. Cognitive stimulation was the frequency and quality of teaching activities with the child. Repertoire assessed the number and types of different approaches or activities used to engage the child. Coders assigned one score for each scale based on the entire task, and overlapped in approximately 25% of the observations (n = 46). Interrater agreement, assessed using intra class correlations, was adequate for the five scales: sensitivity ICC = .70; positive affect ICC = .69; detachment ICC = .67; cognitive stimulation ICC = .80; repertoire ICC = .81. Support for the construct validity of the two transformed scales, detachment and cognitive stimulation, emerged via moderate and strong correlations with the other established scales. As expected, detachment was negatively correlated with sensitivity (r = −.60), positive affect (r = −.55), and repertoire (r = −.48). Cognitive stimulation was positively correlated with sensitivity (r = .49), positive affect (r = .38) and repertoire (r = .82).

The five scales were moderately to strongly intercorrelated (range of r absolute values = .35 to .82). To reflect sensitive and affectively positive behavior, we averaged the sensitivity and positive affect scales, which were highly intercorrelated (r = .66). We labeled this composite maternal sensitivity. To create a cognitive growth-fostering behavior composite, we averaged cognitive stimulation and repertoire (r = .82). Higher scores in this composite reflected a higher frequency and variety of growth-promoting behaviors. Detachment was less strongly correlated with the other scales (r range = −.35 to −.60), supporting its use as a separate variable, with higher values indicating greater disengagement. Thus, we retained three behavior variables for the analyses: maternal sensitivity, cognitive growth-fostering behavior, and detachment.

Data Analyses

Data were analyzed using SPSS 22.0.0 (IBM Corp., Armonk, NY, USA). To address the descriptive goal of the study, we first provide information about mother-reported friend characteristics and support, and their relations to generational status (i.e., born outside vs. born in the mainland US). We then report analyses to address the main goal of the study, starting with preliminary bivariate correlations testing whether control variables should be included in subsequent analyses. Finally, separate hierarchical multiple regression analyses assessed the associations of friend support types and their interactions with maternal age at the time of the study on maternal sensitivity, cognitive growth-fostering behavior, and detachment, respectively. All regressions controlled for the corresponding type of support received from all other members in the mothers’ networks. We computed interaction terms using centered variables and examined significant interactions using R. Chris Fraley’s computational application (supplement to Roisman et al., 2012), including region of significance analyses to identify the specific age ranges for which friend support predicted maternal behavior.

Results

Describing Friend Characteristics and Support

One hundred and four mothers nominated at least one friend (61.9%). Of these mothers, 58 (55.8%) nominated only one friend, 25 (24%) nominated two friends, 16 (15.4%) nominated three friends, and five mothers (4.8%) nominated four or five friends. In all, mothers nominated 178 friends. The mean age of mothers’ friends was 22.3 years (SD = 5.11) and it was significantly positively correlated with the mothers’ own age (r = .21, p = .03). Friends were mostly females who resided relatively close to the mothers and talked with them daily. Emotional support was the most frequent type of support received from friends at 39%; 33% of mothers reported receiving friend socializing support, while only 18% of mothers reported receiving child care assistance.

In regard to differences by generational status, the difference for number of friends did not reach significance (t(166) = −1.88, p = .06; M = .88, SD = 1.13 for immigrant mothers; M = 1.21, SD = 1.11for U.S.-born mothers). Nonetheless, immigrant mothers (M = 4.76, SD = 7.86) reported significantly less overall friend support, t(166) = −2.33, p = .02 than U.S.-born mothers (M = 7.89, SD = 9.28). Cohen’s effect size (1988; d = .36) suggested that practical significance was small to moderate. Parallel analyses for each type of support indicated that immigrant mothers (M = .74, SD = 1.54) reported significantly less emotional support, t(163.21) = −2.99, p =.003, than women born in the U.S. (M = 1.59, SD = 2.14). Cohen’s effect size (d =.46) indicated that practical significance was medium. There was no difference by generational status in socializing support, χ²(1) = 1.92, p = .66 or child care assistance, χ²(1) = 2.09, p = .15.

Selection of Control Variables

Bivariate correlations examined the relations between maternal (e.g., age, education, employment, school enrollment, economic strain, welfare receipt, generational status) and child characteristics (i.e., age, gender), and maternal parenting behaviors. Results indicated that none of these factors were significantly related to parenting behavior. Results for the correlations between these factors and the support variables indicated only two significant correlations (beyond those for generational status reported above). Mothers who were employed were less likely to report socializing support from their networks than their non-working peers, χ²(1) = 5.63, p = .02. Finally, those who reported higher economic strain received less network child care support (r = −.23, p =.03).

Associations between Friend Support and Maternal Behavior

Bivariate correlations examined the relations among friend support, network support, maternal age, and parenting behaviors (Table 1). Results indicated that overall friend and overall network support were unrelated to parenting. Friend emotional, socializing, and child care support were not associated with the parenting behaviors. Network socializing support was significantly related to the display of less sensitive parenting. Age was unrelated to maternal behavior.

Table 1.

Bivariate Correlations, Means, and Standard Deviations for Main Study Variables

	Friend Support				Network Support				Parenting			M(SD)	Observed Range
Variable	1	2	3	4	5	6	7	8	9	10	11
1. Total Friend Support	-											6.5(8.8)	0–41
2. Friend Emotional Support	.87***	-										1.2(1.9)	0–8
3. Friend Socializing Supporta	.19*	−.11	-									.33(.47)	0–1
4. Friend Child Care Supporta	.59***	.41***	.20*	-								.18(38)	0–1
5. Total Network Support	−.14	−.20*	−.09	.04	-							31.9(16.1)	2–91
6. Network Emotional Support	−.09	−.10	−.10	.03	.69***	-						4.6(3.4)	0–16
7. Network Socializing Supporta	−.20**	−.28***	.03	−.09	.16*	−.29***	-					.49(.50)	0–1
8. Network Child Care Support	−.00	−.04	−.05	.06	.68***	.35***	.16*	-				5.5(3.4)	0–18
9. Sensitivityb	.06	.11	−.01	.04	−.04	.03	−.18*	−.04	-			5.1(1.0)	3–8
10. Cognitive Growth-Fosteringb	.02	.09	−.15	−.07	.01	.06	−.09	−.04	.61***	-		3.4(.99)	1.5–7.5
11. Detachmentb	.12	.08	.05	.01	−.02	−.02	.02	−.04	−.63***	−.44***	-	1.32(.58)	1–3
12. Maternal Age	−.10	−.12	.03	.03	−.04	−.02	.03	.07	−.02	.06	−.05	19.5(1.34)	15.8–21.6

Note. (N = 168).

^aDichotomized variable (none = 0; at least some = 1).

^b9-point scale.

^*p <.05;

^**p <.01;

^***p <.001

Multivariate Associations between Friend Support and Maternal Behaviors

To test direct and moderated associations between types of friend support and the parenting behavior variables, we computed hierarchal linear regressions. For all regressions, network support, friend support and maternal age were entered in the first step, and moderator effects were tested by entering the maternal-age-by-friend-support interaction term in the second and last step. Parallel regression models tested whether overall friend support predicted maternal behaviors (controlling for overall network support). Neither the main effect nor the interaction term between total friend support and age approached significance (results available from the authors). Examination of Tolerance and VIF values indicated that multicollinearity was not a problem in any of the models.

The first hierarchical linear regression assessed the association of friend emotional support and maternal age on parenting after accounting for the effect of network emotional support (Table 2). There were no main effects or interaction effects for emotional support on maternal sensitivity or detachment. However, the interaction of friend emotional support and age significantly predicted cognitive growth-fostering behavior, above and beyond total network emotional support. The interaction accounted for 3% of the variance in growth-fostering. The predicted values of growth-fostering were plotted at one standard deviation below and one standard deviation above the mean for maternal age (Figure 1). Simple slope analysis revealed that the slope was significant for older (t = 2.84, p =.01), but not for younger mothers (t = −.113, p = .91). Region of significance analysis indicated that friend emotional support significantly predicted the display of more cognitive growth-fostering for mothers who were 19.5 years and older (range = 19.5 to 21.6; all t’s > 2.00; all p’s < .05; depicted by the shaded area surrounding the + 1 SD slope).

Table 2.

Hierarchical Regression Predicting Parenting from Friend Emotional Support

	Sensitivity						Cognitive Growth Fostering						Detachment
	Model 1			Model 2			Model 1			Model 2			Model 1			Model 2
Predictors	B	SE B	β	B	SE B	β	B	SE B	β	B	SE B	β	B	SE B	β	B	SE B	β
Step 1:
Network Emotional Support	.01	.02	.04	.01	.02	.03	.02	.02	.07	.01	.02	.05	−.00	.01	−.02	−.00	.01	−.02
Maternal Age	.00	.06	.00	.00	.06	.00	.05	.06	.07	.06	.06	.08	−.02	.03	−.04	−.02	.03	−.04
Friend Emotional Support	.06	.04	.12	.08	.04	.15	.05	.04	.10	.08	.04	.17*	.02	.02	.07	.02	.03	.06
Step 2:
Friend Emotional x				.03	.03	.08				.07	.03	.18*				−.00	.02	−.02
Maternal Age
R2		.02			.02			.02			.05*			.01			.01

Note. (N = 168).

^*p <.05

Figure 1.

Maternal age moderates the relation between friend emotional support and maternal growth-fostering behaviors. *p <.05

Regression results for friend socializing support indicated no main or interaction effects on either of the parenting behaviors (Table 3). There was, however, a significant negative main effect of network socializing on maternal sensitivity. In regard to child care, regression indicated friend child care was not related to maternal sensitivity. However, results indicated that the interaction of friend child care and maternal age significantly predicted both cognitive growth-fostering and detachment, above and beyond network child care support (Table 4). The interaction accounted for 3% of the variance in growth-fostering, and 6% of the variance in detachment. Results from simple slope analyses for cognitive growth-fostering (Figure 2) indicated that the slope of friend child care and cognitive growth-fostering was significant for younger (t = −2.23, p = .02), but not for older mothers (t = .97, p = .33). Region of significance analyses indicated that friend child care significantly predicted the less cognitive growth-fostering for mothers who were younger than 18.7 (range = 15.8 to 18.7; all t’s < −1. 97; all p’s < .05).

Table 3.

Hierarchical Regression Predicting Parenting Behavior from Friend Socializing Support

	Sensitivity						Cognitive Growth Fostering						Detachment
	Model 1			Model 2			Model 1			Model 2			Model 1			Model 2
Predictors	B	SE B	β	B	SE B	β	B	SE B	β	B	SE B	β	B	SE B	β	B	SE B	β
Step 1:
Network Socializing Support	−.36	.15	−.18*	−.37	.15	−.19*	−.18	.15	−.09	−.19	.15	−.10	.02	.09	.02	.03	.09	.03
Maternal Age	−.01	.06	−.01	.02	.07	.03	.05	.06	.07	.07	.07	.10	−.02	.03	−.05	−.05	.04	−.12
Friend Socializing Support	−.01	.16	−.00	−.01	.16	−.00	−.31	.16	−.15	−.31	.16	−.15	.06	.10	.05	.06	.10	.05
Step 2:
Friend Socializing x				−.09	.13	−.06				−.07	.13	−.05				.10	.08	.12
Maternal Age
R2		.03			.04			.03			.04			.01			.02

Note. (N = 168).

^*p <.05

Table 4.

Hierarchical Regression Predicting Parenting Behavior from Friend Child Care Assistance

	Sensitivity						Cognitive Growth-Fostering						Detachment
	Model 1			Model 2			Model 1			Model 2			Model 1			Model 2
Predictors	B	SE B	β	B	SE B	β	B	SE B	β	B	SE B	β	B	SE B	β	B	SE B	β
Step 1:
Network Child Care Support	−.01	.02	−.04	−.01	.02	−.02	−.01	.02	−.04	−.00	.02	−.01	−.01	.01	−.04	−.02	.01	−.09
Maternal Age	−.01	.06	−.01	−.05	.06	−.06	.05	.06	.07	−.00	.06	−.00	−.02	.03	−.05	.02	.04	.05
Friend Child Care Support	.10	.20	.04	.08	.20	.03	−.18	.20	−.07	−.21	.20	−.08	.03	.12	.02	.05	.12	.04
Step 2:
Friend Child Care x				.26	.17	.13				.37	.17	.18*				−.31	.10	−.26**
Maternal Age
R2		.00			.02			.01			.04*			.00			.06**

Note. (N = 168).

^*p <.05,

^**p <.01

Figure 2.

Maternal age moderates the relation between friend child care support and maternal growth-fostering behaviors. *p <.05;

Figure 3 depicts age as moderating the child care-detachment association. Simple slope analysis indicated that both slopes were significantly different from zero, but in the opposite directions (for younger mothers, t = 2.55, p = .01; for older mothers, t = −2.2, p = .03). Region of significance analyses indicated that for mothers younger than 18.8 (range = 15.8 to 18.8; all t’s > 1. 97; all p’s < .05) friend child care support was associated with the display of greater detachment. However, for mothers who were at least 20.7 years old (range = 20.7 to 21.6; all t’s < − 1. 98; all p’s < .05), child care support was associated with less detachment.

Figure 3.

Maternal age moderates the relation between friend child care support and maternal detachment. *p <.05

Discussion

We examined social support from friends, a developmentally salient source of support, as a potential protective factor for the parenting adjustment of young Latina mothers. We found that the associations between friend social support and parenting depended on the specific type of support (emotional, socializing, and child care support) and the mother’s age. These differential relations may account for the mixed findings in the previous literature, and highlight the utility of using a developmental lens to provide a more nuanced picture of the role of social support for Latina adolescent mothers’ parenting.

Consistent with our conceptualization, no relation emerged between an overall composite of friend support and any parenting behavior. Instead, the different support types we assessed showed varied associations with parenting. This finding is consistent with research that implicates support type as an important construct in the measurement of social support with adolescent mothers (Nath et al., 1991). Furthermore, although mothers of different ages were similar in parenting behaviors and reported similar levels of friend support, friend provision of emotional and child care support showed differential effects on maternal parenting depending on the mothers’ age. Regarding emotional support, mothers’ utilization of more frequent emotional support from friends related to the display of more desirable parenting during social interaction with their toddlers, but only for mothers aged 19.5 years or older. For younger mothers, emotional support appeared to have no effect on their parenting. The finding of a positive association for emotional support is consistent with parenting and social support models (Collins et al., 1993; Contreras et al., 2002), and previous literature suggesting the importance of friend emotional support to the psychosocial adjustment of parenting and non-parenting adolescents (Richardson, et al., 1991;1995; Way & Chen, 2000). When mothers feel supported emotionally, their parenting may benefit through increased self-esteem and/or parenting confidence. For example, among Mexican-origin adolescent mothers, those who felt more emotionally supported reported higher levels of parenting self-efficacy (Umaña-Taylor et al., 2013).

The lack of association between emotional support and parenting among mothers younger than 19.5 years, although unexpected, is consistent with recent findings indicating that among non-parenting Latina adolescents, perceptions of greater friend support related to lower self-esteem in early adolescence, but related to higher self-esteem for older adolescents (Bámaca-Colbert et al., 2017) . Similarly, Davis (2002) found that friend support related to more depressive symptoms for younger African American adolescent mothers, but to less for older adolescent mothers. Although the reason for the lack of effect for younger mothers in our study is not clear, perhaps, younger Latina mothers are more embedded in their families, thus making familial support more relevant for their parenting adjustment. This differential effect by age may be especially relevant for Latina adolescent mothers given research documenting a stronger reliance on family versus external sources of support among Latina adolescents, compared to those of other ethnic groups (Chang, Chen, & Alegría, 2014). Nonetheless, because few previous studies examined potential differences by age, it remains unclear whether the lack of effect for younger mothers is unique to our sample.

The effects of friend child care support on the young mother’s parenting also depended on her age. Among mothers aged 18.8 years or younger, friend child care support related to the display of less growth-fostering behaviors and more detachment. Although young mothers may feel supported when friends provide child care, they may end up with fewer opportunities to engage in and develop their parenting skills. It may also be that young mothers who are less involved in parenting may solicit more child care support from friends. While this support may allow them to engage in other activities, such as attending school or work, it does not appear to help facilitate their parenting engagement and development of skills. The young mothers’ reliance on friends for child care assistance may also reflect a relative lack of child care support from family members, who are the more normative sources of this type of support for young adolescents (Grau et al., 2012). To the extent that this is the case, factors related to lack of child care support from family members (e.g., family conflict, isolation) may help account for the negative associations we found for friend support among the younger mothers. Longitudinal studies that assess friend and family support across time could clarify the mechanisms that may account for this negative effect.

In contrast to this negative association among younger mothers, among mothers 20 years or older, those reporting child care support from friends interacted in a less detached manner with their children. These mothers may be within the developmental period when relying on friends for child care support may be more normative. As is the case for adult mothers, these mothers appear to be able to benefit from this support (Ceballo & McLoyd, 2002). Our findings suggest that for older mothers, the availability of child care from friends may facilitate their engagement in parenting. However, the negative relations we found for younger mothers raise concern and point to further investigation of the sources of child care available to Latina adolescent mothers and the repercussions of these for their adjustment as adolescents and as parents.

Our results regarding the specific age ranges at which friend support related to maternal behavior indicate important differences between mothers who are less than 19 years of age and those who are above 20 years of age. These different effects may be reflective of cognitive and social-emotional shifts corresponding to developmental transitions between adolescence and adulthood (Steinberg & Morris, 2001). Specifically, the results suggest that friend emotional and child care support are unrelated or negatively related to parenting for mothers in late adolescence, but serve protective roles for parenting during early adulthood. There is limited recent research with European American and African American adolescent mothers, as well as non-parenting Latina adolescents, which demonstrates the differential effects of age (Bámaca-Colbert et al., 2017; Davis, 2002; Ensor & Hughes, 2010). Taken together, these findings argue for the consideration of maternal age in shaping the ways in which friend support influences adjustment. They also highlight the need to assess the types of friend support as separate measures in an ecological context, taking into account the mother’s individual, developmental, and cultural characteristics.

In a complementary goal, the study explored characteristics of maternal friend networks within the cultural context of these young mothers. Mothers reported one to two friends, on average, who were likely to be female neighbors of similar age. Neighborhood-based friendships may be particularly common among ethnic minority adolescents from low-income families (Way & Chen, 2000). The number of friends reported in the current study may reflect a preference for one or two close friendships, rather than a large group of friends, among adolescent mothers, which was described in a qualitative study of African American adolescent mothers in Chicago (University of Chicago, 2011). Although some prior studies reported one to two friends on average (Ensor & Hughes, 2010), others reported averages of two to three friends (Richardson et al., 1995; Voight et al., 1996). Nonetheless, given that these prior studies have not included Latina adolescent mothers, it is not clear how representative of Latina adolescent mothers the current findings are. Friends were a considerable source of support for Latina adolescent mothers in this sample, as 60% of mothers received at least one type of support from friends during the previous month. Mothers utilized friends most frequently for emotional (39%) and socializing (33%) support, which is consistent with the most prevalent types of support from friends reported by African American adolescent mothers (Voight et al., 1996) and non-parenting Latina adolescents (Stanton-Salazar & Spina, 2005).

To consider another layer in the ecological framework of parenting (Contreras et al., 2002), we explored differences between immigrant and nonimmigrant Latina mothers in number of friends and amount of friend support they reported. Although immigrant mothers reported a similar number of friends as US-born mothers, immigrant mothers reported receiving less friend emotional support than US-born mothers. These findings are in line with the traditional cultural value of familismo, which emphasizes the importance of support within the family, and the literature documenting Latino adolescents’ preference for family versus outside support (Suarez-Orozco & Suarez-Orosco, 1995). However, considering that friend emotional support appears to be beneficial, especially for older adolescent mothers, this support may be especially important for immigrant mothers who may not have access to their extended family due to migration. Similarly, future research should consider the generational status of Latina adolescent mothers especially regarding the sources of support that are available to them.

Limitations

Despite its contributions, our study has limitations that need to be considered. Given the cross-sectional design of the study, the direction of these effects cannot be determined. For instance, both competent parenting and more friend emotional support may be predicted by another factor, such as greater maternal socio-emotional competence. To reach conclusions about directionality, longitudinal studies are needed. The current sample is a convenience sample, recruited primarily in pediatric clinics serving the low-income Latino neighborhoods in a large Midwestern city, and is not necessarily representative of all parenting Latina adolescents in that city. Similarly, findings cannot be generalized beyond Latina mothers of primarily Puerto Rican origin, who themselves belong to an under-researched group in comparison to Mexican-American adolescents, and thus more research is needed with other Latina adolescent mothers. We followed prior literature (Bornstein & Putnick, 2007) that used age as a proxy for maternal developmental level or maturity. Nonetheless, it would be helpful for studies to assess other variables that distinguish these levels, such as facets of identity exploration and autonomy development. In measuring social support, the current study relied solely on the mothers’ report. Future studies could be strengthened by including the report of friends, as well as gathering more information regarding the reported supportive interactions. The current study selected types of support that were hypothesized to be especially relevant to friends of adolescent mothers. However, studies may also benefit from examining other types of support (i.e., informational support). We focused on received support given research indicating its relation to perceived value of support (Devereux et al., 2009) and the ease of translation to interventions for high risk populations (Rini et al., 2006). However, perceived support is also valuable in its link to psychological adjustment (Uchino, 2009) and it is worth examining both in future studies. Due to lack of normality, we dichotomized our child care and socializing variables, therefore we cannot draw conclusions about differences between mothers utilizing different levels of these types of support. Only one type of support provided by the adolescents’ networks related to parenting in our sample. Specifically, mothers who reported more socializing support from their networks displayed less maternal sensitivity. Whereas socializing interactions may provide useful opportunities for exploration and identity development, they may not facilitate the adolescents’ development of sensitive parenting behaviors. Nonetheless, this overall lack of association for network support precluded a better understanding of the relative contribution of friend support. Future studies should examine further the ways in which the social support networks of adolescent mothers can promote their adjustment. This research could also be improved by gathering more information regarding individual and demographic characteristics of the adolescents’ friends, especially their parenting and pregnancy status. Pregnant and parenting friends may be particularly valuable sources of support, as they can assist with child care support and exchange advice and emotional support through shared experiences (Herrman, 2008; University of Chicago, 2011).

In spite of these limitations, our findings provide support for Contreras and collaborators’ (2002) model of parenting among Latina adolescent mothers by documenting differential associations of friend support on parenting based on maternal age and the type of support that is provided. Differences in friend support between mothers born in and outside the mainland U.S. also highlight the importance of considering maternal immigration and cultural background to more fully understand social support processes in Latina families (Contreras et al., 2002). Continued research that explores the protective role of friend support for the parenting of young Latina mothers, using a developmental and culturally-informed lens, can provide information to strengthen interventions for this population.