Direct, indirect, and buffering effects of support for mothers on children's socioemotional adjustment.

Defining Social and Formal Support

In common with other researchers examining social support from family and friends ( Bonds, Gondoli, Sturge-Apple, & Salem, 2002 ; Choi & Pyun, 2014 ; Heberle, Krill, Briggs-Gowan, & Carter, 2015 ; Herwig, Wirtz, & Bengel, 2004 ; C.-Y. S. Lee, Anderson, Horowitz, & August, 2009 ; McConnell et al., 2011 ; Östberg & Hagekull, 2013 ; Ryan, Kalil et al., 2009 ), we did not examine enacted (received) support when assessing benefits for child adjustment. This is because of the difficulty in distinguishing preventive support (which may predict better outcomes) from responsive support that is related to need (and consequently often associated with poorer outcomes, especially in cross-sectional analyses). For social support, we followed others cited by drawing on two overlapping constructs: “social embeddedness” (number and quality of social ties) and “perceived support” (availability and adequacy), both presumed to signal a parent’s ability to draw on actual support from family and friends ( Barrera, 1986 , p. 415).

Observational studies of the impact of formal support for parenting from professional health and welfare services on child adjustment are inconclusive, reflecting reliance on measures of enacted support likely to be conflated with family need ( Leventhal et al., 2000 ; Spielberger & Lyons, 2009 ). To measure formal support, we adopted a similar approach to that commonly taken for social support, using parental perceptions of availability and adequacy. Despite widespread provision of family support services in many Western countries, many parents perceive barriers to access ( Ghate & Hazel, 2002 ; Whittaker & Cowley, 2012 ). Barriers partly relate to low awareness or practical problems surrounding travel, opening hours, or administrative procedures. It is important to note, however, that they also relate to perceptions that available support is inadequate to meet needs and to fears of interference and stigma.

Pathways From Support to Child Socioemotional Adjustment

This study tests pathways from support for mothers to children’s socioemotional adjustment via three family stressors: dysfunctional parenting, maternal distress, and economic strain (see Figure 1 ). In this model, most hypothesized effects of support involve reduced dysfunctional parenting. Social embeddedness has long been conceived as having direct benefits for parenting, via mechanisms such as provision of information and advice, modeling of appropriate behavior, and positive affirmation of a parent’s own attitudes and parenting skills ( Belsky, 1984 , 1990 ; McConnell et al., 2011 ). Perceived availability of formal support might have similar benefits, signaling greater engagement with universal provision of professional information and advice to parents that increases their knowledge and skills. Empirical evidence that parenting quality is linked to support availability, however, has been relatively sparse and confined to studies of social support ( Bonds et al., 2002 ; C.-Y. S. Lee et al., 2009 ). Existing studies of young families’ service use have yielded few conclusions on benefits for parenting ( Maupin, Brophy-Herb, Schiffman, & Bocknek, 2010 ; Spielberger & Lyons, 2009 ). This may reflect a cross-sectional design, coupled with measures of enacted support. Further work is therefore needed to substantiate a path from support to adjustment via parenting alone, as shown in our conceptual model (see Figure 1 ).

Conceptual model depicting paths from support for mothers to child socioemotional adjustment via maternal distress, economic strain and dysfunctional parenting. Solid arrows show hypothesized main paths, and broken arrows show potential additional paths.

Parenting process and family stress models ( Belsky, 1984 ; Conger et al., 1994 ) highlight the importance of parental psychological and economic resources for effective parenting. This points to further paths whereby support could reduce maternal distress and economic strain, both of which in turn affect adjustment via improved parenting (see Figure 1 ). Social embeddedness provides families with opportunities for emotional and instrumental support and is thought to be important in regulating maternal psychological functioning, even without any major stressors ( Barrera, 1986 ; Cohen & Wills, 1985 ; Thoits, 2011 ). Emotional support from family and friends may enable parents to share frustrations over minor daily hassles with a sympathetic audience, thereby preventing further escalation of problems ( Lakey & Orehek, 2011 ). Instrumental support from family and friends may also reduce economic strain ( Harknett, 2006 ), which may sustain maternal psychological functioning indirectly ( Conger et al., 1994 ). Although little is known about the possibility, conceivably mothers’ access to professional services could provide similar sustaining emotional and instrumental support.

Although benefits of social support for mothers’ psychological well-being are well established (see, e.g., Brown, Harris, Woods, Buman, & Cox, 2012 ; Manuel, Martinson, Bledsoe-Mansori, & Bellamy, 2012 ), only a few studies have explored complete pathways to child adjustment via these constructs ( Choi & Pyun, 2014 ; Herwig et al., 2004 ; Östberg & Hagekull, 2013 ). Two were cross-sectional in design, so one cannot be sure about the direction of effects. One longitudinal study found that reduced economic strain, as well as lower parenting stress and better parenting, mediated effects of greater perceived social support on young children’s behavior problems ( Choi & Pyun, 2014 ). However, it contained no explicit measurement of maternal psychological functioning and did not adjust for possible confounders of support−child outcome associations.

Additional pathways from social support not involving parenting (see Figure 1 , dashed arrows) might impact both externalizing and internalizing problems. Maternal distress may model dysregulated behaviors and emotions ( Heberle et al., 2015 ), whereas economic strain could lead to greater adjustment problems, even in young children, via negative social comparisons and poverty-related stigma ( Heberle & Carter, 2015 ).

Buffering Effects of Support

Social support has long been theorized to have a protective or buffering effect on psychological functioning, by offering appraisal and/or coping mechanisms to deal with stressors ( Cohen & Wills, 1985 ). Although buffering was originally conceived in relation to adult psychological functioning, it is possible to extend this idea in relation to children’s socioemotional adjustment. The availability of support may protect children from the harmful effects of disruptive family processes, if parents are able to draw on support for appraisal−coping strategies, and/or if other individuals directly offer the child appraisal−coping strategies that dilute or counteract family stressors. Empirical evidence in relation to child socioemotional adjustment has been limited to studies investigating moderating effects of social support on stressors such as maternal depression and suboptimal parenting. These present mixed findings. In two studies, support buffered effects of lower maternal psychological well-being and suboptimal parenting on children’s externalizing or internalizing behavior problems ( Barnett, Scaramella, Neppl, Ontai, & Conger, 2010 ; Heberle et al., 2015 ). In contrast, another study found support was less, rather than more, effective in protecting children from severe maternal depression ( L.-C. Lee, Halpern, Hertz-Picciotto, Martin, & Suchindran, 2006 ), whereas two others found no moderating effects of social support on the effects of parenting stress ( McConnell et al., 2011 ; Östberg & Hagekull, 2013 ; Ryan, Tolani, & Brooks-Gunn, 2009 ). Although not previously explored, perceived availability of formal support may have a similar protective role in times of stress. It may signal access to assistance with parenting problems, as well as access to other forms of support relieving parental psychological distress and economic strain, such as couple relationship counseling, substance abuse treatment, or access to welfare payments. Further research is needed to establish whether formal, as well as social, support generally has protective buffering effects in relation to children’s adjustment or whether its action is compromised by contextual strains such as maternal depression.

Study Hypotheses

Our study explores pathways from support for mothers to two aspects of children’s socioemotional adjustment (externalizing and internalizing behavior problems) via three family stressors in our conceptual model: maternal distress, economic strain, and dysfunctional parenting. We also explore whether support moderates associations between the stressors and child adjustment. We aimed to test the following hypotheses:

Hypothesis 1: Both social and formal support will predict better child socioemotional adjustment.

Hypothesis 2: Positive effects of both types of support will be mediated via less economic strain, maternal distress and dysfunctional parenting.

Hypothesis 3: Both types of support will have buffering effects in reducing the impact of economic strain, maternal distress, and dysfunctional parenting on child adjustment.

Based on existing literature, we put forward similar hypotheses in relation to both aspects of child adjustment studied.

To strengthen causal inference, our study has a prospective design with social and formal support for mothers measured in the vulnerable early years of children’s lives (infancy and toddlerhood, 10–22 months) and before behavioral problems are likely to develop. We examined the influence of support on trajectories of socioemotional adjustment in middle childhood (measured from approximately ages 6 to 10), because as noted this signals adolescent and young adult risk. Potential mediators are measured at an intermediate time point (ages 4–5). To help overcome potential hazards associated with selection effects and maternal bias, our analyses adjusted for a wide range of confounders measured in infancy that are associated with support, mediators, and outcome variables, including baseline maternal mental health and socioeconomic information.

Method

Data were from the Growing Up in Scotland study’s first birth cohort (children born 2004–2005; further details available in Bradshaw, Tipping, Marryat, & Corbett, 2007 ). Baseline data were gathered from 5,217 families in 2005–2006, when children were 10 months old. Families were followed up annually for 5 years (to 70 months) and then at approximately two-year intervals (94 and 122 months). Each data collection sweep was subject to medical ethical review (Scotland “A” Multi Research Ethics Committee), with mothers or caregivers giving informed consent.

This study used data from computer-assisted personal interviews with the main caregiver. We excluded 93 families with multiple births and a further 103 families where the main caregiver interviewed at child age 10 and 22 months was not the natural mother. Of the remaining 5,021 families, 3,598 (72%) were followed up at child age 46 and 58 months. We further excluded 62 families where the child’s natural mother was not the main caregiver interviewed about potential mediators at these ages, giving an eligible sample of 3,536 families. Of these, 3,031 families (86%) were followed to the final time point. To provide consistent reporting of child outcomes, we restricted the analysis sample to cases where the mother was interviewed at all relevant outcome time points (70, 94, and 122 months) and provided outcome information on at least one of these occasions ( n = 2,649; 87% of the complete eligible sample follow-up). The analysis sample contained fewer mothers with low support and low educational qualifications compared to a complete follow-up of the eligible sample but did not differ regarding other covariates.

All measures were based on information supplied by the child’s natural mother.

Adjustment was measured at 70, 94, and 122 months using the Strengths and Difficulties Questionnaire ( R. Goodman, 1997 ). Items ask for agreement with statements concerning the child, with response options rated on this 3-point scale: 0 ( not ), 1 ( somewhat ), and 2 ( certainly true ). Scores are nonnormally distributed ( Stone, Otten, Engels, Vermulst, & Janssens, 2010 ). Externalizing problems used the combined conduct problems and hyperactivity−inattention five-item subscales (Cronbach alphas at each age = .74–.80); internalizing problems used the combined peer relationship and emotional problems five-item subscales (Cronbach alphas = .61–.76). Externalizing and internalizing scores have good convergent and discriminant validity across informants and with respect to clinical disorder ( A. Goodman, Lamping, & Ploubidis, 2010 ).

Support measures

A factor analysis of all support items used in this study found a two-factor solution, with all items described for social support loading onto one factor (loadings = .5–.7) and all items described for formal support (including two items that do not make specific reference to professional services) loading onto the second factor (loadings = .4–.6).

Social support

A standardized scale (Cronbach’s alpha = .65) was created from four items, measured at 22 months, concerning support for mothers from family and friends. The first item asked: “Not counting people who live with you, which of the following statements best describes how many people you have a close relationship with?” rated on this 4-point scale: 1 ( I have close relationships with lots of people ), 2 ( I have close relationships with some people ), 3 ( I have one or two close relationships ), and 4 ( I don’t have any close relationships ). The second asked: “Thinking about your immediate family (parents and brothers or sisters) living elsewhere, can you tell me how much you agree or disagree with the following statement: ‘I feel close to most of my family’?” rated on a 5-point scale ranging from 1 ( agree strongly ) to 5 ( disagree strongly ). The third asked for agreement with the statement “My friends take notice of my opinions,” rated on a 5-point scale ranging from 1 ( agree strongly ) to 5 ( disagree strongly ); mothers reporting no family ( n = 9; .3%) or no friends ( n = 25; 1.2%) were recoded as disagree strongly. 1 Last, mothers were asked: “Overall, how do you feel about the amount of support or help you get from family or friends living elsewhere?” rated on a 3-point scale ranging from 1 ( I get enough help ) to 3 ( I don’t get any help at all 2 ).

Formal support

A standardized scale (Cronbach’s alpha = .64) used six items administered at 10 and/or 22 months (one item was presented at both ages) from a previous study of support among low-income families ( Ghate & Hazel, 2002 ) that asked mothers’ agreement with statements concerning parenting advice available from professionals such as health visitors. Responses were rated on a 5-point scale ranging from 1 ( strongly agree ) to 5 ( strongly disagree ). The item administered at both ages was “If you ask for help or advice on parenting from professionals like doctors or social workers, they start interfering or trying to take over.” Those at 10 months only were “It’s difficult to ask people for help or advice about parenting unless you know them really well” and “It’s hard to know who to ask for help or advice about being a parent.” Those at 22 months only were “Professionals like health visitors and social workers do not offer parents enough advice and support with bringing up their children” and “If other people knew you were getting professional advice or support with parenting, they would probably think you were a bad parent.”

To validate the social support measure, we examined associations with instrumental support received when the child was age 34 months. Mothers perceiving high social support were more likely than mothers reporting low social support to have weekly grandparental child care for an hour or more (72% vs. 32%, p < .001) and to receive grandparental weekly help with household chores or purchases (42% vs. 27%, p < .001). It was not possible to validate attitudes to formal support against later receipt. However, it is important to note that attitudes are unlikely to be based solely on prejudice and hearsay, because all mothers received routine universal postnatal support from health visitors referred to in the interview items. Attitudes were also associated with use and perceptions of universally available antenatal support measured in the survey (details are available on request).