Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Aug 1.
Published in final edited form as: J Marriage Fam. 2019 Apr 8;81(4):953–967. doi: 10.1111/jomf.12573

Informal Networks of Low-Income Mothers: Support, Burden, and Change

Melissa Radey a, Lenore M McWey b
PMCID: PMC6867802  NIHMSID: NIHMS1019222  PMID: 31749503

Abstract

Objective:

The authors examined the support and burden of low-income, urban mothers’ informal networks.

Background:

Living or growing up in poverty strongly predicts barriers and instability across several life domains for mothers and their children. Informal networks can play a critical role in promoting maternal and child well-being particularly in the midst of poverty. Understanding informal support and the reciprocal burden it may create is especially relevant for low-income families living with a reduced public safety net in the post-welfare reform era. Therefore, study aims were to measure support and burden among low-income mothers and determine if support and burden change over time.

Method:

Data were from the Welfare, Children, Families (WCF) project, a longitudinal study of 2,400 low-income, caregivers of children and adolescents living in Boston, Chicago, or San Antonio (http://web.jhu.edu/threecitystudy/index.html)). We applied latent class analyses to support and burden indicators in four domains—emotional, favor, child care, and financial.

Results:

Results supported four profiles of informal networks – healthy, unhealthy, burden only, and support only. Although most mothers had healthy informal networks, approximately one-third experienced no support or support imbalance which related to network changes at later time points. Demographic characteristics largely were not predictive of support profile or profile change.

Conclusion:

Although many mothers had healthy support and burden, the most vulnerable did not have consistently healthy informal networks. The identification of a sizable minority of low-income mothers who cannot consistently rely on informal support is significant in light of diminished formal supports available to children and families.

Keywords: social support, kinship, latent class analysis, low-income families, poverty, family relations

Approximately one in ten U.S. families live in poverty, including 28% of single-mother families and 18% of all children (Fontenot, Semega, & Kollar, 2018). Despite the detrimental impacts of poverty on health, educational outcomes, and lifestyle behaviors (for review see Edin & Kissane 2010), poor families in the U.S. benefit less from the public safety net than families in other industrialized nations (IOM, 2013). In 1996, the Personal Responsibility Work Opportunity Reconciliation Act replaced the formal cash safety net with an employment-based system. This increased poor families’ reliance on informal supports (Levine, 2013). Post-welfare reform, the percentage of families “disconnected” from employment and cash welfare grew from 12% of low-income single mothers in 2004 to 20% in 2008 (Loprest & Nichols, 2011), and welfare redistribution spending across all government programs decreased among the poorest families (Moffitt, 2015). Of eligible families, TANF utilization dropped from 68% in 1996 to 23% in 2017 (Floyd, Pavetti, & Schott, 2017).

Can families and neighbors pick up where government benefits leave off? Low-income mothers rely on family and friends to meet basic needs (e.g., Edin & Lein, 1997; Levine, 2013). Clear and substantial evidence attests to the importance of informal support for maternal and child well-being among low-income families (e.g., Huang, Costeines, Kaufman, & Ayala, 2014; Taylor & Conger, 2017). Yet, limited support in disadvantaged contexts can also create burden, defined as network obligations or relational stress (Offer, 2012). The post-welfare environment may increase mothers’ exposure to family or friends who drain their resources (Levine, 2013). As such, low-income mothers often have to make calculated decisions about whether to call on family or friends based on what is expected in return (Edin & Shaefer, 2015; Levine, 2013; Stack, 1974). Additionally, instability (e.g., family composition change, housing loss) is common among mothers in poverty, and that may translate to unreliable support (Sandstrom & Huerta, 2013; Schenck-Fontaine, Gassman-Pines, & Hill, 2017). Yet, research has largely treated support as unidirectional and time constant. We know little about the potential reciprocation of support over time. Therefore, the aims of this study were to measure informal support and burden among low-income mothers, and determine if support and burden change over time.

Conceptual Framework

Informal support is a tool for low-income families to negotiate their disadvantaged environments (e.g., Harknett & Hartnett, 2011; Taylor & Conger, 2017). The broad concept of informal support captures how resources (e.g., emotional, practical, financial) from one’s informal network foster individual and societal functioning and wellbeing (Brownell & Shumaker, 1984). To consider the functionality of low-income informal networks, we examined both informal support and burden. With our interest in support availability and burden, we define support as the functional components of relationships (Cohen, Mermelstein, Kamarck, & Hoberman, 1985) to meet basic needs, and those who perceive support for basic needs as having “healthy support.” However, low-income informal networks are typically bidirectional (Mazelis, 2017; Dominguez, 2011). Exchange theories contend that to acquire and maintain informal support, individuals must deliberately invest resources (Bourdieu, 1986; Portes, 1998) or assume a level of burden (Offer, 2012). We term this investment as “healthy burden.” Conceptualized here, healthy burden (i.e., some, but not excess), also contributes to network functioning (e.g., Levine, 2013; Offer, 2012). We use the term “healthy” to describe support and burden recognizing “healthy” levels may not—and likely do not—eliminate resource scarcity that mothers experience in low-income environments (Belle, 1983; Mazelis, 2017). Rather, perceptions of informal support promote maternal and child wellbeing (Meadows, 2009; Ryan, Kalil, & Leininger, 2009).

Drawing from economics and behaviorism, people negotiate and mobilize social relationships in order to profit by maximizing material and emotional rewards while minimizing costs (Blau, 1964). To access informal support, individuals enter transactional relationships with expectations, yet without explicit obligations or timelines for such obligations (Portes, 1998). For example, a mother may offer to provide childcare for her niece with the expectation that her sister (or another network member) will reciprocate the favor—either through child care or another form (e.g., loan, favor, housing)—upon request. For long-term functionality, individuals perceive balanced and fair network relationships (e.g., Nelson, 2005).

The concept and dimensions of informal support and burden vary across studies. Typically, studies differ in the included dimensions (e.g., emotional, practical, financial), network sources (e.g., partner, family member, caseworker), and realization (i.e., received, perceived) of support or burden. Based on our interest in informal support and burden associated with maternal and child wellbeing, we considered emotional, practical, child care, and financial dimensions from family or friends. In terms of realization, measures of perceived support capture subjective evaluations of network capacity in specified domains and avoid conflating need with access (Harknett, 2006); they distinguish between mothers who did not receive support due to a lack of need versus due to a lack of access. Perceived support also is more strongly associated with health and wellbeing than received or actualized support (Turner & Turner, 1999; Wethington & Kessler, 1986). Following others examining support of low-income mothers (e.g., Harknett & Hartnett, 2011; Ryan et al., 2009), coupled with our interest in capturing support and burden influential to wellbeing, we examined perceived support and burden.

Although much qualitative evidence attests to the functioning of exchange and reciprocity in informal networks (Edin & Lein, 1997; Levine, 2013; Offer, 2012; Stack, 1974), unbalanced transactions in disadvantaged environments can create, rather than alleviate, stress and burden (Belle, 1983; Levine, 2013; Offer, 2012). Mothers may perceive fragmented and unhealthy networks such that mothers experience stressful relationships and unhealthy burden without the advantage of support availability (Offer, 2012; Radey, 2015). Rather than following a pattern of exchange, resource-constrained environments may mean that mothers perceive support not contingent on healthy burden. Alternatively, mothers may experience unhealthy burden to network members without reciprocity expectations due to the depths of network disadvantage.

Informal Support and Burden of Low-income Mothers

Extensive research highlights the importance of perceived reciprocity in informal networks among low-income families (e.g., Dominguez, 2011; Edin & Lein, 1997; Gazso, McDaniel, & Waldron, 2016; Stack, 1974). In her seminal ethnography of a low-income, Black community in the Midwest, Stack (1974) uncovered complex systems of cooperation among kin including family and close friends. Similarly, in Dominguez’s (2011) ethnography of low-income Latinos, mothers commonly relied on their own mothers to provide childcare in exchange for financial and material contributions to the household. Despite potential race or ethnic differences with Black and Hispanic mothers accessing more (e.g., Brewster & Padavic, 2002; Hogan, Eggebeen, & Clogg, 1993) or less support (Hogan, Hao, Parish, 1990; Turney & Kao, 2009), low-income mothers commonly participate in informal networks to get by. Yet, network reliance often comes at a cost (Offer, 2012). Among low-income mothers both before and after welfare reform (N = 95), Levine (2014) found mothers recognized that receiving support often meant network obligations in some form. Cohesive, trusting networks provided mothers critical resources; disorganized, untrusting ones, often a result of mothers’ experiences with prior unreciprocated support, increased mothers’ deprivation.

Despite ethnographic evidence of informal networks among low-income mothers, often due to data limitations, quantitative studies typically measure informal support without considering potential burden (e.g., Harknett & Hartnett, 2011; Turney & Kao, 2009). Limited studies suggest that mothers most in need and, likely, the least-positioned for reciprocity, perceived less support than their more advantaged counterparts (e.g., Harknett & Hartnett, 2011; Meadows, 2009; Radey & Brewster, 2013). In a longitudinal examination of support availability among mothers of young children, less availability of instrumental and emotional support was related to poverty, poor physical health, and poor mental health (Harknett & Hartnett, 2011). In addition, limited evidence suggests that support provision does not translate to available support (Pilkauskas, Campbell, & Wimer, 2017; Radey, 2015). Using Welfare, Children, Families study, the data used in the current study, Radey (2015) found that mothers with excess burden had lower levels of support compared to mothers with less burden. Similarly, providing money to network members was positively related to experiencing a hardship in a national sample of low-income mothers suggesting that providing support does not necessarily translate to available support (Pilkauskas et al., 2017). Despite the potential burden of networks, a recent analysis of a nationally-representative sample of Black Americans indicated that 87% and 83% of respondents exchanged resources with family members and friends, respectively (Taylor, Mouson, Nguyen, & Chatters, 2016).

Given social network variability, recent qualitative and quantitative work suggest the importance of longitudinal studies (Gazso et al., 2016). Employing a life-course perspective to analyze qualitative interviews for both the give and take of low-income support networks, researchers found “more changeable than durable” networks (Gazso et al., 2016, p. 441). Complimenting earlier findings of economic, household, residential, and child care instability among poor families (Hill, Romich, Mattingly, Shamsuddin, & Wething, 2017; Sandstrom & Huerta, 2013), major life events (e.g., divorce, addiction, violence, and single parenthood), common when juggling life demands and poverty, contributed to the “changing, borderless character” of social networks (Gazso et al., 2016, p. 451). Documenting network changes is important given potential consequences. Although limited work has examined the consequences of network variability, a linear probability model analysis using children of Survey of Income and Program Participation respondents showed that support instability increased child food insecurity and worsened child health (Wolf & Morrissey, 2017).

Study Contribution and Hypotheses

Low-income mothers often rely on informal networks in order to get by, particularly in the post-welfare reform era with limited public assistance availability (Loprest & Nichols, 2011). In this study, we focus on informal support and burden among low-income mothers because of the weak public safety net coupled with support’s influential role in both maternal and child wellbeing (IOM, 2013; Taylor & Conger, 2017). Using exchange theories to improve our understanding of the informal networks of low-income mothers, we examined (a) both support and burden, and (b) changes over time. Based on available evidence, we hypothesized:

  1. the largest profile of low-income mothers will perceive healthy support and burden while the smallest profile will perceive neither healthy support nor burden.

  2. vulnerability will exacerbate network functioning; mothers with particularly low incomes or low educations will face higher levels of unhealthy support or burden than their more advantaged counterparts.

  3. support and burden profiles will vary over time and vulnerable mothers will be more susceptible to change.

Method

We analyzed data from the Welfare, Children, Families (WCF) project, a longitudinal study of low-income, caregivers of children and adolescents living in Boston, Chicago, or San Antonio (93% of whom were children’s biological mothers, hereafter referred to as “mothers”). Designed to provide insight into low-income families’ lives post-welfare reform, sampled mothers had children between ages 0–4 or 10–14 at baseline (Angel et al., 2009). The stratified, random sample targeted Black, Hispanic, and White households living in neighborhoods with at least a 20% poverty rate per the 1990 Census. Within eligible neighborhoods, sampled households included English- or Spanish-speaking respondents living below 200% of the poverty threshold, oversampling respondents living in poverty (<100%). The face-to-face, in-home study first collected data on mothers and their children in 1999 (Wave 1), then again approximately 1.5 years later in 2000–2001 (Wave 2), and again 5 years later in 2005–06 (Wave 3). The project had a Wave 1 response rate of 75% with similar rates across cities with Waves 2 and 3 response rates of 88% and 84%, respectively (Angel, Burton, Chase-Lansdale, Cherlin, & Moffitt, n.d.).

In this study, we used Wave 1 data to describe mothers’ informal networks, to conduct the initial latent class analysis (LCA), and to consider the role of demographic characteristics in contributing to profile membership. Next, we conducted LCAs on later waves of data to consider change in latent profile at later time points. The Wave 1 LCA includes 2,334 mothers (98% of sample), all of those with valid data on the informal network measures and demographic variables. Due to study design and their small number (n = 27), we excluded mothers who did not identify as Hispanic, non-Hispanic Black, or non-Hispanic White. To examine latent profile change, the subsequent analyses included all mothers with valid data at Waves 1 and 2 (n = 2,006) and at Waves 1 and 3 (n = 1,600). In some instances, the caregiver changed at Wave 2 (n = 56) or Wave 3 (n = 52), and we excluded these cases. Bonferroni-corrected t-tests and chi-square tests indicated that included mothers were similar to those with missing data on all support, burden, and individual characteristics with the exception that included mothers at Waves 1 and 3 were approximately one year younger and had children approximately one year younger. At Wave 2, included mothers were less likely to have postsecondary education compared to those with missing data. Therefore, included mothers generally resembled mothers with missing data on the main study variables of interest. All descriptive analyses and post-LCA analyses employed weights to represent mothers living in low-income neighborhoods in Boston, Chicago, and San Antonio and to adjust for the clustered, stratified sample, and item non-response. The LCA analyses used a three-step approach, described the Analytic Strategy section, appropriate for unweighted data that considers the clustered nature of the observations (Vermunt, 2010).

Measures

Support and burden.

We measured mothers’ support and burden in four realms: emotional (i.e., listen to your problems), practical (i.e., do small favors), childcare (i.e., take care of your children), and financial (i.e., loan you money). To measure support, mothers were asked whether they had “no one”, “some,” or “enough” support in each area. To measure burden, mothers were asked if the number of people that needed them to help in each area was “no one,” “too many,” “only a few,” or “as many as you can handle.” To conduct the LCA based on a contingency table of categorical variables (Kim, Zarit, Fingerman, & Han, 2015), we conducted correlations between support and burden variables to consider the relationship between burden and support. The correlational analyses revealed that mothers with “no demands” or “excess demands” perceived less support. Based on prior work on the significance of reciprocity for network functionality (e.g., Offer, 2012), the importance of having at least someone to turn to when in need (e.g., Harknett, 2006), and the correlational analysis, we dichotomized support (enough/some versus no one) and burden (as many as you can handle/only a few versus too many/no one). The dichotomies allowed us to distinguish between those who had at least someone in an area and those with healthy burden from those without; they also ensured a manageable number of cells in the data matrix. To examine latent profile change, we created a variable indicating whether or not mothers changed latent profiles after Wave 1.

Covariates.

Selected for their potential influence on support and burden, we included several maternal demographic characteristics collected at Wave 1 to predict latent profile membership: race and ethnicity, nativity, age, relationship status, number of children in the household, and poverty status. In addition, we included child age in months. Mothers self-reported their race and specified whether or not they were of Hispanic origin. From these categories, we created three mutually-exclusive categories of sufficient size for analysis: Hispanic, non-Hispanic Black, and non-Hispanic White. To measure nativity, mothers specified their country or origin and we distinguished between those born in the US from those who were not. Age was measured in years and calculated from the mother’s birthdate. Relationship status included three categories: single, cohabiting, or married. Education level included less than high school or GED, high school diploma or GED only, and more than high school diploma or GED. We measured number of children based on the household roster and the number of household members under the age of 18. To consider household resources, an income-to-needs ratio was calculated from the household size and the household income from the prior month and imputed from other income variables when necessary (see Angel et al., 2009 for calculation). We also included child age in months and interview city. Table 1 displays the sample’s descriptive statistics.

Table 1.

Weighted Means (Standard Errors) and Percentage Distributions of Demographic and Support Characteristics of Mothers: Welfare, Children Families, A Three-City Study (n = 2,334)

Demographic Characteristics
Race and Ethnicity
 Hispanic 53.8
 Non-Hispanic, Black 41.0
 Non-Hispanic, White 5.2
Maternal age 32.9 (0.33)
Immigrant 21.7
Living arrangement
 Single 62.4
 Cohabiting 6.6
 Married 31.1
Education level
 Less than a high school diploma 42.2
 High school diploma only 34.9
More than a high school diploma 22.9
Poverty level 0.75 (0.02)
Number children in household 2.7 (.05)
Focal child age
 Of younger sample (n = 1,212) 2.6 (0.07)
 Of older sample (n = 1,122) 12.5 (0.07)
Open in a new tab

Analytic Strategy

To address the first aim (i.e., testing informal support and burden), we examined support and burden along with the demographic characteristics among low-income mothers. We conducted Spearman correlations, appropriate for dichotomous variables (Cohen & Cohen, 1975), among support and burden. To classify patterns of support and burden among low-income mothers, we conducted an LCA. LCA suggests that an unobserved, mutually-exclusive grouping variable can be constructed from a set of categorical predictors and considers values on the range of indicators to create a latent class rather than considering each indicator separately (Collins & Lanza, 2010). Prior studies have used LCA to examine support access and transfers (e.g., Hogan et al., 1993; Kim et al., 2015; Lowe & Willis, 2015).

We conducted an LCA using a three-step approach (Vermunt, 2010). The first step builds a set of response variables. Our analysis included support and burden in each of four domains (i.e., emotional, practical, child care, and financial) totaling eight indicators. The cross-classification of the eight variables yielded 256 different cells, or potential patterns of support and burden.

In the second step, each mother was assigned to a particular cell based on her posterior profile membership probabilities (Vermunt, 2010). In this step, LCA uses an iterative process to test k number of latent profiles against k-1 profiles to maximize model fit. Model fit was assessed by a variety of statistical criteria including Akaike Information Criterion (AIC), Bayesian Information, Criterion (BIC), Sample Size Adjusted-BIC (A-BIC), and Lo, Mendell, and Rubin likelihood ratio test (LMRT). Lower levels of AIC, BIC, and A-BIC; and a significant LMRT indicated better fit. We determined the final profiles from the fit indices, model parsimony, and model interpretation based on theoretical relationships of support and burden (Nylund, Asparouhov, & Muthén, 2007). We also reported class entropy that describes the overall classification of mothers into profiles, but we did not use entropy in model selection because it is not designed to do so (Vermunt & Magidson, 2002). Entropy values approaching 0.80, as ours and other examinations of social networks did (e.g., Lowe & Willis, 2015), are acceptable and indicate that mothers were appropriately classified minimizing the concern of misclassification (Clark & Muthén, 2009; Vermunt, 2010). To better understand the latent profile characteristics, we examined the percent of each latent profile that had support and had burden for each domain.

The third step of the process considers the association between specified predictors and latent profile membership (Vermunt, 2010). The multinomial regression analysis tested how demographic characteristics influenced the latent profile distribution and compared how mothers in less advantaged groups fared relative to mothers in the “Healthy support and burden” latent profile. We presented odds ratios in which values greater than one indicate the percentage increase of a particular latent profile relative to the reference category that is associated with a one-unit change in the covariate.

After conducting the cross-sectional LCA, we examined how Wave-1 latent profile related to change-in-profile at later time points. We compared mothers’ Wave-1 latent profile to their Wave 2 and Wave 3 latent profiles, 1.5 and 5 years later, respectively. We considered the most common pathways in which mothers with changing support shifted. We also employed a logistic regression model to consider characteristics associated with latent profile change at either Wave 2 or 3. In preliminary analyses, we considered distinguishing among those who gained healthy support and burden, those who lost it, those who lacked it at all points, and those who had it at all points. With few exceptions, results were remarkably similar to those of the logistic regression. Because the data collection points were rather arbitrary (e.g., 1.5 and 5 years later), variability predicts poor family outcomes (Hill et al., 2017), and the less stable multinomial results due to a small number of mothers in the continuously unhealthy profile, we present the logistic regression results.

We conducted descriptive statistics using Stata statistical software, version 13. We estimated LCA models using the Mplus statistical software, version 7 (Muthen & Muthen, 2012). Following Ing and Nylund-Gibson (2017), we employed Full Information Maximum Likelihood (FIML) estimation with robust standard errors which makes the assumption that item-level missing data are missing at random (Little & Rubin, 1990).

Findings

Description of Low-income Mothers

Table 1 provides a snapshot of the demographic characteristics of low-income mothers of young children and young adolescents. Most mothers were Black (41%) or Hispanic (54%) with an average age of approximately 33 years old. Overall, mothers showed a high level of disadvantage. Over 75% had a high school diploma or less, and, on average, mothers lived at 75% of the poverty level. Regarding relationship status, 62% were single and 31% were married. Mothers had a range of one to nine children with a median of three children.

Table 2 provides Spearman correlations and the percentage of mothers with healthy levels of each type of support and burden. Correlations ranged from .08 to .43 and indicated that support and burden were positively and significantly correlated with one another at the p < 0.01 level. Mothers generally perceived support and burden. In terms of support, 87–88% of mothers perceived having at least someone for emotional support, favors, and child care while 79% of mothers had someone for financial support. In terms of burden, 66–76% of mothers had at least a couple of people calling on them without too much burden. When mothers perceived no or excess burden, the most common area was also for financial support (33%).

Table 2.

Spearman Correlations and Percentages for Support and Burden (N = 2,334)

Variables 1 2 3 4 5 6 7 8
1 Emotional support   _            
2 Support for favors 0.330   _
3 Support for child care 0.353 0.378   _
4 Support for loans 0.295 0.343 0.438   _
5 Healthy emotional burden 0.097 0.075 0.125 0.134   _
6 Healthy burden for favors 0.130 0.082 0.128 0.131 0.390   _
7 Healthy burden for child care 0.090 0.138 0.128 0.103 0.241 0.267   _
8 Healthy burden for loans 0.077 0.196 0.121 0.164 0.216 0.299 0.212   _
% 0.87 0.87 0.88 0.79 0.72 0.76 0.69 0.66
Open in a new tab

Note: With a Bonferroni correction, all bivariate correlations are significant at the p < 0.01 level.

Revealing Profiles of Support and Burden

To better understand informal support and burden of low-income mothers, we tested models with various numbers of latent profiles and selected the model with the best fit (see Table 3). Comparison of model fit statistics indicated that a four-profile model was superior to other options and adding additional profiles was unwarranted. To examine the interplay between support and burden across emotional, practical, child care, and financial domains, we first examined how support related to burden across domains. Two profiles had both healthy or both unhealthy levels of support and burden constituting 65% and 6% of the sample, respectively. The other profiles had unbalanced transactions with healthy support only (18%) or burden only (12%). As depicted in Figure 1, mothers tended to have similar levels of support and burden across emotional, practical, child care, and financial domains, with the exceptions that fewer mothers had financial support or burden.

Table 3.

Fit Indices for Latent Profile Models: Wave 1 (n = 2,334)

Model AIC BIC A-BIC Entropy LRT Profiles: n, %
2 Profile 17628.60 17726.44 17672.42 0.66 p < .001 1. n = 736, 31.53%
2. n = 1598, 68.47%
3 Profile 17163.33 17312.97 17230.36 0.75 p < .001 1. n = 345, 14.78%
2. n = 391, 16.75%
3. n = 1598, 68.47%
4 Profile 17049.12 17250.56 17139.36 0.74 p < .001 1. n = 142, 6.08%
2. n = 274, 11.74%
3. n = 407, 17.44%
4. n = 1511, 64.74%
5 Profile 17027.75 17280.98 17141.18 0.75 p = .17 1. n = 84, 3.60%
2. n = 156, 6.68%
3. n = 291, 12.47%
4. n = 386, 16.54%
5. n = 1417, 60.71%
6 Profile 17015.94 17320.96 17152.58 0.77 p = .06 1. n = 96, 4.11%
2. n = 162, 6.94%
3. n = 211, 9.04%
4. n = 233, 9.98%
5. n = 345, 14.78%
6. n = 1287, 55.14%
Open in a new tab

Note: AI =Akaike Information Criterion; BIC=Bayesian Information Criterion; A-BIC=Sample Size Adjusted BIC; Lo-Mendell-Rubin Adjusted LRT Likelihood Ratio Test=LRT.

Figure 1.

Figure 1.

Open in a new tab

Latent Classes of Informal Networks (Wave 1)

At one end of the profile continuum, the first profile, termed “unhealthy support and burden” contained the fewest mothers (6%). Mothers in this profile had unhealthy levels of support and burden. This group was the most disadvantaged group as measured by each domain measure (i.e., emotional, practical, child care, and financial). For support and burden in each domain, fewer than one half of mothers in this profile had healthy levels, significantly less than mothers in profiles with healthy support or burden (see Table 4). At the other end of profile continuum, in the fourth and largest profile termed “healthy support and burden”, mothers had the most advantaged networks such that 90% or more of mothers had healthy levels of support and 80% or more had healthy levels of burden across domains. The healthy support and burden profile stands out as the most advantaged profile with significantly healthier support and burden than all other profiles in six of the eight variables used in the LCA.

Table 4.

Chi-Square Tests Comparing Differences in Support and Burden Across Latent Classes (n = 2,334)

Unhealthy Support & Burden (6%) Healthy Burden Only (12%) Healthy Support Only (17%) Healthy Support & Burden (65%)
Emotional support abcdef 38.0 56.2 91.2 96.2
Support for favors bcdef 41.6 46.0 94.4 97.0
Support for child care abcde 20.4 46.4 100.0 99.1
Support for loans bcdef 20.4 24.1 83.3 93.1
Healthy emotional burden acdef 26.1 77.7 26.8 86.8
Healthy burden for favors acdf 9.9 95.3 13.8 95.6
Healthy burden for child care acdef 32.4 62.8 38.8 82.1
Healthy burden for loans abcdef 16.2 63.5 32.2 79.6
Open in a new tab
a

The Unhealthy Support & Burden and Healthy Burden profiles are significantly different.

b

The Unhealthy Support & Burden and Healthy Support profiles are significantly different.

c

The Unhealthy Support & Burden and Healthy Support & Burden profiles are significantly different.

d

The Healthy Burden and Healthy Support profiles are significantly different.

e

The Healthy Burden and Healthy Support and Burden profiles are significantly different.

f

The Healthy Support and Healthy Support and Burden profiles are significantly different.

The remaining profiles experienced unbalanced transactions such that one profile was high on burden and low on support (termed “healthy burden only”) and the other profile was low on burden and high on support (termed “healthy support only”). As depicted in Figure 1 and Table 4, while these two profiles experienced imbalance, their healthy and unhealthy levels were more moderate than the profiles with balanced transactions (i.e., unhealthy support and burden and healthy support and burden).

Demographic Characteristics Associated with Support and Burden Profiles

In the third step of the LCA, we examined how demographic characteristics predicted latent profile membership. The multinomial regression model was statistically significant. Overall, few demographic characteristics predicted latent profile. Specifically, in limited instances, poverty status and marital status significantly predicted profile membership. Married mothers and those in deeper poverty had lower odds of membership in the healthy burden only profile relative to the healthy support and burden profile than their single, and less impoverished counterparts, respectively. No tested demographic characteristic distinguished mothers in the unhealthy support and burden or support only profiles from the healthy support and burden profile. Race and ethnicity, nativity, age, education level, number of children in the household, and child age were not significantly related to latent profile.

Changes in Support and Burden

To examine latent class change, we conducted two additional, separate LCAs on Wave 2 and Wave 3 data and a logistic regression predicting change. Identical to the Wave 1 LCA, we tested models with various latent profiles and selected the model with the best fit. Appendices 1 and 2 provide the model fit criteria and profiles of each latent profile. In both analyses, fit statistics were not always congruent such that the BIC suggested a four-profile model while the LRT suggested a five-profile model. In light of fit indices, model parsimony, and model interpretation (Nylund et al., 2007), a four-profile option provided the best fit to the data. The profiles of each latent profile in the subsequent waves followed a similar pattern as the Wave 1 data resulting in our maintaining the same names for each profile. In addition, the profile proportions approximated each other regardless of wave.

Similar proportions in each profile at each wave hide an important finding: mothers with unhealthy support or burden lacked profile stability. Given similar findings between any two waves, Figure 2 illustrates movement between latent profiles from Waves 1 and 3. To interpret Figure 2, the four panels reflect the four latent profiles at Wave 1. For example, Panel A includes the 6% of mothers with unhealthy burden and support at Wave 1; at the other end of the continuum, Panel D includes the 66% of mothers with healthy support and burden at Wave 1. The extrapolated portion of the pie is the percentage of mothers who remained in their Wave 1 profile at Wave 3. In Panel A, only 24% of mothers with unhealthy support and burden remained in the profile at Wave 3 with relatively equal proportions joining healthy burden only (20%), healthy support only (24%) and healthy support and burden (32%) latent profiles. Similarly, only 16% and 32% of those with healthy burden only or healthy support only maintained these perceptions; the majority, 53% in both profiles, joined the healthy support and burden profile. Alternatively, Panel D indicates that 71% of mothers in the healthy support and burden profile at Wave 1 retained this health in Wave 3. Together, the panels of Figure 2 illustrate that mothers with unhealthy support or burden tended to enter a healthier profile at Wave 3 while mothers with healthy support and burden tended to retain it.

Figure 2.

Wave 3 Latent Class Distribution by Wave 1 Latent Class Distribution*

Figure 2.

*Exprapoled sections reveal the percentage of mothers of each latent class who remained in that class at Wave 3.

Open in a new tab

We next examined the individual characteristics that predicted at least one latent profile change after Wave 1. Logistic regression results in Table 6 confirm the greater movement among mothers without healthy support and burden levels. Similar to the limited predictors of latent profile, age was the only demographic characteristic predicting latent profile change. With each year of age, mothers had 3% lower odds of change.

Table 6.

Odds Ratios (and Confidence Intervals) of Logistic Regression Analysis Predicting Latent Class Change

Experienced a change at Wave 2 or Wave 3
Latent Class at Wave 1
   No healthy support or burden 35.89**
(11.62–110.86)
   Healthy burden only 20.01**
(5.66–70.74)
   Healthy support only 8.32**
(reference: healthy support and burden) (4.44–15.60)
Mother’s Race/Ethnicity
   Black 0.88
(0.41–1.93)
   Hispanic 0.96
(Reference=White) (0.44–2.10)
Immigrant (Yes) 1.06
(0.62–1.81)
Relationship status
    Cohabiting 1.14
(0.55–2.39)
    Married 0.76
    (Reference=single) (0.45–1.27)
Mother’s age 0.97*
(0.95–1.00)
Poverty level 1.11
(0.79–1.56)
Highest level of education
   HS diploma/GED only 1.12
(0.68–1.84)
   More than HS diploma 0.78
   (reference less than HS) (0.47–1.29)
Number of kids in household 1.09
(0.95–1.26)
Child age 1.03
(0.98–1.08)
Open in a new tab

Discussion

Research has demonstrated the importance of informal support perceptions for maternal and child well-being among low-income families (e.g., Huang et al., 2014; Taylor & Conger, 2017); however, little quantitative work has considered the potential for informal burden that may accompany support. Moreover, we do not know much about informal support over time. These gaps in knowledge are especially important considering that welfare spending has decreased post-welfare reform among the poorest families (Moffitt, 2015). With a diverse sample of mothers, many of whom were single and living in poverty, this study demonstrated distinguishable profiles of informal networks and changes in support and burden over three time points and seven years. Findings from this large, low-income sample of mothers yield two central contributions to the literature: (a) healthy support relies upon healthy burden, and (b) support and burden change over time, particularly among the most vulnerable low-income mothers.

Results supported four profiles-–healthy support and burden, unhealthy support and burden, healthy burden only, and healthy support only. Congruent with our hypothesis, among this largely disadvantaged group of mothers, the healthy support and burden group was the largest profile while the unhealthy support and burden group was the smallest. Although healthy informal networks as operationalized in this study do not mean that mothers had enough informal support to prevent scarcity or hardship, the large percentage of mothers with at least some support and burden is encouraging. Mothers and their children benefit from informal support availability across multiple domains, including health, maternal employment, and child cognitive development and behavior (Huang et al., 2014; Meadows, 2009; Ryan et al., 2009). Although most mothers perceived healthy informal networks, a sizable minority (20%−30%) did not. Moreover, mothers were particularly lacking in terms of financial support.

The LCAs of mothers at Waves 2 and 3 indicated that weak or unbalanced informal networks changed over time. Congruent with our hypothesis derived from exchange theory and Offer’s (2012) qualitative work, the first central contribution of this study is the finding that healthy networks were conditional upon burden. Healthy networks required not only available support, but, also, manageable network contributions. Mothers with unhealthy networks—either in support or burden—could not reliably turn to others for basic support. Among this study sample, approximately one-fourth of the mothers experienced support imbalance. Those with imbalance were in precarious situations vulnerable to change at a later study wave.

Building on the importance of immediate reciprocity, the second central contribution of this study is the finding that networks change over time, particularly among the most disadvantaged. Congruent with our hypothesis, results suggest that mothers without healthy support and burden experienced the bulk of network change. Despite 18 months between Wave 1 and Wave 2 and six years between Wave 1 and Wave 3, mothers’ movement between profiles appeared remarkably similar suggesting that frequent fluctuations occur among those without a healthy network. Gaining healthy support and burden was more typical than losing them, but switching profiles was common among those without a healthy balance.

Theory generated from ethnographic work can help to explain the importance of network balance and stability. Specifically, when describing the burden of reciprocity, Offer (2012) suggests that among the most impoverished, the pressure to reciprocate support can be a significant impediment to support. This can result in social fragmentation whereby members (a) are excluded from supports if network members perceive them as an undue burden upon already limited resources; or (b) withdraw from support by self-imposing restrictions to support because of the inability to reciprocate (Offer, 2012). Our findings support these ideas, at least in part, in that one-sided relationships tended to change over time. In addition, the movement among those with unhealthy support or burden reinforces earlier research that demonstrates the most disadvantaged have the least access to support (e.g., Harknett & Hartlett, 2011).

Counter to our hypothesis (i.e., specific vulnerabilities will exacerbate network functioning), most demographic characteristics were not related to informal network functioning. Although married mothers and those in deep poverty had lower odds of membership in the healthy burden only profile, no other covariates reached statistical significance. Although work examining the contribution of demographic characteristics to support profiles is limited, our findings are congruent with studies examining older adults (Hogan & Eggebeen, 1995) and low-income women following a natural disaster (Lowe & Willis, 2015). For example, among low-income women who survived Hurricane Katrina, the examined demographic characteristics (i.e., employment status, age, race/ethnicity) were not related to class profile. Rather than pre-determined networks, perhaps low-income mothers face situations (e.g., evictions, partner incarceration, child behavior problems) that inhibit the development and maintenance of healthy support networks (Gazso et al., 2016). Alternatively, demographic characteristics may not be influential in network health among low-income mothers. Future research can benefit from modeling longitudinal trajectories of informal network profiles to consider the role of demographics and life events for informal network health among diverse populations.

These findings are important because they indicate that we cannot assume that the most vulnerable families will be able to rely on informal networks for help. Moreover, we cannot assume that mothers with support will maintain that support over time. Whereas U.S. public safety net policies are designed to aid when families’ own formal and informal resources are insufficient, our study identifies a group of families who lack adequate informal networks. This is consistent with past research that demonstrates those who are most need of support are often those without it (Harknett & Hartlett, 2011; Radey & Brewster, 2013). In such cases, our U.S. public safety nets may not be meeting the intended needs of some families (Edin & Shaefer, 2015; Loprest, 2011). By identifying profiles of families and understanding the change and fluctuations in those profiles over time, we can begin to specifically target services to better meet the needs of the most vulnerable families. In targeting services, mothers may benefit from an enhanced public safety net with the capacity to bolster families’ abilities to both access and contribute to informal support networks. Doing so may mitigate the social fragmentation that seems to be occurring among the most disadvantaged.

Limitations and Future Directions

Despite the contributions of this study, there also are limitations. First, while this study tested both support and burden, we did not measure exchanges between people. Rather, we assessed mothers’ reports of their emotional, practical, child care, and financial supports, as well as their burden in those same domains. Second, we tested demographic characteristics known to be linked with support and support change; however, there are other key variables (e.g., maternal depression, low self-efficacy) that are also predictive of support (Harknett, 2006). Future research could test the extent to which other known variables predict support profile and profile change. Third, the study had limited and rather arbitrary data points. The finding that movement between profiles was similar 1.5 years and 5 years later suggests that these intervals may be too distant to capture all movement. Future work with additional time points provides an opportunity to examine the frequency of movement, including how long mothers who enter the healthy support and burden profile stay. Fourth, although strengths of this study include a diverse sample of families living in poverty, the measures of emotional, practical, child care, and financial support and burden are rather basic. Although past research has used similar variables (e.g., Hogan et al., 1993; Kim et al., 2015), measures that more comprehensively capture the depth and variations of support would further contribute to the literature. Furthermore, our focus on perceived support and burden provides little insight into received support and burden. Despite these limitations, our findings identify discrete profiles of support networks. The identification of a sizable minority of low-income mothers who cannot consistently rely on informal support is significant in light of diminished formal economic supports available to children and families. Future research should determine how circumstances (e.g., marriage, additional child birth, health issue) influence support profile trajectory and the effect of these profiles on family and child outcomes.

Supplementary Material

Supp AppendixS1

Table 5.

Odds Ratios (and Confidence Intervals) of Multinomial Regression Analysis Predicting Latent Class Membership

Unhealthy Support & Burden vs. Healthy Support & Burden Healthy Burden Only vs. Healthy Support & Burden Healthy Support Only vs. Healthy Support & Burden
Mother’s Race/Ethnicity
   Black 0.85 1.76 0.78
(0.39–1.89) (0.60–5.21) (0.47–1.29)
   Hispanic 1.40 2.87 0.65
(Reference=White) (0.63–3.08) (0.99–8.32) (0.38–1.12)
Immigrant (Yes) 0.51 0.98 1.19
(0.25–1.01) (0.60–1.57) (0.78–1.82)
Relationship status
    Cohabiting 0.80 0.68 0.79
(0.31–2.07) (0.30–1.52) (0.43–1.48)
    Married 0.60 0.49* 0.68
    (Reference=single) (0.29–1.26) (0.25–0.98) (0.43–1.08)
Mother’s age 1.01 0.99 0.99
(0.99–1.03) (0.97–1.01) (0.97–1.01)
Poverty level 1.15 0.64* 1.02
(0.78–1.68) (0.45–0.91) (0.76–1.37)
Highest level of education
   HS diploma/GED only 0.78 0.76 0.79
(0.48–1.28) (0.49–1.18) (0.56–1.12)
   More than HS diploma 0.70 0.66 0.80
   (reference less than HS) (0.38–1.28) (0.40–1.08) (0.55–1.19)
Number of kids in household 1.13 1.03 1.05
(0.99–1.30) (0.89–1.18) (0.94–1.18)
Child age 1.02 1.01 1.03
(0.97–1.07) (0.97–1.05) (0.99–1.07)
Open in a new tab

Acknowledgements:

The research was supported by National Institute of Health (NIH), National Institute of Child Health and Development (NICHD) 5R03HD092706–02. The authors also are grateful to the helpful comments of the Ming Cui, the JMF editor, and the anonymous reviewers.

References

  1. Angel R, Burton L, Chase-Lansdale PL, Cherlin A, Moffitt R, & Wilson WJ (2009). Welfare, children, and families: A three-city study Inter-university Consortium for Political and Social Research. [Google Scholar]
  2. Angel R, Burton L, Chase-Lansdale PL, Cherlin A, & Moffitt R (n.d.). Welfare, children, and families: A three-city study, Wave 3 Users Guide Ann Arbor, MI: Inter-University Consortium for Political and Social Research. [Google Scholar]
  3. Belle D (1983). The impact of poverty on social networks and supports. Marriage and Family Review, 5, 89–103. doi: 10.1300/J002v05n04_06 [DOI] [Google Scholar]
  4. Blau P (1964). Exchange and power in social life Transaction Publishers: New Brunswick, NJ. [Google Scholar]
  5. Bourdieu P (1986). The forms of capital. In Richardson JG (Ed.), Handbook of theory and research for the sociology of education (pp. 241–258). New York: Greenwood. [Google Scholar]
  6. Brewster KL, & Padavic I (2002). No more kin care? Change in Black mothers’ reliance on relatives for child care, 1977–94. Gender & Society, 16, 546–563. doi: 10.1177/0891243202016004008 [DOI] [Google Scholar]
  7. Brownell A, & Shumaker SA (1984). Social support: An introduction to a complex phenomenon. Journal of Social Issues, 40(4), 1–9. doi: 10.1111/j.1540-4560.1984.tb01104 [DOI] [Google Scholar]
  8. Clark S, & Muthén B (2009). Relating latent class analysis results to variables not included in the analysis Retrieved from https://www.statmodel.com/download/relatinglca.pdf [Google Scholar]
  9. Cohen J, & Cohen P (1975). Applied multiple regression/Correlation analysis for the behavioral sciences Hillsdale, NJ: Lawrence Erlbaum. [Google Scholar]
  10. Cohen S, Mermelstein R, Kamarck T, & Hoberman HM (1985). Measuring the functional components of social support. In Social support: Theory, research and applications (pp. 73–94). New York, NY: Springer. [Google Scholar]
  11. Collins LM, & Lanza ST (2010). Latent class and latent transition analysis: With applications in the social, behavioral, and health sciences New York: John Wiley & Sons. [Google Scholar]
  12. Domínguez S (2011). Getting ahead: Social mobility, public housing, and immigrant networks New York: NYU Press. [Google Scholar]
  13. Dunkel-Schetter C, & Bennett TL (1990). Differentiating the cognitive and behavioral aspects of social support. In Sarason BR, Sarason IG, & Pierce GR (Eds.), Wiley series on personality processes. Social support: An interactional view (pp. 267–296). Oxford, England: John Wiley & Sons. [Google Scholar]
  14. Edin K, & Kissane RJ (2010). Poverty and the American family: A decade in review. Journal of Marriage and the Family, 72, 460–479. doi: 10.1111/j.1741-3737.2010.00713 [DOI] [Google Scholar]
  15. Edin K, & Lein L (1997). Making ends meet: How single mothers survive welfare and low-wage work New York: Russel Sage Foundation. [Google Scholar]
  16. Edin K, & Shaefer HL (2015). $2.00 a day: Living on almost nothing in America New York: Houghton Mifflin Harcourt Publishing Company. [Google Scholar]
  17. Floyd I, Pavetti L, & Schott L (2017). TANF reaching few poor families Washington, DC: The Center on Budget and Policy Priorities. [Google Scholar]
  18. Fontenot K, Semega J, Kollar M (2018). Income and Poverty in the United States: 2017 Washington, DC: U.S: Census Bureau: Current Population Reports, P60–263, U.S. Government Printing Office. [Google Scholar]
  19. Gazso A, McDaniel S, & Waldron I (2016). Networks of social support to manage poverty: More changeable than durable. Journal of Poverty, 20, 441–463. doi: 10.1080/10875549.2015.1112869 [DOI] [Google Scholar]
  20. Harknett K (2006). The relationship between private safety nets and economic outcomes among single mothers. Journal of Marriage and Family, 68, 172–191. doi: 10.1111/j.1741-3737.2006.00250 [DOI] [Google Scholar]
  21. Harknett KS, & Hartnett CS (2011). Who lacks support and why? An examination of mothers’ personal safety nets. Journal of Marriage and Family, 73, 861–875. doi: 10.1111/j.1741-3737.2011.00852 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Hill HD, Romich J, Mattingly MJ, Shamsuddin S, & Wething H (2017). An introduction to household economic instability and social policy. Social Service Review, 91, 371–389. doi: 10.1086/694110 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Hogan DP, & Eggebeen DJ (1995). Sources of emergency help and routine assistance in old age. Social Forces, 73, 917–936. doi: 10.1093/sf/73.3.917 [DOI] [Google Scholar]
  24. Hogan DP, Eggebeen DJ, & Clogg CC (1993). The structure of intergenerational exchanges in American families. American Journal of Sociology, 98, 1428 – 1458. doi: 10.1086/230194 [DOI] [Google Scholar]
  25. Hogan DP, Hao L, & Parish W (1990). Race, kin networks, and assistance to mother-headed families. Social Forces, 68, 797–812. doi: 10.1093/sf/68.3.797 [DOI] [Google Scholar]
  26. Huang CY, Costeines J, Kaufman JS, & Ayala C (2014). Parenting stress, social support, and depression for ethnic minority adolescent mothers: Impact on child development. Journal of Child and Family Studies, 23, 255–262. doi: 10.1007/s10826-013-9834-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Ing M, & Nylund-Gibson K (2013). Linking early science and mathematics attitudes to long-term science, technology, engineering, and mathematics career attainment: Latent class analysis with proximal and distal outcomes. Educational Research and Evaluation, 19, 510–524. doi: 10.1080/13803611.2013.806218 [DOI] [Google Scholar]
  28. Institute of Medicine (IOM) (2013). U.S. health in international perspective: Shorter lives, poorer health, Report Brief Washington, DC: National Academy of Sciences. [Google Scholar]
  29. Kim K, Zarit SH, Fingerman KL, & Han G (2015). Intergenerational exchanges of middle‐aged adults with their parents and parents‐in‐law in Korea. Journal of Marriage and Family, 77, 791–805. doi: 10.1111/jomf.12185 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Levine J (2013). Ain’t no trust: How bosses, boyfriends, and bureaucrats fail low-income mothers and why it matters Oakland CA: University of California Press. [Google Scholar]
  31. Little RJA, & Rubin DB (1990). Statistical analysis with missing data New York: Wiley. [Google Scholar]
  32. Loprest P, & Nichols A (2011). Dynamics of being disconnected from work and TANF Washington, DC: The Urban Institute. [Google Scholar]
  33. Lowe SR, & Willis M (2015). Trajectories of perceived social support among low-income female survivors of Hurricane Katrina. Journal of Social and Personal Relationships, 32, 1034–1055. doi: 10.1177/0265407514558958 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Mazelis JM (2017). Surviving poverty: Creating sustainable ties among the poor New York, NY: NYU Press. [Google Scholar]
  35. Meadows SO (2009). Is it there when you need it? Mismatch in perception of future availability and subsequent receipt of instrumental social support. Journal of Family Issues, 30, 1070–1097. doi: 10.1177/0192513X09333759 [DOI] [Google Scholar]
  36. Moffitt RA (2015). The deserving poor, the family, and the U.S. welfare system. Demography, 52, 729–749. doi: 10.1007/s13524-015-0395-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Muthén LK, Muthén BO (2012). Mplus user’s guide (Version 7) Los Angeles, CA: Muthén & Muthén. [Google Scholar]
  38. Nelson MK (2005) The social economy of single motherhood: Raising children in rural America New York: Routledge. [Google Scholar]
  39. Nylund KL, Asparouhov T, & Muthén BO (2007). Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural Equation Modeling, 14, 535–569. doi: 10.1080/10705510701575396 [DOI] [Google Scholar]
  40. Offer S (2012). The burden of reciprocity: Processes of exclusion and withdrawal from personal networks among low-income families. Current Sociology, 60, 788–805. doi: 10.1177/0011392112454754 [DOI] [Google Scholar]
  41. Pilkauskas NV, Campbell C, & Wimer C (2017). Giving unto others: private financial transfers and hardship among families with children. Journal of Marriage and Family, 79, 705–722. doi: 10.1111/jomf.12392 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Portes A (1998). Social capital: Its origins and applications in modern sociology. Annual Review of Sociology, 24, 1–24. doi: 10.1146/annurev.soc.24.1.1 [DOI] [Google Scholar]
  43. Radey M (2015). The role of race/ethnicity and acculturation in the functioning of disadvantaged mothers’ social support networks. Family Relations, 64, 592–605. doi: 10.1111/fare.12167 [DOI] [Google Scholar]
  44. Radey M, & Brewster K (2013). Predictors of stability and change in private safety nets of unmarried mothers. Journal of Social Service Research, 39, 397–415. doi: 10.1080/01488376.2013.775090 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Ryan RM, Kalil A, & Leininger L (2009). Low‐income mothers’ private safety nets and children’s socioemotional well‐being. Journal of Marriage and Family, 71, 278–297. doi: 10.1111/j.1741-3737.2008.00599 [DOI] [Google Scholar]
  46. Sandstrom H, & Huerta S (2013). The negative effects of instability on child development: A research synthesis Washington, DC: The Urban Institute. [Google Scholar]
  47. Schenck-Fontaine A, Gassman-Pines A, & Hill Z (2017). Use of informal safety nets during the Supplemental Nutrition Assistance Program benefit cycle: How poor families cope with within-month economic instability. Social Service Review, 91, 456–487. doi: 10.1086/694091 [DOI] [Google Scholar]
  48. Stack C (1974). All our kin: Strategies for survival in a Black community New York: Harper & Row. [Google Scholar]
  49. Taylor J, & Turner RJ (2001). A longitudinal study of the role and significance of mattering to others for depressive symptoms. Journal of Health and Social Behavior, 42, 310–325. doi: 10.2307/3090217 [DOI] [PubMed] [Google Scholar]
  50. Taylor RJ, Mouzon DM, Nguyen AW, & Chatters LM (2016). Reciprocal family, friendship and church support networks of African Americans: Findings from the National Survey of American Life. Race and Social Problems, 8, 326–339. doi: 10.1007/s12552-016-9186-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Taylor ZE, & Conger RD (2017). Promoting strengths and resilience in single‐mother families. Child Development, 88, 350–358. doi: 10.1111/cdev.12741 [DOI] [PubMed] [Google Scholar]
  52. Turney K, & Kao G (2009). Pre-kindergarten child care and behavioral outcomes among children of immigrants. Early Childhood Research Quarterly, 24, 432–444. doi: 10.1016/j.ecresq.2009.07.007 [DOI] [Google Scholar]
  53. Vermunt JK (2010). Latent class modeling with covariates: Two improved three-step approaches. Political Analysis, 18, 450–459. doi: 10.1093/pan/mpq025 [DOI] [Google Scholar]
  54. Vermunt JK, & Magidson J (2002). Latent class cluster analysis. In Hagenaars JA & McCutcheon AL (Eds.), Applied latent class analysis (pp. 89–106). Cambridge, U.K.: Cambridge University Press. [Google Scholar]
  55. Wolf S, & Morrissey T (2017). Economic instability, food insecurity, and child health in the wake of the Great Recession. Social Service Review, 91, 534–570. doi: 10.1086/694111 [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp AppendixS1
NIHMS1019222-supplement-Supp_AppendixS1.docx (19.3KB, docx)

RESOURCES