Hispanic Couples in the Supporting Healthy Marriage Evaluation: How Representative are they of Low-Income Hispanic Couples in the United States?
The quality and dynamics of mother-father couple relationships shape the experiences and well-being of the entire family.1,2 A large body of research, for example, suggests that children thrive when their parents are in stable, high-quality, low-conflict relationships.3-6 Despite the size and growth of the Latinoa population in the United States, little is known about the relationship dynamics of Hispanic couples or how they might differ from other racial and ethnic groups. Few studies have examined how Hispanic couples interact and communicate with one another, how they resolve arguments or disagreements, or the relevance of these factors for their family’s well-being.
One reason that research on this topic is lacking is that data on these issues are scarce for the U.S. Hispanic population. A recent review of more than 20 (primarily national) data sets revealed that only one—the Supporting Healthy Marriage (SHM) evaluation data set—includes both a large enough sample of Latinos and information on an array of domains of family life. This data set therefore has the potential to address critical gaps in knowledge about Hispanic family life. Still, the SHM evaluation data set, while rich in information on married couplesb,c and their children, is a convenience sample of couples that may not represent the general population of Hispanic couples in the United States.
This brief assesses the extent to which Hispanic participants in the SHM evaluation data set represent the broader U.S. population of Hispanic couples. Specifically, we examine how representative the study’s Latino couples are of low-income Hispanic couples with children (under age 18) in the general population by comparing them on socio-demographic characteristics such as age, education, number and ages of children, earnings, and so forth. We also assess how Latino couples in SHM compare to the broader population of low-income Hispanic couples across multiple domains of marital quality and mental health. Through these analyses, we highlight the SHM dataset’s potential for generating knowledge about Hispanic families, and identify potential limitations.
This brief relies on data from the American Community Survey (ACS) and the Fragile Families and Child Wellbeing Study (Fragile Families). The ACS provides a national benchmark against which to assess the representativeness of SHM couples in terms of their demographic, employment, economic, and family characteristics, as well as their human capital. In turn, the Fragile Families data set contains a sample of families that are low-income and have young children. Unlike the ACS, Fragile Families collected data on couples’ relationship dynamics and well-being that can be used to compare Hispanic couples in the two data sets on those dimensions.
Compared with low-income Hispanic couples in the general population (ACS sample), Hispanic couples in the SHM evaluation data set:
- Were younger, had been married for fewer years, and had fewer and younger children.
- Were less likely to be immigrants. More than half of the Hispanic couples in SHM had at least one partner who was born in the United States.
- Had more human capital. Hispanic couples in SHM were more likely to report speaking English well and to have at least a high school education.
- Had higher earnings but lower labor force participation.
In addition, relative to low-income Hispanic couples in the Fragile Families study, Hispanic couples in SHM:
- Were less likely to express their love and affection often and to feel satisfied with the way they handled relationship problems; they were also more likely to argue frequently.
- Were more likely to be sad or have a substance abuse problem.
The differences across samples limit the ability to use the SHM data set to make generalizable statements about Hispanic family life in the United States. Nevertheless, the large sample size of Hispanics in SHM and the breadth of data collected (as well as the lack of current alternatives) allow for critical exploratory research on the well-being of Hispanic families. Our findings have several implications for programs and research:
- National data are needed on couples’ relationships and family dynamics.
- National data collection efforts should also include questions that can represent the diversity of Hispanic family life to deepen the field’s understanding of the varied experiences of Hispanic families in the United States (e.g., nativity, country of origin, educational attainment).
- Programs that target Hispanic couples like those who participated in the SHM evaluation should develop new strategies to enroll segments of the Hispanic population that were not well-represented in the SHM evaluation. In particular, additional efforts are needed to recruit couples in which both partners are foreign-born, who have limited English proficiency, and who have low levels of education.
Data and Methods
Description of data sets
The Supporting Healthy Marriage (SHM) evaluation contains data from a federally funded experimental evaluation of SHM programs implemented in eight sites from 2007 to 2009.7,8
Participants were recruited from different sources, including through community partners and referrals. To be eligible to participate in the study, couples had to be married and have a child under age 18. Most couples in the SHM evaluation study were low-income (approximately 82 percent of all couples had incomes below 200 percent of the federal poverty line); however, low-income family status was not a program or study requirement. Participants were randomly assigned to either an SHM experimental program or a control group. Data from male and female partners—in both the experimental and control groups—were collected at baseline (prior to the intervention) and at two later times: 12 and 30 months post-intervention. Analyses for this brief were conducted using baseline datad and restricted to couples in which at least one partner identified as Hispanic (N=3,402 couples).
The American Community Survey (ACS) is an annual survey conducted by the U.S. Census Bureau that is nationally representative of U.S. households.e The survey collects detailed demographic information from one household respondent, including age, gender, place of birth, and employment status. The responding adult answers these questions for all household members, including their partner (if they have one). This brief used data from the ACS three-year estimates for the same years during which the SHM study was conducted (2007–2009).f We limited our analyses to respondents who were married or living with their partner, who identified as Hispanic (or who indicated that their partner was Hispanic), and who had a child under age 18 living in the household (N=59,787 couples).
The Fragile Families and Child Wellbeing Study (Fragile Families) is a federally funded longitudinal study of the families of nearly 5,000 children born in large cities in the United States from 1998 to 2000.9 Parents were recruited from hospitals following childbirth. Interviews were conducted with both mothers and fathers of focal children. The study oversampled unmarried parents, whom they refer to as “fragile families.” When the data are weighted, the sample is representative of births in urban U.S. populations. The characteristics examined in this brief were assessed at the time of the focal child’s birth, with a few exceptions noted in Table 1. The total number of Hispanic couples included in our analyses was 474.
To make the ACS and Fragile Families samples comparable to the target population of the SHM evaluation study, we restricted the samples in each of the comparison data sets to couplesg with the following characteristics:
- At least one partner identified as Hispanic
- Married or cohabitingh
- Both partners are age 18 or older
- Had at least one child under age 18i
- Family income under 200 percent of federal poverty linej,k
To see how low-income Hispanic couples in SHM compare to those in the ACS and Fragile Families studies, we compared the distribution of a range of demographic and relationship characteristics across the three data sets. Table 1 defines the characteristics used to compare the samples and indicates whether similar or comparable measures were available in ACS and Fragile Families. To the extent possible, we defined variables at the couple level and used variables that were as equivalent as possible across data sets.
Sample weights were applied to the ACS and Fragile Families data sets. We present means and proportions for each variable across the three data sets when available. However, we based our comparisons of demographic characteristics of couples in SHM with the general population on those that compared SHM couples to couples in ACS (but not Fragile Families), as ACS is a nationally representative sample of U.S. households. Demographic information for the Fragile Families set is presented to provide context for the comparisons made. However, because these data are representative of a more targeted population (births in large cities from 1998 to 2000), they should not be used for comparisons of demographic characteristics of Hispanic couples in SHM. Instead, Fragile Families contains information about relationship quality and family functioning that allowed us to compare the data set to SHM. We used t-tests and chi-square tests to test for significant differences across data sets. Descriptive statistics were generated using Stata.
Age and race/ethnicity: Hispanic couples in SHM differed from Hispanic couples in the general population (ACS) on several demographic characteristics (Table 2).
Specifically, in terms of age and race/ethnicity, couples in SHM were:
- Younger and closer in age to each other. Hispanic men in SHM were, on average, five years younger (32 vs. 37 years old) than Hispanic men in the general population; Hispanic women in SHM were four years younger (30 vs. 34 years old) than Hispanic women in the general population. In addition, the difference in age between men and women was smaller among SHM couples than in the general population of Hispanic couples (1.6 vs. 4.4 years).
- More likely to be in a mixed-race relationship. Twenty-one percent of couples in the SHM sample were in a mixed- race relationship, compared to 14 percent in the national sample.
Nativity status: There were fewer immigrant couples in SHM than in the general population. Additionally, SHM couples were:
- More likely to have both partners born in the United States (31 percent), relative to Hispanic couples in the general population (19 percent). The SHM sample was less likely to include couples in which both partners were foreign-born (46 percent), relative to Hispanic couples in the general population (66 percent).
- More likely to be recent immigrants if they had been born outside the United States. A larger proportion of Hispanic men and women in SHM had been in the United States for five years or fewer (18 and 26 percent, respectively) compared to those in the general population (10 and 15 percent, respectively).
Language: In both samples (SHM and ACS), most couples included at least one partner who spoke Spanish. However, a lower percentage of couples in the SHM sample had both partners report speaking Spanish, relative to the national sample (72 percent compared to 80 percent). In general, compared with the general population, SHM couples were:
- More likely to have only one partner who spoke Spanish and to both report being proficient in English. Twenty-two percent of all couples in SHM had only one partner who spoke Spanish, relative to eight percent of all Hispanic couples in the United States. In contrast, a larger percentage of SHM couples had both partners report speaking English well (55 percent) than among the general population of Hispanic couples (44 percent).l
Education: Couples in the SHM sample had higher levels of education compared with Hispanic couples in the general population. Specifically:
- More than half of the couples in the SHM sample (55 percent) had a high school diploma, compared to 35 percent of Hispanic couples in the national sample.
Employment and earnings: Couples in the SHM sample earned, on average, over $3,000 more than Hispanic couples in the general population ($26,932 vs. $23,515, respectively);m however, their participation in the labor force was lower.
- A smaller proportion of Hispanic couples in the SHM sample had at least one employed partner (84 percent), relative to the national sample (92 percent). This difference appears to be driven by lower labor force participation among men in the SHM sample, relative to Hispanic men in the general population. The proportion of couples in which both partners were employed was similar in the two samples (31 percent in SHM and 32 percent in ACS).n However, the proportion of SHM couples in which both partners were unemployed was 16 percent, compared to just eight percent of Hispanic couples in the general population.
Children: Hispanic couples in SHM lived with fewer, and younger, children than Hispanic couples in the general population.
- Nearly one third (31 percent) of the couples in the SHM sample lived with one child or fewer, compared to 23 percent in the national sample. Conversely, a larger proportion of couples in the general population lived with two, three, or four children. There were no differences between the samples in the proportion of families living with five or more children in the household.
- On average, SHM couples had more children under age 5 and fewer children age 5 to 17, relative to Hispanic couples in the general population.
Relationship length: Hispanic couples in the SHM sample had been married for fewer years than Hispanics in the general population.
- The average couple in the SHM sample had been married for six years, compared to nearly 12 years in the national sample.
Relationship quality: Compared to Hispanic couples in the Fragile Families sample, Hispanic couples in the SHM sample had more strained relationships.o Specifically, SHM couples were:
- Less likely to express love and affection often. Over half (52 percent) of the couples in the SHM sample reported expressing love and affection toward one another often, compared to 71 percent of the Hispanic couples in Fragile Families.
- More likely to argue, not handle problems well, and report physical violence. Two thirds (68 percent) of the couples in the SHM sample had at least one partner who reported that the couple argued often, compared with 37 percent of Hispanic couples in Fragile Families. A smaller percentage of couples in the SHM reported handling their problems well than those in Fragile Families (29 vs. 37 percent). In addition, reports of physical violence were three times higher in SHM than in Fragile Families (26 vs. 8 percent, respectively).
Feelings of sadness and substance use: Hispanic couples in SHM were more likely to feel distressed/sad and to have a substance use problem. Specifically, couples in SHM were:
- More than twice as likely to report feeling sad or distressed than couples in Fragile Families (21 vs. 9 percent, respectively).
- Nearly six times as likely to report a substance abuse problem than couples in Fragile Families (36 vs. 6 percent, respectively).
The Supporting Healthy Marriage evaluation data set contains a wealth of information about couples’ relationship dynamics—more than any other large data set with a high proportion of Hispanics. The dataset can help address important gaps in knowledge about how Hispanic couples communicate and how they resolve disagreements, as well as provide insights into the overall quality of their relationships.
However, Hispanic couples in SHM differ in some important ways from Hispanic couples with low incomes in the general population. Hispanic couples in the SHM sample were somewhat better off than those in the general population—they were better educated, had higher earnings, and were more likely to speak English well. Partners in SHM study couples were also more likely to both be U.S.-born. On average, Hispanics with these characteristics tend to have more resources, be more connected to social networks, and have greater access to social institutions (such as the community organizations from which many couples from this study were recruited).10,11
Couples in the SHM were also younger, had been married for fewer years, and had fewer (and younger) children in their households. In short, SHM couples tended to be in the early stages of the family life cycle. Self-selection into SHM programs among younger couples—who, on average, had been married for six years—is not surprising. During the early years of marriage, many couples struggle to settle into their roles as spouses/partners and parents and to assume new financial responsibilities.12 This makes the early years a likely time for couples to seek support from relationship programs such as those evaluated through the SHM project.13,14 Indeed, SHM couples reported poorer relationship characteristics than those in the Fragile Families sample. These differences, while expected given that SHM couples were seeking programs to strengthen their relationships, must be kept in mind when interpreting findings derived from SHM data.
The differences identified here are consistent with those found in analyses comparing the entire SHM sample (not just Hispanics) to the general U.S. population.15 Although neither the SHM programs nor the evaluation study were designed to be representative of the general population, future efforts to enroll Hispanic participants might need to develop new strategies to engage specific segments of the Hispanic population. Specifically, recruitment should target couples in which both partners were born outside of the United States, who do not speak English well, and who have lower levels of education.
Together, our findings suggest that the SHM evaluation data set can fill in some gaps in knowledge about Hispanic family life. Our companion brief, How Well Do National Surveys Measure Hispanic Family Life?, documents the wealth of information on family life collected through the SHM evaluation. The data offer a unique opportunity to examine the interrelationship between couples’ characteristics, their relationship dynamics, and children’s outcomes among Hispanics. However, caution is needed when using the data set to develop profiles of relationship dynamics and family functioning in a “typical” low-income Hispanic family in the United States. Our findings highlight the need for nationally representative data with information about couples’ relationship and family dynamics, as well as information that can better represent Hispanic diversity. In turn, these data would contribute to a better understanding of the varied experiences and changing characteristics of Hispanic families in the United States.16
About the Supporting Healthy Marriage Evaluation Data Set
The SHM data set includes data from a recent experimental evaluation of the effectiveness of federally funded healthy marriage and relationship education programs targeting low-income married couples. The data set contains extensive information on various dimensions of couples’ marital quality and child outcomes, as well as key social, demographic, and economic measures. Notably, roughly 40 percent of the couples in the SHM sample had at least one partner who identified as Hispanic.
The ACS is a nationally representative sample of households and only captures residential unions. Thus, comparisons to couples in ACS exclude couples that are separated and couples that have children who do not live in the household. Given the lack of more recent data, we relied on the Fragile Families data set to assess how SHM couples fared in terms of their relationship quality and mental health compared to other Hispanic couples. Fragile Families is weighted to represent births occurring in large U.S. cities from 1998 to 2000, and is therefore not representative of the entire U.S. population. Comparisons to couples in Fragile Families do not capture the characteristics of Hispanic couples in smaller cities, rural areas, or those who have not had a child recently. Additionally, the Fragile Families sample was assembled nearly 10 years earlier than SHM, and does not capture the characteristics of Hispanic families that migrated to the United States in more recent years. Notwithstanding, the Fragile Families was the best data set available to assess how Hispanic couples in the SHM study compared to other couples in the general population in terms of marital quality and family functioning. his brief extends other efforts by the National Research Center on Hispanic Children and Families to assess and inventory the capacity of our country’s data infrastructure to measure and describe the characteristics and experiences of Hispanics in the United States.
aWe use “Hispanic” and “Latino” interchangeably throughout the brief. Consistent with the U.S. Census Bureau definition, this includes individuals having origins in Mexico, Puerto Rico, and Cuba, as well as other “Hispanic, Latino or Spanish” origins.
bPrograms targeted married couples but varied in the procedures used to verify marital status. Some programs asked couples whether they considered themselves married, while others placed more emphasis on legal marriage. At the 12-month follow-up, 81 percent of all couples in the SHM evaluation reported being legally married.
cFor simplicity, we refer to these couples as “married couples” from this point forward.
dLanguage spoken at home was assessed at the 12-month follow up.
eData are from the Integrated Public Use Microdata Series.
fWe used the 2008 ACS data set for “number of years married,” which was unavailable in the 2007 and 2009 ACS data sets.
gWe focused on heterosexual couples because the SHM programs did not include same-sex couples.
hWe included cohabiting couples to represent the full spectrum of families included in the SHM evaluation.
iTwo programs served couples who were pregnant or had recently had a child; the other six programs served couples with children under age 18.
jIncome-to-poverty calculations in the Fragile Families study were based on household income because family income was not available.
kWe conducted sensitivity analyses that restricted the SHM sample to those who met this income criterion. The results were generally the same as those conducted with the entire sample, with a few exceptions noted in the text.
lTo be eligible for SHM, both partners had to be able to speak either English or Spanish well.
mIn the sensitivity analyses that restricted the SHM sample to those with incomes 200 percent below the poverty line, the average combined earnings for SHM couples was $23,867, which is not significantly different from couples in the general population.
nIn the sensitivity analyses that restricted the SHM sample to those with incomes 200 percent below the poverty line, the proportion of couples in which both partners were employed was 27 percent, which is significantly lower than the proportion in the ACS.
oAdditional analyses restricted the SHM sample to couples who had a child age one or younger, to make the SHM sample more comparable to Fragile Families. The conclusions from those analyses were the same as those reported for the entire SHM Hispanic sample.
About this Brief
This brief extends other efforts by the National Research Center on Hispanic Children and Families to assess and inventory the capacity of our country’s data infrastructure to measure and describe the characteristics and experiences of Hispanics in the United States. Other efforts in this series include:
Improving Data Infrastructure to Recognize Hispanic Diversity in the United States:This brief and associated data tool identify which key data elements needed to capture the diversity (e.g., country of heritage, time in the United States, legal status, generational status, etc.) of the Hispanic population are available in nationally representative, large-scale data sets used to examine key topics related to child and family well-being.
Using Existing Large-Scale Data to Study Early Care and Education among Hispanics:A series of briefs and respective interactive tools inventories and critically assesses the availability of data elements related to early care and education search, access, decision-making, and utilization that have been measured in large-scale data sets with sizeable Latino samples.
This brief is a companion to the Hispanic Family Life Data Scan, which reviews surveys from more than 20 recent national data sets to assess the extent to which they collect information about Hispanic family/household composition, family formation and stability, relationship dynamics, and parenting and co-parenting.
1Amato, P. R. (2000). The consequences of divorce for adults and children. Journal of Marriage and the Family, 62(4), 1269-1287.
2Peterson, J., & Zill, N. (1986). Marital disruption, parent-child relationships, and behavior problems in children. Journal of Marriage and the Family, 62(4), 1269-1287.
3Brown, S. L. (2010). Marriage and child well-being: Research and policy perspectives. Journal of Marriage and Family, 72(5), 1059-1077.
4Thomson, E., & McLanahan, S. S. (2012). Reflections on “Family structure and child well-being: Economic resources vs. parental socialization”. Social Forces, 91(1), 45-53.
5Moore, K. A., Kinghorn, A., & Bandy, T. (2011). Parental relationship quality and child outcomes across subgroups. Research Brief. Washington, DC: Child Trends.
6Amato, P. (2005). The impact of family formation change on the cognitive, social, and emotional well-being of the next generation. Future of Children, 15(2), 75-96.
7Hsueh, J., Alderson, D. P., Lundquist, E., Michalopoulos, C., Gubits, D., Fein, D., et al. (2012). The Supporting Healthy Marriage evaluation: Early impacts on low-income families. Washington, DC: Office of Planning, Research, Evaluation, Administration for Children and Families, U.S. Department of Health and Human Services. Retrieved from https://www.acf.hhs.gov/sites/default/files/documents/opre/early_impacts_low.pdf
8Lundquist, E., Hsueh, J., Lowenstein, A. E., Fausceta, K., Gubits, D., Michalopolous, C., et al. (2014). A family-strengthening program for low-income families: Final impacts from the Supporting Healthy Marriage evaluation. Washington, DC: Office of Planning, Research and Evaluation, Administration for Children and Families, and U.S. Department of Health and Human Services. Retrieved from https://www.acf.hhs.gov/sites/default/files/documents/opre/shm2013_30_month_impact_reportrev2.pdf
9Reichman, N., E., Teitler, J., O., Grafinkel, I., & McLanahan, S., S. (2001). Fragile Families: Sample and design. Children and Youth Services Review, Vol. 23(Nos.4/5), 303-326.
10Viruell-Fuentes, E. A., & Schulz, A. J. (2009). Toward a dynamic conceptualization of social ties and context: Implications for understanding immigrant and Latino health. American Journal of Public Health, 99(12), 2167-2175.
11Mulvaney-Day, N. E., Alegria, M., & Sribney, W. (2007). Social cohesion, social support, and health among Latinos in the United States. Social Science & Medicine, 64(2), 477-495.
12Risch, G. S., Riley, L. A., & Lawler, M. G. (2003). Problematic issues in the early years of marriage: Content for premarital education. Journal of Psychology and Theology, 31(3), 253-269.
13Cherlin, A. (2009). Marriage, divorce, remarriage. Harvard University Press.
14Gottman, J. M., & Levenson, R. W. (2000). The timing of divorce: Predicting when a couple will divorce over a 14‐year period. Journal of Marriage and Family, 62(3), 737-745.
15Miller, G. J., Grbitis, D., Principe, A. D., & Knox, V. (2012). The Supporting Healthy Marriage evaluation: Final implementation findings. Washington, DC: Office of Planning, Research and Evaluation, Administration for Children and Families, U.S. Department of Health and Human Services. Retrieved from https://www.mdrc.org/sites/default/files/shm%20impl%20full%20pdf_1.pdf
16U.S. Department of Health and Human Services. (2014). Survey data elements to unpack diversity of Hispanic populations. Washington, DC: Office of Planning, Research and Evaluation, Administration for Children and Families, U.S. Department of Health and Human Services. Retrieved from https://www.acf.hhs.gov/sites/default/files/documents/opre/brief_survey_data_to_unpack_hispanic_final_03_27_2014.pdf