Love, Health, and the ‘Hood: An Examination of Romantic Relationship Adjustment and Perceived Neighborhood Quality as Predictors of Partnered Black Americans’ Long-Term Psychological Health

Abstract

Existing disparities regarding Black Americans’ psychological health warrant further investigation of socioecological factors that may be associated with negative and positive dimensions of psychological health in this population. Romantic relationship functioning and neighborhood context are two domains relevant to Black Americans’ mental health. However, less is known about how they may serve as independent and interactive prospective predictors of Black Americans’ psychological health and potentially in distinctive ways for Black men and women. Using data from 333 partnered Black Americans who participated in the Midlife in the United States study (MIDUS), we investigated relationship adjustment and neighborhood quality as independent and interactive predictors of negative and positive affect 10 years later and examined gender differences in these linkages. Higher neighborhood quality predicted lower levels of negative affect and higher levels of positive affect for both men and women a decade later. Additionally, for Black men, the longitudinal association between relationship adjustment and negative affect differed by neighborhood quality such that better relationship adjustment predicted higher subsequent negative affect only for men in lower quality neighborhoods. Findings demonstrate the connections among romantic relationship functioning, ecological resources, and gender in this population and highlight the importance of incorporating socioecological and intersectional perspectives for predicting Black Americans' long-term psychological health.

General Scientific Summary:

This study suggests that living in higher quality neighborhoods is related to greater positive aspects and lower negative aspects of psychological health over time for Black Americans. Moreover, for Black men specifically, better romantic relationship functioning is connected to poorer psychological health when they reside in lower quality neighborhoods.

Due to the historical and institutional legacies of racism in the United States, Black Americans are vulnerable to and experience disparities in psychological-emotional health (Barnes & Bates, 2017; Carter et al., 2021). To mitigate these disparities, it is important to utilize inclusive, multi-dimensional conceptualizations of psychological health and to employ an intersectional lens in the investigation of interpersonal- and community-level factors that may affect Black Americans’ psychological health over time (Alvidrez et al., 2019). In keeping with this perspective, the current study was designed to investigate romantic relationship adjustment and neighborhood quality as prospective independent and interactive predictors of psychological health for partnered Black Americans and to examine how these associations may differ for men and women.

Psychological Health in Black Americans

Compared to adults from other racial groups, Black adults are more likely to experience frequent symptoms of emotional distress, including feelings of sadness, hopelessness, and worthlessness (CDC, 2019; Mental Health America, 2022). Recent reports show that 4.8 million Black Americans contend with serious mental illness, with these rates increasing over the last decade (Mental Health America, 2022; SAMHSA, 2019). At the same time, Black Americans have similar or lower prevalence rates of most mental disorders compared to White Americans (Erving et al., 2019), suggesting that clinical measures of psychological disorders may not be sufficient to fully characterize Black Americans’ psychological functioning.

Examination of broader dimensions of psychological functioning that are less stigmatizing and common to multiple mental health disorders may afford a more holistic understanding of psychological-emotional health among Black Americans and aid in redressing disparities. The Research Domain Criteria (RDoC) framework proposed by the National Institute of Mental Health (Sanislow et al., 2010) provides a useful framework for such investigations by enhancing a dimensional understanding of mental health that extends beyond the typical categorical boundaries of psychopathology (Carter et al., 2021; Sanislow et al., 2010). Negative and positive affect are two constructs that fall within domains highlighted in RDoC (i.e., negative and positive valence systems) and may be relevant for further investigation, as disturbances in affect are common to a range of psychopathologies (e.g., depression, anxiety, borderline personality disorder, posttraumatic stress disorder; Carcone & Ruocco, 2017; Gore & Widiger, 2018; Sanislow et al., 2010). Better understanding of these affective domains may have important transdiagnostic implications for Black Americans and aid in advancing efforts to redress disparities in psychological distress. Further, attention to both negative and positive affect subverts the historical tendency to pathologize Black individuals and families by only attending to negative dimensions of psychological health. The current study was designed to extend the literature on Black Americans’ psychological health by investigating predictors of both negative and positive affect among Black Americans over 10 years.

Socioecological Determinants of Black Americans’ Psychological Health

Attempts to understand Black Americans’ psychological (and physical) health require situating individuals’ health within the interpersonal and institutional structures that sustain disparities in these areas and contribute to poorer outcomes throughout the life course (Alvidrez et al., 2019). This involves illuminating salient social and ecological factors that influence functioning in this population to leave individuals more at risk for (or more resilient to) poorer psychological health and that reveal potential points for intervention. Moreover, socioecological approaches to this work benefit from applying an intersectional framework (Crenshaw, 1989) to uncover nuances in the ways systems of oppression converge to position individuals who hold multiple marginalized identities (e.g., Black women) as most vulnerable to poorer health (del Río-González et al., 2021). Thus, the current study investigates romantic relationship functioning and neighborhood quality (interpersonal- and community-level factors in domains Black Americans experience disparities; Raley et al., 2015; Williams & Collins, 2001) as predictors of Black Americans’ long-term psychological health and examines gender differences in these associations.

Romantic Relationships

A growing literature on Black adults’ relationship functioning and psychological health demonstrates the beneficial effects of better relationship quality for long-term psychological health (Barton et al., 2022; King et al., 2022). The strength and strain model of marriage and health (Slatcher & Selcuk, 2017) posits that both relationship strengths and strains impact psychological health, including negative and positive affect, as well as cognitive appraisals of stress, perceptions of partner/relationship functioning, and other health behaviors. Understanding of how relationship functioning is associated with positive versus negative dimensions of psychological health is relatively limited, with much of the extant research overlooking the multi-dimensional nature of psychological health and relying primarily on measures of depressive symptoms as indicators of well-being (e.g., King et al., 2021). Disentangling these linkages will provide a more complete picture of these associations and inform efforts to decrease psychological distress and enhance psychological flourishing for Black adults.

The sociocultural family stress model (McNeil Smith & Landor, 2018) advocates for Black intimate relationships to be considered in light of the intertwined oppressive systems that Black families face, as family processes can transpire in specific ways for Black women and men. Societal racial inequality can amplify gender inequalities within romantic relationships, with Black couples describing how experiences of racism can precipitate men’s turning to their intimate relationships to reclaim a sense of power but also pressure women to minimize their power or intelligence in relationships (Cowdery et al., 2009). Black women are subject to cultural stereotypes that encourage prioritizing the needs of their loved ones, displaying perpetual strength and resilience, and denying psychological distress (i.e., Black superwoman schema; Abrams et al., 2019; Woods-Giscombé, 2010). As a result, Black women may be disadvantaged relative to men in capitalizing on the psychological benefits of better relationship adjustment, particularly in the context of socioecological stress (Jenkins et al., 2020).

Neighborhood Context

Neighborhood quality is another salient socioecological factor for Black Americans’ psychological health, as residential context and segregation in the housing domain are primary mechanisms perpetuating health disparities (Williams et al., 2019). The implementation of mortgage redlining, predatory lending, discriminatory housing policies, and other segregationist practices have created the historical and current landscape of the residential separation of Black Americans (Williams & Collins, 2001). Previous research has shown that neighborhood disadvantage is related to clinical depression for Black Americans (particularly those who are lower-income; Hastings & Snowden, 2019) and that exposure to neighborhood stressors is associated with poorer long-term mental health outcomes for Black adults, even when they move to different neighborhoods (Russell et al., 2018). Residential areas that afford community resources may therefore be important for enhancing individual psychological health for this population. According to social disorganization theory (Kubrin & Weitzer, 2003) and theories on neighborhood disorder (Ross & Mirowsky, 1999), the social cohesion and control, physical order and repair, and safety of a community contribute to perceptions of neighborhood quality and residents’ well-being. Research demonstrates that neighborhood quality is an ecological resource that can be leveraged for Black Americans’ psychological health, with some work demonstrating its relevance for both negative and positive aspects of well-being (depressive symptoms, life satisfaction; Erving & Hills, 2019).

Gender may play a salient role in the associations between neighborhood quality and psychological health. Although some research within the general population (Mair et al., 2010a) and with Black Americans specifically (Cutrona et al., 2005; Lei & Simons, 2021) has highlighted the salience of neighborhood factors for women’s mental health, accumulating evidence supports Black men’s particular psychological sensitivity to their neighborhood context (Hale et al., 2019; Mair et al., 2010b). For instance, studies have found that fears of neighborhood violence predicted increases in depressive symptoms for Black emerging adult men but not for women (Assari et al., 2015) and that perceived neighborhood crime was associated with psychological distress for older Black men but not Black women (Erving, 2022). Because Black men face cultural perceptions related to criminal behavior and societal pressure to live up to masculine ideals of toughness and fearlessness, it may be more difficult or distressing for them to maintain psychological health in the context of neighborhood disorder or violence (Assari et al., 2015; Erving, 2022).

Interactive Effects of Romantic Relationships and Neighborhood Quality

Neighborhood quality is also an important environmental context for Black romantic relationships (Cutrona et al., 2003), which may interact with relationship adjustment to impact psychological health. Black adults living in urban neighborhoods have discussed how neighborhood violence impacted their romantic relationships and mental health, as distressed environments contributed to a lack of emotional vulnerability with romantic partners and reinforced feelings of worry and anxiety (Hollie & Coolhart, 2020). However, because neighborhoods are typically conceptualized as a stressor for Black Americans, relatively less is known about how one’s neighborhood can serve as a resource that interacts synergistically with romantic relationship functioning to potentiate Black Americans’ psychological health. Via provisions of greater personal safety/security and additional social capital/support, which enrich both relationships and individual well-being (Bryant & Wickrama, 2005), higher neighborhood quality could attenuate the links between poorer relationship functioning and negative dimensions of psychological health or strengthen the linkages between higher relationship functioning and positive dimensions of psychological health.

Gender may further intersect with relationship functioning, neighborhood quality, and psychological health for Black Americans. Wickrama et al. (2010) found that African American husbands’ (but not wives’) perceptions of community disorder predicted their own and their wives’ hostile spousal behaviors; in turn, hostile spousal behaviors predicted poorer mental health for both individuals and their partner. The authors note that women may have been better able to protect their relationships from direct influences of harmful external environments but were still psychologically susceptible to these effects indirectly through their husbands’ reactions and the resulting negative spillover to their relationship functioning.

Current Study

The current study was designed to advance the literature on Black Americans’ psychological health by utilizing a dimensional approach to examine negative and positive affect among Black Americans across 10 years, investigating relationship adjustment and perceived neighborhood quality as independent and interactive predictors of psychological health and examining gender differences in these associations. We hypothesized that neighborhood quality would interact with relationship adjustment to predict psychological health and that these linkages would vary by gender. Specifically, we anticipated that higher perceived neighborhood quality would attenuate the associations between poorer relationship adjustment and negative affect and strengthen the associations between better relationship adjustment and positive affect, with these effects being stronger for Black men, given prior research that they may be more sensitive to their neighborhood context.

Method

Data for the current investigation were from the Midlife in the United States study (MIDUS; http://www.midus.wisc.edu). MIDUS is a longitudinal epidemiological study of health and aging that includes assessments of psychological factors, including both negative and positive psychological health. The second wave (M2, conducted from 2004 to 2006) and third wave of MIDUS (M3, conducted from 2013-2015) oversampled African Americans from Milwaukee, Wisconsin, to increase the racial diversity of the sample. Milwaukee participants were recruited by using a stratified probability sampling design to include participants from census blocks that varied in median household income (e.g., areas with median household income above or below $40,000) and racial composition (e.g., areas with at least 40% of Black/African American residents) to capture the experiences of Black Americans from a variety of neighborhoods.

Transparency and Openness

Consistent with TOP Guidelines, we report how we determined our sample size, all data exclusions (if any), all manipulations, and all measures in the study. Data and materials used in the current study are publicly available at https://www.icpsr.umich.edu/web/ICPSR/series/203. Analytic code for this study is available upon request. This study was not pre-registered.

Participants

Of the total 5,555 participants at M2, individuals who were not partnered (i.e., not married or in cohabitating relationships; n = 1,637) and those who did not identify as Black/African American (n = 3,585) were excluded, resulting in the current analytic sample of 333 participants.¹ Fifty-four percent of the analytic sample was male. Participants ranged from 34 to 83 years of age at M2 (men: M = 53.02, SD = 11.61; women: M = 49.86, SD = 11.27). Most participants (84%) were married; the others were in cohabitating relationships. At M2, the median household income was $56,650.

Procedure

Details on the recruitment and study procedures for MIDUS are described elsewhere (Ryff & Krueger, 2018). MIDUS participants were initially recruited in 1995-1996 via random digit dialing. Of the original 4,963 participants who completed data collection during M2, 3,294 participants completed data collection during M3 (approximately 10 years apart). Of the 592 Milwaukee participants from M2, 389 participants completed data collection at M3. Data collection procedures included measures of psychosocial functioning, personal relationships, and physical and mental health collected via computer-assisted phone interviews and self-administered questionnaires. The University of Wisconsin-Madison Institutional Review Board approved study protocols and procedures, and all participants provided written informed consent before being assessed.

Measures

Psychological Health

Psychological health was assessed at M2 and M3 via measures of negative and positive affect using two 6-item scales. For negative affect, participants rated how frequently in the past 30 days they felt “so sad nothing could cheer [them] up, nervous, restless or fidgety, hopeless, that everything was an effort, and worthless” (i.e., K6 nonspecific psychological distress scale; Kessler et al., 2002, 2003). For positive affect, participants rated how frequently in the past 30 days they felt “cheerful, in good spirits, extremely happy, calm and peaceful, satisfied, and full of life” (Assari et al., 2018; Mroczek & Kolarz, 1998). Response options ranged from 1 (none of the time) to 5 (all of the time). Items were averaged within each domain, with higher scores reflecting greater negative and positive affect, respectively. Cronbach's alpha was .84 at M2 and .85 at M3 for negative affect and .91 at M2 and M3 for positive affect.

Relationship Adjustment

Relationship adjustment was assessed at M2, with six items related to relationship support and six items related to relationship strain (Grzywacz & Marks, 1999). Questions related to relationship support encompassed feelings about being appreciated, cared for, and understood by one’s partner and being able to relax, open up, and rely on them. Questions related to relationship strain encompassed perceptions of tension, arguments, being let down, demands, criticism, and irritability. Response options ranged from 1 (often/a lot) to 4 (never/not at all). Items were averaged, with items related to relationship support recoded so that higher scores reflected better relationship adjustment. Cronbach’s alpha was .89.

Neighborhood Quality

Neighborhood quality was assessed at M2 via six items related to neighborhood characteristics. We used Keyes's (1998) 4-item measure of neighborhood quality, which includes two items assessing neighborhood social cohesion (“People in my neighborhood can trust each other;” “I could call on a neighbor for help if I needed it”) and two items assessing safety (“I feel safe being out alone in my neighborhood during the daytime;” “I feel safe being out alone in my neighborhood at night”). Two additional items assessing the physical environment (“Buildings and streets in my neighborhood are kept in very good repair;” “My neighborhood is kept clean”) were also included for the current study to enhance measurement reliability and because of prior research and theory highlighting the relevance of the physical environment as an aspect of neighborhood quality (Leung & Takeuchi, 2011; Ross & Mirowsky, 1999). Participants rated their agreement with these statements using options ranging from 1 (not at all) to 4 (a lot). Scores were summed, with higher scores reflecting better neighborhood quality. Cronbach’s alpha was .78.

Attrition, Missing Data, and Covariates

Approximately 60% of individuals in the analytic sample who provided data at M2 also completed data collection at M3 10 years later (n = 187), which is comparable to attrition rates for other longitudinal studies of aging (e.g., Singh et al., 2019) and consistent with findings regarding attrition rates for non-White participants (Radler & Ryff, 2010). We explored all key variables and found that the M2 scores for relationship adjustment, neighborhood quality, and psychological health did not differ significantly between those who provided data at both M2 and M3 and those who only provided data at M2 (ps > .220). However, gender was related to missingness (p = .003), as men were less likely to provide data at M3. Attrition analyses indicated that higher age, lower income, and not working were each associated with greater likelihood of missing M3 data (ps ≥ .003), and these variables were all included as covariates. Because gender, age, income, and work status were included in the model, data were treated as missing at random, meeting the assumptions of Maximum Likelihood approaches. Covariates were coded as age (in years), education (highest level participants completed, ranging from 1= not completing any school or only some grade school to 12 = completing a graduate degree), household income (log-transformed), and work status (0 = currently working, 1 = not working).

Data Analysis

Analyses were conducted using path modeling in Mplus version 8.2 (Muthén & Muthén, 1998-2017). We used full information maximum likelihood parameter estimates with Mplus defaults (i.e., using the observed variance-covariance matrix to estimate the likelihood function for each case) and robust standard errors (MLR). These estimates take into account missing data and nonnormality to compute unbiased parameter estimates based on the likelihood of the observed values and adjusted for patterns of missingness due to included auxiliary covariates (noted above). This is considered an appropriate alternative to multiple imputation and produces unbiased estimates of parameters and standard errors under conditions up to missing at random (as in the current analyses; Graham, 2009), allowing all 333 participants with complete data at M2 to be included in analyses. We estimated a model in which M2 psychological health, relationship adjustment, and neighborhood quality were entered as simultaneous predictors of M3 psychological health (Figure 1). A two-way interaction term between neighborhood quality and relationship adjustment was also included as a predictor of M3 psychological health. Negative affect and positive affect were tested as predictors and outcomes in separate models. All variables were centered before creating the interaction terms, and the covariance of the interaction term with its components was constrained to improve model convergence (constraints did not change the size or direction of effects). All other predictors were allowed to covary. Significant interactions were probed at low and high (one SD below and above the mean) levels of the moderator (i.e., neighborhood quality).

Figure 1.

Prospective prediction of psychological health at MIDUS 3 by relationship adjustment, neighborhood quality, and psychological health at MIDUS 2.

Note. Age, education, work status, and household income were also included as covariates in the models.

To determine if the associations were moderated by gender, we conducted multi-group analyses using the chi-square difference test (adjusted for MLR) to compare models in which paths were constrained to be equivalent across gender (i.e., the same for men and women) versus freely estimated. When the chi-square test suggested that the constrained (i.e., more parsimonious) model was a significantly poorer fit for the data, individual paths were examined to determine which could be constrained. If model fit did not significantly degrade when paths were constrained to be the same, those constraints were retained in the final model to enhance model parsimony. Consistent with recommendations by Hu and Bentler (1999), good model fit was interpreted in terms of a non-significant chi-square test, Root Mean Square Error of Approximation (RMSEA) less than or equal to .06, Standardized Root Mean Square Residual (SRMR) less than or equal to .08, and a Comparative Fit Index (CFI) greater than or equal to .95.

Results

Means, standard deviations, and correlations among study variables are presented in Table 1. Independent samples t-tests indicated that men reported significantly higher levels of relationship adjustment (t(244) = 2.58, p = .011) and positive affect (t(299) = 2.24, p = .026) than women. There was not a significant difference between women and men for negative affect (t(296) = −1.91, p = .057) or neighborhood quality (t(272) = 1.35, p = .179). In general, correlations among most study variables were significant. For both women and men, relationship adjustment and neighborhood quality at M2 were positively correlated. Relationship adjustment was positively associated with M2 and M3 positive affect for both women and men but was not significantly associated with negative affect at either M2 or M3 for women or at M3 for men. For men, neighborhood quality was negatively associated with M2 and M3 negative affect; neighborhood quality was positively associated with M2 and M3 positive affect for both women and men.

Table 1.

Means, Standard Deviations, and Correlations Among Relationship Adjustment, Neighborhood Quality, and Negative and Positive Affect for Men and Women

Men
Variables	M	SD	1	2	3	4	5	6	7	8	9
1. RA (M2) ‡	3.34	0.45	-
2. NQ (M2)	19.69	3.62	.17*	-
3. NA (M2)	1.59	0.71	−.28*	−.22*	-
4. PA (M2) ‡	3.77	0.72	.36*	.24*	−.32*	-
5. NA(M3)	1.56	0.72	.05	−.39*	.62*	−.17*	-
6. PA (M3)	3.68	0.74	.25*	.43*	−.30*	.41*	−.51*	-
7. Age (M2) ‡	53.02	11.61	.17*	.16	−.22*	.03	−.16	.12	-
8. Education (M2) a	5.96	2.34	−.07	.07	−.16*	−.05	−.38*	.11	−.02	-
9. Work Status (M2) b	1.39	0.49	.24*	−.05	.09	−.10	.30*	−.24*	.38*	−.18*	-
10. Income (M2)	57,331	39,287	.02	.16	.27*	.04	−.10	−.09	−.04	.39 *	−.24*
Women
Variables	M	SD	1	2	3	4	5	6	7	8	9
1. RA (M2)	3.17	0.66	-
2. NQ (M2)	19.10	3.62	.22*	-
3. NA (M2)	1.75	0.77	−.16	−.19*	-
4. PA (M2)	3.57	0.82	.28*	.31*	−.58*	-
5. NA(M3)	1.71	0.74	−.02	−.14	.58*	−.32*	-
6. PA (M3)	3.62	0.76	.21*	.30*	−.35*	.61*	−.55*	-
7. Age (M2)	49.86	11.27	.06*	.26*	−.23*	.16	−.11	.14	-
8. Education (M2) a	6.32	2.44	−.04	.10	−.13	.07	−.37*	.11	.01	-
9. Work Status (M2) b	1.42	0.49	−.02	.06	.23*	.01	.14	.04	.21*	−.23*	-
10. Income (M2)	61,804	48,614	.02	.12	−.30*	.23*	−.23*	.08	−.14	.29*	−.32*

Note. N = 333 participants. RA = Relationship Adjustment, NA = Negative Affect, PA = Positive Affect, NQ = Neighborhood Quality, M2 = MIDUS 2, M3 = MIDUS 3.

^‡Significant difference between men and women.

^aThe highest level of education participants completed, ranging from 1 = completing no school or only some grade school (grades 1-6) to 12 = completing a terminal graduate degree (e.g., Ph.D., Ed.D., M.D.).

^bCurrently working = 0, Not working = 1. Spearman correlations were used for work status.

^*p < .05

Prospective Associations Among Relationship Adjustment, Neighborhood Quality, and Negative Affect

The unstandardized estimates, p-values, and 95% confidence intervals for the model focused on relationship adjustment, neighborhood quality, and negative affect are presented in Table 2. With the exception of the interaction between M2 relationship adjustment and neighborhood quality in predicting M3 negative affect and the path from M2 work status predicting M3 negative affect, paths were constrained to be equivalent across men and women. Thus, the final model provided excellent fit to the data, (χ²(10) = 5.73, p = .837; RMESA = .00; SRMR = .03; CFI = 1.00). Regarding the covariates, education was negatively associated with M3 negative affect for men and women, and work status was positively associated with M3 negative affect for men; there were no other significant covariates. There was not a main effect of relationship adjustment in predicting later negative affect, but there was a main effect of neighborhood quality, such that higher neighborhood quality was associated with lower negative affect for both men and women 10 years later.

Table 2.

Unstandardized Coefficients, Standard Errors, P Values, and 95% Confidence Intervals for Multigroup Analyses of the Longitudinal Associations Among Relationship Adjustment, Neighborhood Quality, and Psychological Health

Effect	Men				Women
	B	SE	p-value	95% CI	B	SE	p-value	95% CI
M3 Negative Affect (Model 1)
NA (M2)	0.54	0.08	<.001	[0.38, 0.69]	0.54	0.08	<.001	[0.38, 0.69]
RA (M2)	0.08	0.06	.224	[−0.05, 0.20]	0.08	0.06	.224	[−0.05, 0.20]
NQ (M2)	−0.03	0.01	.007	[−0.06, 0.01]	−0.03	0.01	.007	[−0.06, 0.01]
RA X NQ (M2)	−0.08	0.04	.035	[−0.15, −0.01]	0.00	0.03	.863	[−0.05, 0.05]
Age (M2)	0.00	0.00	.930	[−0.01, 0.01]	0.00	0.00	.930	[−0.01, 0.01]
Education (M2)	−0.07	0.02	<.001	[−0.11, −0.03]	−0.07	0.02	<.001	[−0.11, −0.03]
Work Status (M2)	0.34	0.12	.007	[0.09, 0.58]	0.01	0.13	.961	[−0.24, 0.25]
Income (M2)	0.03	0.03	.260	[−0.02, 0.09]	0.03	0.03	.260	[−0.02, 0.09]
M3 Positive Affect (Model 2)
PA (M2)	0.48	0.08	<.001	[0.31, 0.64]	0.48	0.08	<.001	[0.31, 0.64]
RA (M2)	0.02	0.08	.793	[−0.14, 0.18]	0.02	0.08	.793	[−0.14, 0.18]
NQ (M2)	0.04	0.02	.004	[0.01, 0.07]	0.04	0.02	.004	[0.01, 0.07]
RA X NQ (M2)	0.01	0.03	.729	[−0.05, 0.07]	0.01	0.03	.729	[−0.05, 0.07]
Age (M2)	0.00	0.00	.414	[0.00, 0.01]	0.00	0.00	.414	[0.00, 0.01]
Education (M2)	0.03	0.02	.063	[−0.01, 0.07]	0.03	0.02	.063	[−0.01, 0.07]
Work Status (M2)	−0.06	0.10	.586	[−0.25, 0.14]	−0.06	0.10	.586	[−0.25, 0.14]
Income (M2)	−0.05	0.03	.111	[−0.10, 0.01]	−0.05	0.03	.111	[−0.10, 0.01]

Note. B = Unstandardized Coefficient, SE = Standard Error, CI = Confidence Interval, RA = Relationship Adjustment, NA = Negative Affect, NQ = Neighborhood Quality, PA = Positive Affect, M2 = MIDUS 2, and M3 = MIDUS 3. Significant associations are noted in bold.

Multi-group analyses revealed a significant gender difference in the interaction between relationship adjustment and neighborhood quality in predicting subsequent negative affect. Specifically, the relationship adjustment X neighborhood quality interaction was significant for men but not women. The interaction was probed post hoc at lower (1 SD below mean) and higher (1 SD above mean) levels of neighborhood quality to examine the nature of the associations between relationship adjustment and later negative affect (Figure 2). At lower levels of neighborhood quality, better relationship adjustment was associated with higher negative affect for men 10 years later (b = 0.37, p = .022), whereas, at higher levels of neighborhood quality, relationship adjustment was not associated with negative affect (b = −0.21, p = .127).

Figure 2.

Association between relationship adjustment (at MIDUS 2) and negative affect (10 Years Later at MIDUS 3) for men and women at low (1 SD below) versus high (1 SD above) levels of neighborhood quality.

Prospective Associations Among Relationship Adjustment, Neighborhood Quality, and Positive Affect

Results for the model focused on relationship adjustment, neighborhood quality, and positive affect are also presented in Table 2. All paths were constrained to be equivalent across men and women, with the final model providing good fit to the data, (χ²(12) = 12.831, p = .382; RMESA = .02; SRMR = .05; CFI = .99). There were no significant covariate findings. Relationship adjustment was not significantly related to subsequent positive affect for men or women. There was a main effect of neighborhood quality such that higher neighborhood quality was associated with greater positive affect 10 years later for both men and women. Multi-group analyses did not indicate a significant interaction between relationship adjustment and neighborhood quality in this model for men or women.

Sensitivity Analyses

To better understand how the specific aspects of neighborhood quality (social cohesion, safety, and physical environment) were related to psychological health and/or interact with relationship adjustment to predict psychological health, we conducted additional analyses using the three sub-components of neighborhood quality, resulting in three additional models per outcome. Models were otherwise identical to the ones presented above, exchanging the composite measure of neighborhood quality with the three subcomponents of neighborhood quality in separate models. Tables summarizing the unstandardized estimates, p-values, and 95% confidence intervals for all sensitivity analyses are in the supplemental material. After accounting for the significant covariates of education and work status, we found that, as a main effect, perceptions of neighborhood safety were related to less negative affect and more positive affect at M3 for both men and women and that neighborhood social cohesion was related to more positive affect at M3 for both men and women. For the models including neighborhood safety and physical structure, multi-group analyses did not indicate any significant interactions with relationship adjustment for either men or women predicting either M3 negative or positive affect. For the model with social cohesion, the gender difference in the path for the M2 relationship adjustment X social cohesion interaction predicting M3 negative affect approached significance (χ² (1) = 3.64, p = .057). When this path was unconstrained, the relationship adjustment X cohesion interaction was significant for men (B = −0.17, p = .032) but not for women (B = 0.02, p = .759). Specifically, at lower levels of social cohesion, better relationship adjustment was associated with higher subsequent negative affect for men (b = 0.68, p = .018), whereas, at higher levels of neighborhood quality, relationship adjustment trended towards lower subsequent negative affect (b = −0.51, p = .070).

Discussion

Given disparities in psychological distress for Black Americans, attempts to understand their psychological health from a dimensional perspective (as consistent with RDoC) are valuable (Carter et al., 2021), particularly when coupled with socioecological and intersectional perspectives to highlight the heterogeneity in risk for, or resilience to, poorer mental health outcomes. This study was designed to utilize such an approach by investigating relationship adjustment and neighborhood quality as prospective independent and interactive predictors of negative and positive dimensions of psychological health in a racially homogenous sample of partnered Black Americans, examining gender differences in these associations. Results demonstrate that, even when accounting for prior levels of the outcomes a decade earlier, perceived neighborhood quality is a long-term predictor of Black Americans’ psychological health and that the 10-year interactive associations between relationship adjustment and neighborhood quality for negative dimensions of health among Black Americans differ between men and women.

Psychological Health in the Context of Relationship Adjustment and Neighborhood Quality

We found that, for Black men, the linkages between relationship adjustment and negative affect varied as a function of neighborhood quality. Men who reported better relationship adjustment and lived in lower-quality neighborhoods (particularly those with less social cohesion) endorsed higher levels of negative affect 10 years later. These findings are in line with previous research demonstrating Black men’s relative psychological sensitivity to neighborhood stress and the ways external environments relate to their psychological health through or in combination with their own relationship functioning (Erving, 2022; Wickrama et al., 2010).

It could be that men in better quality relationships wish to provide the most desirable residential environments for their partners. Yet, the experience of poorer neighborhood quality may be an ambient environmental stressor that interferes with attempts to fulfill their role as the provider and protector for their families. As documented in qualitative research, some married Black men view protecting and providing for their families and displaying affection, love, and partnership as salient relationship characteristics and part of their primary responsibilities within the family (Griffith et al., 2011; Hurt et al., 2017). As such, men in better adjusted relationships who reside in lower-quality neighborhoods may be distressed not being able to provide high-quality residential environments for their partners and family, as evidenced by poorer perceptions of neighborhood safety, cohesion, and physical environment (Thorpe et al., 2020), resulting in heightened emotional distress over time. In contrast, men residing in higher-quality neighborhoods may not experience the same external stress and be better positioned to capitalize on the beneficial effects of their romantic relationships and the psychosocial resources afforded by their neighborhoods, resulting in less emotional distress over time. Future studies that assess Black men’s cognitions about their roles as familial protectors, and the ways that these are related to environmental stress, would add additional nuance for understanding the linkages among relationship adjustment, neighborhood quality, and negative affect among Black men.

Notably, relationship adjustment and neighborhood quality did not interact to predict Black women’s psychological outcomes in the current study. This may be due, at least in part, to the cultural and societal pressure Black women face that can make it harder for them to capitalize on the psychological benefits of relationship adjustment (Woods-Giscombé, 2010), even in higher quality neighborhoods. Additional research is needed to better understand the ways intimate relationships and neighborhood characteristics may interact to predict Black women’s mental health, with future work benefiting from examining dyadic linkages in this process and attending to the potential spillover from romantic partners’ distress to women’s psychological health (Wickrama et al., 2010).

Neighborhood Quality

In addition to the interactive associations between relationship adjustment and neighborhood quality in the prediction of negative affect, we found evidence of a main effect for neighborhood quality for psychological health. For both negative and positive dimensions of psychological health, higher neighborhood quality was associated with better psychological functioning 10 years later—even after accounting for initial levels of psychological health and other sociodemographic characteristics, including education and income. Further, the absence of gender differences in these associations suggests the salience of ecological resources for Black men and women. These findings support theoretical frameworks emphasizing “neighborhood- and community-level factors …[as] robust predictors of health outcomes” (p. S19, Alvidrez et al., 2019) and are consistent with empirical research demonstrating the impact of community participation and social contact for both negative and positive dimensions of concurrent psychological health (Erving & Hills, 2019). The current work extends this research by underscoring the long-term associations between perceptions of neighborhood resources and individuals’ psychological health a decade later. Given the impact of structural forces on Black residential environments (Williams & Collins, 2001), these findings suggest that one potential avenue for communities to cultivate resilience to institutional racism may be through investment in social ties/connections, safety resources, and physical or structural improvements that inform neighborhood perceptions and contribute to mental health. Additional research that disentangles the specific mechanisms linking neighborhood characteristics to long-term mental health outcomes for Black Americans may uncover opportunities for intervention (via leveraging community resources) to decrease psychological distress and enhance flourishing in communities.

Romantic Relationships

Contrary to expectation, we did not find a main effect of relationship adjustment for either negative or positive affect for men or women. These results differ from the growing literature demonstrating the prospective effects of relationship quality on Black Americans’ psychological health over periods of 1-2 years (e.g., Barton et al., 2022). One potential explanation for this divergent pattern of results is the relatively long period between assessments in the current study (approximately 10 years), as the links between relationship functioning and psychological health may be more salient on shorter timescales. Indeed, a meta-analysis found that the association between depression and marital quality is strongest when time frames are 12 months or shorter (Egan & Smith, 2018). Alternatively, it could be that affective dimensions of psychological health, as utilized in the current study, have weaker linkages with intimate relationships than symptoms of psychopathology (e.g., depressive symptoms), which have typically been used in previous prospective research (e.g., Jenkins et al., 2020). Additional work is necessary to examine which (or all) of these explanations best characterize the ways relationship adjustment relates to long-term psychological outcomes in this population. Thus, the current results should be viewed as one piece of evidence contributing to the growing knowledge base on Black Americans’ relationship adjustment and psychological health, highlighting the benefits of employing ecological-intersectional perspectives to understand how these constructs relate over time in this population. Indeed, our results suggest that the nature of the long-term associations between relationship quality and Black Americans’ psychological health (particularly negative dimensions) are not universal but depend on both neighborhood context and gender.

Limitations and Future Directions

The current investigation has multiple methodological strengths, including the use of a large, national, racially homogenous sample of partnered Black Americans at midlife, the separate examination of negative and positive dimensions of psychological health, and a longitudinal design that included data collected 10 years apart, controlling for prior levels of the outcome variables. There are also limitations. For instance, it is possible that some of the associations observed in this study were due to shared method variance resulting from reliance on single-reporter, self-report measures. Future work should include longitudinal assessments of both romantic partners to examine the linkages among relationship adjustment, neighborhood quality, and psychological health in a dyadic context, and, where possible, utilize multiple methods of measurement (e.g., self-report, interview-based, and observational measures) for individual and relationship functioning and include objective measures of neighborhood quality (e.g., area deprivation index; Kind et al., 2014) to decrease shared method variance. Additionally, given that MIDUS is not a nationally representative sample and that we utilized data from an oversample of individuals from Milwaukee, these findings may not necessarily generalize to the entire Black American population. More research is needed to examine these linkages with more geographically and ethnically diverse samples (e.g., Afro-Caribbeans; Erving & Hills, 2019). Future work in this area would also benefit from examining these processes among Black gender-minoritized populations. Other considerations include the fact that, given the 10-year timespan between assessments, it is possible some participants changed neighborhoods by M3, moving to neighborhoods of either lower or higher quality than the ones on which they originally reported. Data on this were not available for the current study. Still, extant research (Russell et al., 2018) suggests that, regardless of such residential changes, the quality of the neighborhoods individuals report on can have long-term associations with their psychological health. Future research that replicates this pattern of findings with participants who are stable in terms of their residential environment would further demonstrate the robustness of the current results. Relatedly, 71 participants in the current study experienced a change in their relationship status between assessment periods; however, it is unknown how many individuals who reported being partnered at both M2 and M3 were in a relationship with the same partner at each wave. Because the primary research questions for the study included investigating whether participants’ previous reports of relationship adjustment predicted their long-term psychological health (as well as potential interactions between relationship adjustment and neighborhood quality and gender differences in these associations), our ability to answer our research questions was not contingent on whether participants remained in stable relationships between assessments, and we retained all participants who reported being partnered at M2. However, future work may explore whether the current findings replicate in samples of Black adults in enduring relationships and/or if these associations differ between individuals who experience changes in their relationship status over time and those in stable relationships.

Clinical and Policy Implications

These limitations notwithstanding, the findings from this study have clinical and policy implications. Findings highlight the importance of attending to social and ecological factors. In collaboration with clients, clinicians should survey the environmental resources available to Black individuals and couples, as these may be existing strengths that can be used to promote psychological health. Practitioners can work with clients to develop strategies to minimize the adverse effects of low neighborhood quality on psychological health and explore the potential benefits of becoming active participants in their communities (Kelly et al., 2020). Practitioners can also help individuals in identifying available community resources and assist their clients, particularly men, in recognizing the relevance their residential surroundings may have for their relational/psychological functioning, contextualizing relational and psychological difficulties within impinging institutional structures outside the individual or family (Hollie & Coolhart, 2020). Regarding public policy, additional attention should be devoted to the connections among neighborhoods, romantic relationships, and psychological health. Policy initiatives and interventions aimed at remediating neighborhood disadvantage and improving neighborhood quality may enhance individuals’ psychological health. Targeting community perceptions, potentially through establishing community-building interventions, neighborhood block coalitions, or community beautification programs, may serve as an opportunity for intervention at the neighborhood level (Robinette et al., 2016). Further, offering mental health services and relationship resources may be especially beneficial in disadvantaged neighborhoods.

Conclusion

This study illustrates the prospective linkages between neighborhood quality and both negative and positive dimensions of psychological health and highlights gender differences in the interactive, longitudinal associations between relationship adjustment and neighborhood quality for negative affect among Black Americans. Future work that continues to examine both negative and positive dimensions of psychological health will advance current representations of Black Americans in the literature, and investigations of the relevance of Black intimate relationships and ecological resources for functioning in this area will enhance our understanding of how they can be used to potentiate individual health and well-being over time.