Parents’ Experiences of Discrimination and Family Relationship Qualities: The Role of Gender

Abstract

Mothers and fathers in 156 African American families reported on racial discrimination experiences, gendered traits, and warmth and conflict in family relationships. Discrimination was linked with relationship quality, but links differed for mothers and fathers. More expressive parents and less instrumental fathers had more positive relationships in the face of discrimination, but for more instrumental fathers, discrimination–relationship quality links were negative. Findings imply consideration of sociocultural and individual characteristics for family relationships.

Racial discrimination is common in the lives of African Americans, with 60–90% of African American adults reporting at least one incident of discrimination per year (Kessler, Mickelson, & Williams, 1999; Landrine & Klonoff, 1996). In turn, many studies show that discrimination is a stressor that is linked to men’s and women’s reports of poor psychological functioning and mental health (Bennett, Merritt, Edwards, & Sollers, 2004; Fischer & Shaw, 1999; Lincoln, Chatters, Taylor, & Jackson, 2007). In the face of research on individuals’ reactions to discrimination, however, we know little about how experiences of discrimination affect families. Research on racial discrimination also documents consistent gender differences, with males reporting more frequent experiences of discrimination than females (Kessler et al.; Stevenson, Cameron, Herrero-Taylor, & Davis, 2002). Although a body of research suggests that males and females react differently to stressful experiences (Taylor 2006; Taylor, Klein, Lewis, Gruenewald, Gurung, & Updegraff, 2000), we know very little about the potentially different responses of males and females to the stressors of racial discrimination.

Links between stressful experiences and family relationships have been documented in European-American families (e.g. Taylor et al., 2000), but research examining the implications of psychosocial stressors for African American family relationships is limited. Although some empirical work has documented associations among stressful events, discrimination experiences, and mothers’ family relationships (Murry, Brown, Brody, Cutrona, & Simons, 2001; Murry et al., 2008), there are gaps in this literature. First, we know little about the potential impact of racial discrimination for African American families who do not face chronic stressors. In addition, research on African American families has often focused on single parents, who tend to be mothers, and we know little about discrimination experiences and relationship qualities in two-parent African American families. Fathers’ transactions within the family may have different, and possibly stronger, implications for family well-being than those of mothers, which may be due to differences in mothers’ versus fathers’ family involvement (Brody & Flor, 1997). Furthermore, examining the links between racial discrimination and mothers’ and fathers’ family relationships has practical implications as family interventions may be more effective for African American families when sociocultural stressors are better understood.

Grounded in research and theories of stress and gender, the overarching goal of this study was to assess how racial discrimination experiences, one type of contextual stressor, influence the nature and quality of parent-adolescent, coparenting and marital relationships. Further, to determine whether discrimination–family relationship links were different for mothers and fathers, we tested the moderating role of parent gender, and we also examined parents’ gendered personality qualities, specifically expressivity and instrumentality, as potential moderators of the links between discrimination and family relationships.

The Role of Gender in Stress Responses

Our examination of differences between mothers’ and fathers’ reactions to discrimination is grounded in research on males’ and females’ responses to stress (Taylor et al., 2000). Whereas males tend to exhibit an active problem focused “fight-or-flight” response in the face of stressors, Taylor and colleagues also proposed that females are more likely to display “tend-and-befriend” responses; that is, women tend to lend and seek support in response to stressors. Empirical support for gender differences in stress responses comes from several studies. Aneshensel, Rutter, and Lachenbruch (1991) found that males were more likely to report active, problem-focused responses to perceived stress, whereas females reported more expressive, emotional responses when they perceived more stress. Based on the work of Neighbors and Howard (1987), some empirical research has found that men were less inclined than women to seek social support in the face of psychological distress (e.g., Lane & Addis, 2005). Women tended to report better psychological adjustment than men in the face of reported stress, which may be due to their inclinations to seek out social support (e.g., Black, Cook, Murry, & Cutrona, 2005). Specific to the current study, in the context of discrimination, African American adolescent boys exhibited more externalizing behaviors than girls (Brody et al., 2006).

In the face of this research, however, most of what we know about gendered responses to stressors has not yet been applied to African American families, and no research has explored whether the stressor of racial discrimination is linked to men’s and women’s family relationships. Thus, in developing our hypotheses, we drew on studies of work-family stress, which document the processes through which stressful experiences at work affect the quality of family life, and highlight gender differences in the links between work stressors and family relationships. Findings from two such studies revealed that on days when fathers reported higher stress at work, they were more likely to engage in conflict or to withdraw from their families (Repetti, 1989). In contrast, when mothers reported more work stress, they engaged in more nurturing behaviors at home (Repetti & Wood, 1997). One reason for these differences may be the nature of family roles: Women’s roles in the family may be more scripted and less influenced by outside experiences, whereas men’s roles are less scripted, and therefore, more susceptible to extra-familial experiences (Crouter, Helms-Erikson, Updegraff, & McHale, 1999).

Although the role of race/ethnicity in stress responses has not been a focus of substantial research, affiliation in the presence of stress may be especially evident among racial ethnic groups for whom family and extended kinship networks are salient. In stressful situations, external resources may be scarce, and family networks can provide emotional support and assistance for African Americans (e.g., Lincoln, Chatters, & Taylor, 2005). Indeed, some research shows that African American women report greater use of informal support networks to cope with the stressor of discrimination as compared to women from other racial ethnic groups (e.g., Thompson, 2006). Further, findings from a study on the psychological impact of racial discrimination on African American men suggest that social support may play a role in buffering individuals from racism-related stress (Utsey & Payne, 2000). However, there is less empirical evidence about the circumstances that contribute to men’s tendencies to seek support from family members. Given the salience of family relationships in some racial ethnic groups, our primary goal was to understand whether and how mothers’ and fathers’ experiences of discrimination were associated with their family relationships. Knowledge of differences between African American mothers’ and fathers’ responses to discrimination has important implications for interventions that may reduce the negative effects of such stressors.

The Role of Gender-Typed Personality Traits in Responses to Stressors

Gender-typed traits have been studied in terms of instrumentality, a stereotypically masculine set of qualities including assertiveness and independence, and expressivity, a stereotypically feminine set of attributes including tendencies to display emotionality and affection (Huselid & Cooper, 1994). Although men are more instrumental and women are more expressive, on average (Orlofsky & O’Heron, 1987), the relations between sex and gendered traits are not absolute: There is variability in gender-typed characteristics within each sex (Hoffman, Powlishta, & White, 2004). Thus, it may be that biological sex reflects the gendered traits (i.e., instrumentality and expressivity) that may be responsible for observed differences in men’s and women’s responses to stressors. For example, tendencies to engage in conflict or withdrawal from family relationships may be an artifact of instrumental, problem-focused attributes. Particularly in the case of a stressor such as racial discrimination, which is beyond an individual’s personal control, attempts to manage stress using a problem-focused strategy may promote frustration, which in turn spills over into conflict with or withdrawal from family members. From another perspective, individuals with more instrumental traits (i.e., males) may be less inclined to seek support when faced with a stressor because this response may be incompatible with a tendency towards self-reliance or independence (Addis & Mahalik, 2003). In contrast, the tendency to affiliate with others in the face of a stressor may be a function of having more expressive traits. In this study, we tested whether gender-typed traits accounted for observed sex differences in stress responses.

Research on gender and mental health suggests that individuals who are both highly expressive and instrumental (i.e., androgynous) may have the best mental health outcomes, compared to individuals who are less expressive or instrumental (Williams & D’Alessandro, 1994). In the context of relationships, however, an expressivity hypothesis suggests that being more expressive is a key factor in positive family relationship qualities (Ickes, 1993). Given the potential implications of gender-typed attributes, namely expressive traits, for relationship qualities, we were interested in whether and how mothers’ and fathers’ gendered attributes (i.e., more expressive/less instrumental and less expressive/more instrumental, respectively) moderated associations between their discrimination experiences and relationship qualities.

The Present Study

Prior research has linked experiences of discrimination to individual well-being. We extended this work by investigating the implications of discrimination for family relationships. Our first aim was to explore the associations between mothers’ and fathers’ reports of discrimination and their experiences of warmth and conflict in parent-adolescent, coparenting, and marital relationships. However, we did not have specific expectations about the patterns of association between discrimination and relationship qualities for each type of family relationship. The second aim of this study was to explore the role of gender in these associations. First, to assess the role of biological sex, we tested the moderating effects of parent gender on discrimination–family relationship links. Next, to illuminate one basis for observed gender differences, we tested whether gendered traits moderated racial discrimination–family relationship linkages.

We tested three hypotheses. Consistent with research on biologically based sex differences in males’ and females’ stress responses, our first hypothesis was that mothers would report more warmth in their relationships with family members (i.e., a “tend-and-befriend” reaction), but that fathers would report more conflict and less warmth in family relationships (i.e., a “fight-or-flight” reaction) when they experienced more frequent racial discrimination. To understand whether gendered traits were responsible for observed sex differences, our second hypothesis was that mothers and fathers with more expressive traits would exhibit more warmth and less conflict in family relationships in the face of discrimination. Thirdly, we hypothesized that mothers and fathers with more instrumental traits would report more conflict and less warmth in their relationships in the face of discrimination.

Method

Participants

The data came from mothers, fathers, and adolescents in 156 families that participated in the second phase of a three-year longitudinal study of relationships in two-parent African American families. Given the goals of the larger investigation, a study of family gender socialization, we did not seek a representative sample. Rather, we targeted families that self-identified as Black or African American and included a mother and father who were living together with at least two adolescent aged offspring. Recruitment took place in two urban centers in the northeast with substantial African American populations, and we used two strategies to generate the sample (see McHale et al., 2006 for more details on recruitment procedures). First, we hired African Americans residing in targeted communities to recruit families by advertising in businesses, churches, and at community events. Approximately half of the sample was recruited in this way. To recruit the rest of the sample, we purchased a marketing list of names and addresses of families with offspring in grades 4–7. We sent letters describing the study, and interested and eligible families called a toll free number or returned a postcard.

Of the original 202 families participating in the larger study, 11 mothers and fathers who were not African American were omitted, 27 fathers refused to participate, 6 families did not participate, and 2 fathers were deployed in phase 2 of the study. These families did not differ from the sample as a whole on any key variables. Of the 156 families in the present analyses, mean ages of mothers and fathers were 41.91 (SD = 6.12), and 44.59 (SD = 7.79), respectively. In terms of marital status, 84.62% of parents were married and all couples lived together for at least 4 years at the time of the interview. Mothers’ and fathers’ education was 14.59 years (SD = 1.90) and 14.23 years (SD = 2.37), respectively, indicating, on average, parents had completed “some college”. Parents’ combined income was $86,614.90 (SD = $57,572.73), which falls between the medians for two-earner families in the states from which data were collected ($98,163 and $74,884, respectively; US Census Bureau, 2007). Together, these figures are indicative of a largely working to middle-class sample. Most parents were employed (88.64% of fathers and 82.69% of mothers), with fathers working more hours per week (M = 42.27, SD = 19.75), on average, than mothers (M = 31.29, SD = 17.46). The early adolescents in these families were, on average 11.43 years of age (SD = 1.09). The majority of youth were biologically related to both parents (94.87% to mothers, 78.21% to fathers), and the sample was approximately equally divided by gender (n = 84 girls, n = 72 boys).

Procedures

Mothers and fathers from each family were interviewed individually in their homes by a team of two interviewers, almost all of whom were African American. Interviewers began by obtaining informed consent. Family members reported on relationship experiences, individual characteristics and attitudes, and individual well-being during the past year. Interviews generally lasted 2 hours. Following the completion of interviews, families were sent a $200 honorarium.

Measures

Racial discrimination experiences were assessed by the 11-item Experiences with Discrimination Scale, developed by Murry and colleagues (2001). Mothers and fathers used a 4-point rating scale (1 = never to 4 = several times) to indicate how often they had experienced different types of discrimination during the past year (e.g., “How often has someone said something derogatory or insulting to you just because you are African American?”). Cronbach’s αs were .86 for mothers’ and .90 for fathers’ reported discrimination.

Parent-adolescent relationship warmth and conflict were assessed using two separate measures. Parents reported their warmth towards offspring (e.g., “I am a person who makes my child feel better after talking over his/her worries with me”) using an eight-item, 5-point rating scale (1 = not at all to 5 = very much) from the parent version of the Child’s Report of Parental Behavior Inventory (Schwarz, Barton-Henry, & Pruzinsky, 1985). Cronbach’s αs were .83 for fathers’ reports of warmth, and .86 for mothers’ reports of warmth with adolescent offspring. Parents also reported the frequency of conflict with their offspring across 11 domains (Smetana, 1998) such as chores and social life, on a 6-point scale (1 = not at all to 6 = several times a day). Cronbach’s αs were .87 for mothers’ and .89 for fathers’ reports of conflict with offspring.

Marital satisfaction was measured using the Couple Relationship Domains scale (Huston, McHale, & Crouter, 1986). Each parent rated satisfaction in eight domains, such as household work, decision making, and communication, on a 9-point scale, with responses ranging from 1 = extremely dissatisfied to 9 = extremely satisfied. Cronbach’s αs were .89 for mothers and .93 for fathers. Marital conflict was assessed with a 5-item subscale from the Relationships Questionnaire (Braiker & Kelley, 1979). Items such as “How often do you feel angry or resentful toward your partner” were rated on a 9-point scale, with responses ranging from 1 = not at all to 9 = very often. Cronbach’s αs were .81 for mothers and .78 for fathers.

Coparent cooperation and conflict were assessed with a measure developed by Margolin, Gordis, and John (2001). All items were rated on a 5-point scale, ranging from 1 = not at all to 5 = almost always. The cooperation scale, consisting of 5 items (e.g., “My spouse asks my opinion on issues relating to parenting”), reflects the degree to which mothers and fathers support each other as parents. Cronbach’s αs were .77 for mothers and .73 for fathers. Conflict specific to parenting matters was assessed with six items (e.g., “My spouse argues with me about our children”). Cronbach’s αs were .81 for mothers and .79 for fathers.

Gender-typed traits were assessed using the Bem Sex Role Inventory (Bem, 1974). Using a 7-point scale (1 = never, or almost never true to 7 = always, or almost always true), mothers and fathers indicated how well a list of 20 instrumental, stereotypically masculine adjectives (e.g. “self-reliant”, “assertive”) and 20 expressive, stereotypically feminine adjectives (e.g., “tender”, “warm”) applied to their personalities. Cronbach’s αs ranged from .80 (fathers’ expressive qualities) to .87 (fathers’ instrumental qualities).

Background characteristics

Family background characteristics, reported by parents, included family members’ ages, genders, education levels, and parental employment status.

Results

Means, standard deviations, and correlations for all variables are presented in Table 1. These data suggest that family relationship quality was generally positive, that is, above the scales’ midpoints, which is indicative of a well-functioning sample. Also of note, expressivity and instrumentality had positive correlates for mothers and fathers, suggesting that in general both types of gendered traits were related to positive relationship qualities. In terms of discrimination experiences, mothers’ and fathers’ ratings were well below the midpoint of the 4-point rating scale, indicating that, on average, parents had “never” to “rarely” experienced discrimination during the past year. Despite high correlations between coparenting and marital qualities, a body of research (e.g., Bonds & Gondoli, 2007) documents that these are unique relationships in family systems; therefore, we retained all relationship measures in the following models.

Table 1.

Means, standard deviations, and correlations for study variables. (N = 312)

Variables	1	2	3	4	5	6	7	8	9	10	Mothers (M)	(SD)
1. Experiences of Discrimination	.06	−.04	.15*	−.01	.13*	.18**	.04	.06	.11	.04	1.58	(.48)
2. Expressivity	.05	−.08	.15*	.35**	−.13*	.20**	−.07	.19**	−.13*	−.12*	5.11	(.65)
3. Instrumentality	.13*	.24**	.05	.14*	.06	.24**	−.02	.11	.10	−.12*	5.08	(.76)
4. Parent-Adolescent Warmth	.22**	.37**	.16**	.18**	−.24**	.33**	−.23**	.28**	−.09	−.01	4.25	(.52)
5. Parent-Adolescent Conflict	.10	.03	.04	−.05	.28**	−.02	.24**	−.28**	.11	−.07	2.35	(.78)
6. Coparent Cooperation	.05	.41**	.19**	.34**	−.11	.23**	−.39**	.57**	−.18**	−.02	3.57	(.80)
7. Coparent Conflict	.25**	−.16**	.02	−.27**	.28**	−.47**	.36**	−.53**	.44**	.07	2.21	(.74)
8. Marital Satisfaction	−.15*	.31**	.18**	.37**	−.15**	.62**	−.59**	.37**	−.48**	−.16**	6.85	(1.59)
9. Marital Conflict	.26**	−.09	−.04	−.15**	.19**	−.35**	.57**	−.60**	.36**	.01	3.93	(1.59)
10. Family Income	.10	−.10	.10	−.03	−.07	−.06	.07	−.02	−.05	--	--	--
Fathers (M)	1.79	4.67	5.58	4.04	2.20	3.95	2.26	7.12	.69	--	--	--
(SD)	(.62)	(.67)	(.76)	(.57)	(.83)	(.67)	(.79)	(1.55)	(1.53)	--	--	--

Note. Mothers’ correlations above the diagonal, fathers’ correlations below the diagonal, correlations between parents on the diagonal.

^*p < .05,

^**p < .01

In a preliminary step, we examined mother-father differences in discrimination and in gender-typed traits using a series of repeated measures analyses of variance (ANOVAs). Consistent with prior work, fathers experienced discrimination more often than mothers, F(1, 311) = 24.18, p < .01. As expected, mothers had more expressive traits than fathers, F(1, 309) = 63.95, p < .01, and fathers had more instrumental traits than mothers, F(1, 309) = 68.73, p < .01.

Parents’ Experiences of Discrimination and Family Relationship Quality

To address the non-independence of the data (parents nested within families) we tested a series of multi-level models (MLM). An MLM approach extends ordinary least squares regression by accounting for data clustering and allowed us to examine within-family differences in mothers’ and fathers’ experiences and family relationship outcomes (e.g., Raudenbush & Byrk, 2002). For our first aim, we explored the main effects of mothers’ and fathers’ discrimination experiences on warmth and conflict in parent-adolescent, coparenting, and marital relationships. Our second aim was to test the moderating roles of gender and gender-typed traits in these links.

We tested three hypotheses in separate models. To determine whether there was a difference between mothers and fathers in the links between discrimination and family relationship qualities, we first tested the interaction between parent gender and discrimination. To determine whether gender-typed traits moderated these interactions, we tested parents’ expressive and instrumental attributes as moderators of discrimination–relationship quality links. To control for the potential effects of unmeasured stressors, and because prior work has evidenced a link between discrimination and socio-economic status (e.g., Kessler et al., 1999), family income was included as a control variable. We also tested for effects of fathers’ biological relatedness to offspring but this variable proved non-significant so is not included in the models. Given our directional hypotheses, we interpreted hypothesized interactions using a 1-tailed statistical test.

Discrimination was positively related to four dimensions of family relationships. Consistent with our expectations, discrimination predicted parent-adolescent conflict, β = .17, SE = .08, t = 2.09, p < .05, d = .34, coparenting conflict, β = .21, SE = .07, t = 2.84, p < .01, d = .46, and marital conflict, β = .48, SE = .15, t = 3.21, p < .01, d = .53. Inconsistent with the idea that discrimination is related to negative relationship qualities, findings revealed a positive link to parent-adolescent warmth, β = .12, SE = .06, t = 2.22, p < .05, d = .37.

The Moderating Role of Parent Gender

To test the first hypothesis, that parent gender would moderate discrimination family relationship linkages, parent gender was dummy coded (0 = mothers and 1 = fathers). Three Discrimination × Parent Gender interactions were significant, for parent-adolescent warmth, β = .22, SE = .10, t = 1.99, p < .05, coparenting cooperation, β = −.32, SE = .15, t = −2.09, p < .05, and marital satisfaction, β = −.65, SE = .31, t = −2.08, p < .05. We followed up these interactions by testing separate models for mothers and fathers. Follow-ups revealed that, for fathers, β = .23, SE = .07, t = 3.08, p < .01, d = .50, but not mothers β = .01, SE = .09, t = .06, ns, experiences of discrimination were related to more warmth with adolescents. Follow-ups also showed that for mothers, β = .30, SE = .14, t = 2.15, p < .05, d = .35, but not fathers, β = .04, SE = .09, t = .42, ns, discrimination was positively related to coparenting cooperation. In the case of marital satisfaction, follow-ups revealed that the discrimination–marital satisfaction link was positive for mothers and negative for fathers but neither coefficient was significant, β = .26, SE = .29, t = .90, ns, and β = −.32, SE = .21, t = −1.55, ns, respectively.

The Moderating Role of Gender-Typed Personality Traits

We next examined the moderating role of gender-typed personality traits in the links between discrimination and family relationship quality. To follow-up interactions, we used procedures outlined by Aiken and West (1991). Specifically, we performed a tertile split to distinguish high (one standard deviation above the mean for each parent) versus low (one standard deviation below the mean for each parent) levels of the moderator in each model.

Beginning with hypothesis two, that expressivity would moderate discrimination–relationship quality links for mothers and fathers, analyses revealed no evidence of three-way interactions, so these terms were dropped. Three significant two-way interactions between expressivity and discrimination emerged (Table 2). In terms of parent-adolescent conflict, for parents with lower expressivity, the link between discrimination and parent-adolescent conflict was positive, β = .27, SE = .11, t = 2.53, p < .01, d = .41, but the effect was non-significant for highly expressive parents, β = .01, SE = .06, t = .18, ns, (Figure 1, Panel A). Also as expected, for parents who were high in expressivity, the link between discrimination and coparent cooperation was positive, β = .39, SE = .09, t = 4.13, p < .01, d = .67 but non-significant for low expressivity parents, β = −.02, SE = .10, t = −.19, ns, (Figure 1, Panel B). Also as hypothesized, for parents low in expressivity, the link between discrimination and marital satisfaction was negative at trend level, β = −.35, SE = .21, t = −1.68, p < .10, d = .28, but non-significant for parents high in expressivity, β = .28, SE = .19, t = 1.44, ns, (Figure 1, Panel C).

Table 2.

Multi-level model results for discrimination predicting parents’ family relationship quality moderated by expressivity and instrumentality using residual maximum likelihood estimation (N = 312)

Family Relationship Quality
	Parent-Adolescent Warmth		Parent-Adolescent Conflict		Coparenting Cooperation		Coparenting Conflict		Marital Satisfaction		Marital Conflict
Predictors	β	(SE)	β	(SE)	β	(SE)	β	(SE)	β	(SE)	β	(SE)
Expressivity Moderator
Intercept	4.18**	(.04)	2.38**	(.06)	3.53**	(.06)	2.24**	(.06)	6.77**	(.14)	4.03**	(.13)
Family Income	.00	(.00)	−.00	(.00)	.00	(.00)	.00	(.00)	.00	(.00)	−.00	(.00)
Discrimination	−.04	(.09)	.27*	(.14)	.24*	(.12)	.22**	(.07)	.21	(.26)	.48**	(.15)
Gender	−.10†	(.06)	−.20*	(.09)	.47**	(.07)	−.05	(.07)	.52**	(.16)	−.47**	(.15)
Expressivity	.28**	(.04)	−.04	(.07)	.31**	(.06)	−.11†	(.06)	.50**	(.13)	−.22†	(.12)
Discrimination × Gender	.26*	(.11)	−.15	(.18)	−.23	(.15)	--	--	−.53t	(.32)	--	--
Discrimination × Expressivity	--	--	−.20*	(.10)	.23**	(.09)	--	--	.31t	(.19)	--	--
Instrumentality Moderator
Intercept	4.26**	(.04)	2.38**	(.07)	3.64**	(.07)	2.21**	(.06)	6.89**	(.14)	4.04**	(.13)
Family Income	−.00	(.00)	−.00	(.00)	−.00	(.00)	.00	(.00)	−.00	(.00)	−.00	(.00)
Discrimination	−.03	(.09)	.17	(.13)	.27*	(.13)	.09	(.12)	.33	(.26)	.21	(.25)
Gender	−.25	(.06)	−.20	(.09)	.27**	(.08)	.01	(.07)	.19	(.15)	−.42**	(.15)
Instrumentality	.08	(.06)	−.04	(.09)	.23**	(.09)	−.04	(.08)	.11	(.17)	.21	(.16)
Discrimination × Gender	.33	(.12)	−.15	(.18)	−.16	(.16)	--	--	−.46	(.34)	.13	(.33)
Discrimination × Instrumentality	.05	(.11)	−.19	(.06)	.14	(.14)	--	--	.26	(.23)	−.33	(.22)
Discrimination × Gender × Instr.	−.26†	(.15)	.50*	(.23)	−.45*	(.21)	--	--	−.91*	(.44)	.74†	(.43)

Note. All non-significant two and three-way interactions are omitted

^†p < .10,

^*p < .05,

^**p < .01.

Figure 1. The moderating role of expressivity for parents’ discrimination experiences (N = 312).

Panel A. Parents’ discrimination experiences × expressive traits predicting parent-adolescent conflict.

Panel B. Parents’ discrimination experiences × expressive traits predicting coparenting cooperation.

Panel C. Parents’ discrimination experiences × expressive traits predicting marital satisfaction.

We next tested hypothesis three to assess the moderating role of instrumentality. Analyses revealed three significant three-way interactions and two trend-level interactions between instrumental traits, experiences of discrimination, and parent gender (Table 2). In a first step, we followed-up the three-way interactions by testing each Discrimination × Instrumentality interaction separately for mothers and fathers. These analyses revealed that instrumentality moderated the effects of discrimination experiences for fathers only. In the case of marital conflict, interactions between discrimination and instrumentality were not significant for mothers, β = −.16, SE = .34, t = −.46, ns, or for fathers, β = .34, SE = .29, t = 1.16, ns, so we did not perform further follow-up for this analysis.

Follow-ups showed that, for fathers who were more instrumental, the link between discrimination experiences and father-adolescent conflict was positive, β = .25, SE = .12, t = 2.09, p < .05, d = .38; this effect was non-significant for low instrumental fathers, β = −.22, SE = .23, t = −.99, ns, (Figure 2, Panel A). In the absence of discrimination, however, having more instrumental qualities was related to less conflict. In the case of father-adolescent warmth, an interaction between discrimination and instrumentality reached trend level; so, given our directional hypothesis, we conducted a follow-up test. This showed that, for less instrumental fathers, the link between discrimination and father-adolescent warmth was positive, β = .47, SE = .15, t = 3.18, p < .01, d = .51 but non-significant for more instrumental fathers, β = .14, SE = .08, t = 1.29, ns, (Figure 2, Panel B). As with father-adolescent conflict, in the absence of discrimination experiences, more instrumental fathers had warmer adolescent relationships. Follow-ups revealed similar patterns for father’s marital relationships. For fathers who were less instrumental, the link between discrimination and coparenting cooperation was positive, β = .36, SE = .18, t = 2.01, p < .05, d = .33, but this effect was not significant for more instrumental fathers, β = −.07, SE = .10, t = −.75, ns, (Figure 2, Panel C). For fathers who never experienced discrimination, however, more instrumental fathers had greater coparenting cooperation. For more instrumental fathers, the link between discrimination and marital satisfaction was negative, β = −.55, SE = .22, t = −2.47, p < .01, d = .40 but non-significant for less instrumental fathers, β = .28, SE = .42, t = .67, ns, (Figure 2, Panel D). In the absence of discrimination, more instrumental fathers had greater marital satisfaction. In sum, these findings suggest that instrumentality may be protective under non-stressful circumstances but detrimental for men’s relationships when they experience stressful events.

Figure 2. The moderating role of fathers’ instrumentality for fathers’ discrimination experiences (N = 312).

Panel A. Fathers’ discrimination experiences × instrumental traits predicting father-adolescent conflict.

Panel B. Fathers’ discrimination experiences × instrumental traits predicting father-adolescent warmth.

Panel C. Fathers’ discrimination experiences × instrumental traits predicting coparent cooperation.

Panel D. Fathers’ discrimination experiences × instrumental traits predicting marital satisfaction.

Exploring Alternative Hypotheses

Given the literature on androgyny and individual well-being, we conducted analyses to explore whether androgyny played a role in the links between mothers’ and fathers’ discrimination experiences and family relationships. Toward this end, we tested whether androgyny moderated discrimination–family relationship links. Consistent with the idea that androgyny is linked to positive well-being, analyses revealed a positive relationship between androgyny and parental warmth, β = .13, SE = .05, t = 2.45, p < .05, d = .39. However, we found no evidence that androgyny moderated discrimination–relationship linkages.

Discussion

A body of work has established that individuals who experience racial discrimination are at risk for poorer psychological well-being. The present study built on this literature by examining whether racial discrimination was linked to family relationship qualities of African American mothers and fathers, and whether sex or gender-typed traits moderated these links. Although discrimination occurred at a low frequency for the sample as a whole, results were consistent with our expectations: (a) Fathers reported more frequent discrimination experiences than mothers, and (b) discrimination was associated with qualities of parent-adolescent, coparent, and marital relationships. Differences between mothers and fathers were evident in some discrimination–relationship quality links; however, these differences were qualified by parents’ expressive and fathers’ instrumental traits. Specifically, gender-typed traits moderated the links between discrimination experiences and family relationship qualities for both parents such that more expressive mothers and fathers generally reported more positive family relationships in the context of discrimination. In addition, more instrumental fathers reported more negative family relationships in the face of discrimination, although instrumentality was associated with positive family relationships for fathers who did not report discrimination. Taken together, our findings imply that there may be more variability for fathers than mothers in the ways that discrimination experiences are associated with their family relationships. At the most general level, our findings -- that discrimination was linked to family relationships in a sample of middle-class, well-functioning African American couples who reported low rates of discrimination -- attest to the significance of discrimination as a socio-cultural stressor for African American families. In the following pages, we expand upon these conclusions, highlighting the differences between mothers and fathers in discrimination–family relationship associations and the distinct patterns of linkages between discrimination and dimensions of family relationships.

Results related to our first hypothesis were somewhat consistent with our expectations about sex differences in stress responses. The positive association between mothers’ discrimination experiences and coparent cooperation suggests that mothers may perceive more family support in the face of a stressor, but this is the only “tend-and-befriend” response that emerged for mothers. In general, there were fewer links between discrimination and mothers’ (N = 1) as compared to fathers’ (N = 5) family relationship qualities. Consistent with the idea that mothers’ relationships in the family may be more scripted than fathers’ (Crouter et al., 1999), it is possible that their extra-familial experiences are less likely to come into play in their family relationships. In contrast, fathers’ family experiences may be more variable, and links between contextual stressors and family relationships, more evident.

For fathers, results were generally consistent with our expectation that discrimination would have negative implications for their relationships, but also revealed that discrimination was positively related to father-adolescent warmth. This pattern suggests that when fathers experience discrimination they may exhibit more affection and support, that is, “tend-and-befriend” in relationships with their offspring. This is consistent with the racial socialization literature (e.g., Hughes et al., 2003; McHale et al., 2006), which finds that (a) fathers are key socializers of discrimination in African American families, and (b) parents who experience discrimination are more likely to prepare offspring for racial biases, compared to parents who do not encounter discrimination. Importantly, these experiences may foster closer bonds between fathers and youth. Future work may reveal if this type of response is specific to African American fathers’ discrimination experiences.

In line with our second hypothesis, expressivity moderated discrimination–parent-adolescent conflict. Moreover, when expressive traits were accounted for, differences between mothers and fathers in links between discrimination and coparent cooperation and marital satisfaction became non-significant. This finding highlights the idea that sex alone may not explain the association between discrimination experiences and relationship qualities. Consistent with the expressivity hypothesis (Ickes, 1993), in the face of discrimination, parents with more expressive traits reported more positive relationship qualities, whereas parents with fewer expressive traits reported fewer positive family relationship qualities. These findings draw attention to the potential protective function of expressivity for women’s and men’s family relationships in the face of discrimination. Findings related to expressivity also have implications for family interventions in that mothers and fathers who draw on expressive behaviors in the face of a socio-cultural stressor may experience more positive family relationships.

When stereotypically masculine, instrumental, traits were examined as moderators, links between discrimination and family relationships emerged for fathers only. Results supported our third hypothesis, that in the face of discrimination, fathers with fewer instrumental traits reported more positive qualities in parent and coparent relationships, whereas those who were more instrumental had poorer family relationships. Importantly, for men who never experienced discrimination, being more instrumental was related to more positive relationship qualities, suggesting that effects of instrumentality may differ by context for men. Instrumental traits, which include attributes such as “self-reliant,” “aggressive,” and “defends beliefs”, may characterize men with problem-focused approaches to handling stressors, despite the fact that discrimination is a type of stressor that is usually beyond the individual’s control. Thus, fathers with more of these characteristics may engage in strategies aimed at eliminating discrimination, only to face increased frustration and negative arousal that spills over into interactions with their family members. Another possibility is that more instrumental fathers depend less on informal support networks when faced with discrimination. In contrast, less instrumental fathers may not resort to “fighting” discrimination. Instrumentality was not a factor in the links between mothers’ discrimination experiences and family relationship quality. For mothers, who were less instrumental than fathers on average and who have more scripted family roles, these types of traits do not appear to have implications for family relationships in the face of discrimination. These findings may be useful to practitioners in drawing attention to the susceptibility of fathers’ relationships, and highlighting the utility of tailoring interventions to individual characteristics.

Overall, our findings revealed that the negative effects of discrimination were most consistent for conflict in couple relationships: There was no evidence that sex differences or gendered traits moderated associations between discrimination and either coparent or marital conflict. It is important to note that levels of couple conflict in this sample were low. Thus, variability in coparent and marital conflict was limited, which may explain why moderators did not emerge for these links. Alternatively, it may be that the stress of discrimination experiences spills over into interactions between parents, and the nature of parents’ gendered traits are less effective buffers for marital relationships. In contrast, the implications of discrimination for positive family relationships and for parent-youth relationships, in general, were more complex: Results differed for mothers versus fathers and also depended on parents’ gendered qualities.

Like prior research that has found links between reports of discrimination and self-reported mental health problems (e.g., Fischer & Shaw, 1999), a limitation of the present study was the reliance on self-reports. Furthermore, our correlational design did not allow for inferences of causality. Examining these links longitudinally, rather than cross-sectionally, would provide insight into the sequence of events and could prove useful in ruling out some alternative explanations for the findings. It is also of note that 21% of fathers were not biologically related to youth, though all parents lived together for 16 years, on average. Finally, our small convenience sample of two-parent families limits the generalizability of our results.

Despite these limitations, the present study extends existing research in several important ways. First, much of the literature on African American families focuses on mothers. Our findings not only highlight African American fathers’ family experiences, but also draw attention to differences between mothers’ and fathers’ family roles and relationships when they experience racial discrimination. In addition, whereas prior research on gender and stress has focused on biological sex differences, this study examined differences between mothers and fathers in addition to within-sex variability to shed light on the implications of gender-typed attributes for links between discrimination and family relationship quality. Overall, the pattern of results suggests that, for mothers and fathers, expressivity is important for positive family interactions, and for fathers, having high levels of instrumentality may be detrimental for family relationships in the face of a significant stressor for African American individuals in the US: racial discrimination.