Shades of Black: Gendered Denominational Variation in Depression Symptoms Among Black Christians

Abstract

Religion and spirituality (R/S) play a central role in shaping the contextual experiences of many Black people in the United States. Blacks are among the most religiously engaged groups in the country. Levels and types of religious engagement, however, can vary by subcategories such as gender or denominational affiliation. Although R/S involvement has been linked to improved mental health outcomes for Black people in general, it is unclear whether these benefits extend to all Black people who claim R/S affiliation irrespective of denomination and gender. Data from the National Survey of American Life (NSAL) sought to determine whether there are differences in the odds of reporting elevated depressive symptomology among African American and Black Caribbean Christian adults across denominational affiliation and gender. Initial logistic regression analysis found similar odds of elevated depressive symptoms across gender and denominational affiliation, but further analysis revealed the presence of a denomination by gender interaction. Specifically, there was a significantly larger gender gap in the odds of reporting elevated depression symptoms for Methodists than for Baptists and Catholics. In addition, Presbyterian women had lower odds of reporting elevated symptoms than Methodist women. This study’s findings highlight the importance of examining denominational disparities among Black Christians, and suggest that denomination and gender may work in tandem to shape the R/S experiences and mental health outcomes of Black people in the United States.

For the Black community in the United States, religion and spirituality (R/S) typically plays an important and unique role in shaping contextual experiences. Black people are more likely to believe in God, attend church, and report that religion has a higher salience in their lives than other groups in the U.S. (Chatters, Taylor, Bullard, et al., 2008; Chatters et al., 1999; Hudson et al., 2015), and Black women tend to be more involved than Black men on nearly every R/S measure (Taylor & Chatters, 2010). Although much of the literature has linked religious involvement to improved mental health outcomes (Koenig, 2012), it is unclear whether greater engagement in religious activities is associated with better mental health outcomes in Black Christian adults and whether these effects are uniform.

Denominational affiliation and gender represent two demographic classifications that provide opportunities for divergent religious experiences. Denominations differ in the gender norms and social behaviors expected of men and women (Maselko & Kubzansky, 2006). They often are socialized and expected to engage in R/S activities differently, which may influence how religious involvement impacts their mental health (McFarland, 2009). Maselko and Kubzansky (2006) found that gender differences in the association between R/S and mental health varied by denomination. However, due to issues with the Black sample in that study, it is unclear whether the gendered denominational findings extend to Black Christians.

Prior research has examined the broad influence of R/S on the health of Black adults, but most do not distinguish, in a nuanced way, how denomination and gender also may influence outcomes. Researchers frequently cluster all Black communities of the Christian faith into a single entity (Steensland et al., 2000; Sternthal et al., 2010; Sullivan, 2010) and refer to the group generically with names like “Black Protestants.” This broad classification assumes there is an insignificant variation between denominational cultures and often ignores the heterogeneity of this group. There is presently a dearth of literature examining the existence of mental health disparities among the diverse community of Black Christians. This study addresses this gap by exploring the moderating influence of gender on the relationship between denomination and prevalence of elevated depression symptoms in Black adults.

Depression, Cardiovascular Disease, and Black Adults

There have been inconsistent findings regarding the prevalence of depression within racial groups in the United States. A recent study of older adults found that Blacks have higher rates of being at-risk for major depressive disorder (MDD) than Whites (Rodriquez et al., 2018). Another determined that Blacks have lower rates of depression than Whites (Williams et al., 2007), concluding that although Blacks experience depression less frequently, their bouts of depression were more likely to be disabling, severe, and to go untreated. These conflicting reports about the prevalence of depression in the Black community are likely a result of systemic and cultural flaws in symptom recognition, diagnosis, and treatment (Bryant et al., 2014; Hunn & Craig, 2009; Wharton et al., 2018).

Although there are challenges in accurately assessing depression rates for Black people, some have noted the potential risk depression poses to their long-term health (Lewis et al., 2011; O’Brien et al., 2015; Sims et al., 2015). Weinstein et al. (2011) found that experiencing more depressive symptoms was related to the presence of cardiovascular disease (CVD) risk factors such as larger waist-to-hip ratios and more body fat for African Americans. Sims et al. (2015) determined that Blacks who reported elevated depressive symptoms had a higher risk of experiencing coronary heart disease during a long-term follow-up. In addition, there is a longitudinal association between CVD risk and mortality in Blacks in the U.S. (Lewis et al., 2011; O’Brien et al., 2015) The evidence makes it clear that Blacks who experience higher depressive symptomology are at risk for worse health outcomes than those who are less depressed. The association between depression and CVD—one of the leading causes of death in the Black community—compels a deeper investigation into contextual factors that contribute to our understanding of the experiences and health outcomes of Black adults living with depression.

Past studies have identified R/S as potential protective factors against depression, CVD, and other health risks. Research suggests that social support (Chatters et al., 2015; Ellison et al., 2017), an optimistic worldview (Koenig & Larson, 2001), positive coping (Assari, 2014), and lower smoking and alcohol consumption (Aldwin et al., 2014) may serve as pathways through which R/S supports healthier outcomes. Also, Black Americans often employ religious resources as a strategy for coping with the racial and life stressors that are linked to negative CVD outcomes (Chatters et al., 2011; Ellison et al., 2017; Mattis & Grayman-Simpson, 2013). Given the prominence of R/S within the Black community, religious contexts are potential sites for collective depression reduction among Black adults. Additional evidence is needed to determine whether the mental health benefits of R/S extend to all Black adults who are religious.

Religion and Mental Health

Much of the research on the association between R/S and health outcomes finds that religious people tend to be happier (Koenig & Larson, 2001), healthier, and live longer (Chida et al., 2009). However, the mental health benefits associated with R/S do not extend uniformly to all people, as several scholars determined that those with the highest levels of religiosity experience the greatest benefit (McFarland, 2009; Mochon et al., 2011). Researchers also have found that those who attended church exactly once per week had fewer depressive symptoms than those who attended less than once per month or never (Sternthal et al., 2010). Yet, they concluded that attending church more frequently than once per week was unrelated to the number of depressive symptoms. Relatedly, other research finds that religious involvement and depression are unrelated (Schnittker, 2001).

R/S activities also have been associated with greater levels of depression. Sternthal et al. (2010) concluded that praying privately with greater frequency was linked to more depressive symptomology. Baetz et al. (2004) found that those who believed in spiritual values were most salient and those who viewed themselves as very religious were more likely to be depressed. Some have argued that these findings are likely due to people using R/S to manage their difficulties, rather than spiritual practices or beliefs negatively effecting mental health (Salmoirago-Blotcher et al., 2013). Overall, past studies emphasize the complex and undetermined connection between R/S and depression among religious people in the U.S.

Research on the association between R/S and mental health in Black Americans has yielded similarly mixed findings. One article concluded that attending church more frequently was associated with lower odds of elevated depression symptoms for Blacks (Jagers et al., 2007). Yet, several authors found a curvilinear relationship between church attendance and depression (Taylor et al., 2012, 2013). These studies indicate that regular, but not too frequent, church attendance may be associated with improved mental health for Black Americans.

Other research suggests that R/S is related to poorer health or has no impact on Black health outcomes. More frequent prayer (Jagers et al., 2007), listening to religious radio (Taylor et al., 2013), and reading religious materials (Taylor et al., 2012) have been linked to worse mental and physical health outcomes for Black people. Attending church, subjective religiosity, and various forms of nonorganizational religious participation were unrelated to the odds of a depression diagnosis for African Americans and Caribbean Blacks (Hudson et al., 2015; Taylor et al., 2013). The inconsistencies within the literature create ambiguity about which subcultures of Black people benefit from R/S activities and support the need to examine, with greater precision, who receives mental health benefits from religion.

Denomination, Gender, and Mental Health

Within churches, many of the traditions and teachings have a history of viewing the roles of men and women differently, and typically, place men in positions of power and leadership (Maselko & Kubzansky, 2006; Mattis & Grayman-Simpson, 2013). Since rigidity of gender norms and the associated social behaviors vary by denomination (Maselko & Kubzansky, 2006), gender differences in the impact of R/S on mental health outcomes also may be distinct. This effect could be amplified for groups in which men and women engage in R/S activities in markedly different ways.

Previous researchers have acknowledged gender differences in R/S participation and identification. Among Christians in the U.S., women tend to attend church more frequently and typically report higher scores on R/S measures than men (Jones et al., 2011; Maselko & Kubzansky, 2006; Pew Research Center, 2014). Researchers also have found gender differences in the types of R/S activities in which Christians participate (Jones et al., 2011; Taylor et al., 2009). Women appear to be more active religiously, though it is uncertain whether the differences in R/S activities and experiences yield dissimilar mental health outcomes (Maselko & Kubzansky, 2006; McFarland, 2009).

There is less certainty about gender differences in the potential mental health benefits of R/S among Black Americans. Blacks are among the most religious groups in the U.S. and Black women score higher than Black men on nearly all religious involvement measures, including religious salience, private prayer, service attendance, church membership, and reading religious materials (Pew Research Center, 2014; Taylor & Chatters, 2010). Adding to the gender differences in engagement, Taylor et al. (2009) found that although Black women reported higher religious participation, Black men spent more time engaged in other activities at churches. Due to their level of involvement in church support structures, Black women are more likely to use religious resources to cope with psychological and other serious problems (Chatters, Taylor, Jackson, et al., 2008; Chatters et al., 2011). Which means Black women’s ability to cope with depression could be more directly influenced by variations in churches’ ability to provide support. Thus, differences in gender role expectations (Mattis & Grayman-Simpson, 2013), R/S engagement (Taylor & Chatters, 2010), and coping (Chatters et al., 2011) necessitate further inquiry into whether R/S is similarly protective for Black women and men.

Quantitative research has considered how gender differences in mental health can vary by denomination, though this work has been done using a mainly White sample of Catholics, and Evangelical, and Mainline Protestants (Maselko & Kubzansky, 2006). Maselko and Kubzansky (2006) identified gender differences in the relationship between R/S and psychological well-being, but found that these disparities only existed among Catholics. That study included separate gender analyses to establish differences by denomination, but excluded Black Protestants due to a small cell size (Maselko & Kubzansky, 2006). The findings are informative about gender differences within the broader group of Christians in the U.S., but are not generalizable to Black Christians.

Differences in mental health outcomes could vary between and within denominations as parishioners participate to a greater or lesser degree in mental health supporting and coping behaviors (Chatters et al., 2011; Lukachko et al., 2015; Park et al., 2018). For instance, Black Pentecostals and women are more likely to rely on religious coping to deal with stressful life issues (Chatters, Taylor, Jackson, et al., 2008; Chatters et al., 2011). Therefore, these groups might be particularly likely to use religious coping to manage stressors and depressive symptoms. In addition, religious activities and physical and mental health behaviors may vary due to gendered denominational expectations, which means that men and women from the same or across denominations could experience disparate health outcomes. However, among Black Christian adults, this potential interaction between denomination and gender has not been adequately examined regarding its mental health effects.

Depression and Social Support Within Churches

Rather than attributing mental health benefits to religious activities, several researchers have stated that social support is a potential mechanism for improving mental health outcomes among Black churchgoers (Assari, 2013; Chatters et al., 2015). For some, establishing social support networks with other church members may substitute for professional help-seeking and may provide protection against depressive symptoms (Chatters et al., 2011, 2015; Hankerson et al., 2011).

It should be noted that sometimes churchgoers do not receive support in this context. At religious services attended by the majority of Black churchgoers, depressive symptomology runs counter to behaviors that are expected in church settings. Actions such as singing, dancing, and expressing joy might be difficult for a person who is experiencing depression. People who attend church while feeling depressed may encounter stigmatizing behaviors or beliefs from fellow congregants or clergy members, which might discourage them from acknowledging their depression or seeking help (Baetz et al., 2004; Bryant et al., 2014; Chatters et al., 2015).

Also, there might be denominational differences in the behavioral expectations for women and men. For example, members of the Black Pentecostal churches may expect men and women to express praise and worship with similar enthusiasm (Casselberry, 2017), whereas other denominations might encourage men and women to be differently vocal. In some churches, a lack of enthusiasm may be viewed as a signal for intervention, whereas others might perceive this person as ungrateful or lacking faith (Payne, 2008). Discord between depressive symptomology and expected worship activities, as well as the offering of social support in the event of this sort of conflict, likely varies by church and possibly by denomination.

In addition, clergy have a significant role in communicating and reinforcing church attitudes regarding mental health (Allen et al., 2009). They provide messages about culturally acceptable ways of dealing with depression and might actively discourage churchgoers from acknowledging depression symptoms or seeking help (Bryant et al., 2014). Payne (2008) examined the ways in which African American Pentecostal preachers spoke about depression during their sermons. Many preachers made remarks that discouraged parishioners from feeling depressed or seeking assistance from psychiatrists or psychotropic medications. Although acknowledging that this sample of preachers was from the same denomination, the author suggests that many of their messages were shaped by denominational doctrine (Payne, 2008).

Clergy can play an important part in the lives of Black people who seek informal support and often provide guidance for those who encounter barriers to formal help seeking (Allen et al., 2009; Chatters et al., 2011). Their messages can encourage parishioners to treat depressive symptoms. Some preachers openly endorse help-seeking or serve as gatekeepers who partner with mental health and social service providers to ensure that churchgoers have access to assistance (Hankerson et al., 2013; Wharton et al., 2018). Ministers who use their position to address mental health issues proactively within their church may help to reduce stigma, while encouraging parishioners to treat depression sooner, thereby reducing the likelihood long-term adverse health impacts (Anthony et al., 2015). Therefore, churches and denominations that provide a supportive environment can work to reduce or increase depression symptoms for Black people of faith.

Current Study

The present study examines the Christian denominational and gender differences in the prevalence of elevated depressive symptoms in Black adults in the United States. This is an important area of inquiry given the higher levels of Black religiosity and religious coping, gender differences in depressive symptomology among Black people, and the diverse subcultures within the Black community. Although previous research includes adequate assessments of denominational and gendered denominational differences in mental health (Maselko & Kubzansky, 2006; Park et al., 1990), this has not sufficiently been examined within Black Christians. Thus, there has been limited discourse in the literature about how denominational and gender group membership may impact Black mental health outcomes.

Prior studies on the influence of gender on mental health outcomes for Black churchgoers primarily have been qualitative and often focus on men and women separately, making limited comparisons within or between groups (Bryant et al., 2014; Hunn & Craig, 2009; Payne, 2008). Some work explores gender differences in depression among Black adults, but does not account for denominational variation (Wharton et al., 2018). This study addresses these gaps in the literature by using data from the National Survey of American Life (NSAL) to measure whether the risk of elevated depressive symptoms differs by Christian denominational affiliation and if this relationship varies by gender for Black adults.

Methods

Participants

The Institutional Review Board (IRB) of Duke University approved this study. The present study includes an analysis of secondary self-reported data collected as part of the NSAL (Jackson et al., 2004). The NSAL is an epidemiological examination of the mental health of Black people and those who live in neighborhoods with Blacks in the United States. The nationally representative, multistage probability samples of adults were interviewed about various topics, ranging from R/S beliefs and behaviors to mental and physical health background (Jackson et al., 2004).

The present article will focus only on NSAL participants who self-identified as Black [i.e., African American (n = 3,570) or Afro-Caribbean (n = 1,438)]. Caribbean Blacks who live in the U.S. were included in the current analysis because they are part of the Black diaspora and they share many social and religious similarities to African Americans (Chatters et al., 2009; Taylor & Chatters, 2010). African American and Afro-Caribbean participants’ respective demographic profiles were designed to mirror those of the national population with regards to characteristics such as gender, education, marital status, region, income, and urbanicity (Jackson et al., 2004). For additional information on the larger NSAL survey design and sample, see Jackson et al. (2004).

Measures

Dependent Variable

Depressive symptoms were measured using a short version of the Center for Epidemiological Studies-Depression scale (CES-D), which assesses the severity of depressive symptomology in people outside of clinical settings (Radloff, 1977). This 12-item version has been shown to be valid and reliable with Black adults, and has a similar factor structure as the original version (Foley et al., 2002; Torres, 2012).

The questionnaire asked how frequently participants experienced depressive symptoms, such as crying spells or a lack of focus during the past week. Response options ranged from 0 (rarely/never) to 3 (most of the time) and the alpha was equal to 0.76 for this scale. Positively worded items were reverse coded and responses to all 12 items were summed to create an index score with a maximum score of 36. These scores were used to determine whether participants reported experiencing elevated (i.e., clinically significant) levels of depressive symptoms.

Because the previously designated threshold of 16 for the original measure was inappropriate to use for the shorter version of the scale, the formula proposed by Kohout et al. (1993) was used to establish a new cut score for clinically significant depressive symptoms. The shorter version has 12 items, so the following calculations were made to select a score that is proportionate to meeting the recommended threshold of 16 on the original scale: 36/60 = 0.60 and 16 × 0.60 = 9.6, which was rounded to 10. This approach to generate a clinically meaningful cutoff score also was utilized in a study that used NSAL data to test the psychometric properties of this shorter CES-D in a Black sample (Torres, 2012).

To identify denominations with higher odds of reporting clinically significant levels depressive symptoms, the outcome variable was treated as dichotomous. Participants with scores greater than or equal to the new cutoff of 10 were categorized as having elevated depressive symptomology and those with scores less than 10 were classified as having depressive symptoms within a normal (i.e., clinically minimal) range.

Key Independent Variables

Religious denomination was measured by asking participants to indicate the current religion and denomination to which they are affiliated. NSAL investigators recoded over 40 different responses into 14 categories (Taylor & Chatters, 2010). Of these categories, this study reclassified most participants into one of six groups that were included in the current analysis: Baptist, Catholic, Methodist, Pentecostal, Presbyterian, and Protestant (other). All other participants were excluded from this analysis, as either they were unaffiliated or not religious, their group’s sample size was too small, or it was not possible to ascertain their specific religious affiliation due to the secondary nature of this analysis.

Participants were categorized according to their self-reported gender. This dichotomous variable was used to determine the odds of reporting elevated depressive symptoms for men and women, while ensuring that denominational effects are not due to potential gender differences in denominational affiliation or depression symptoms. Together, denomination and gender variables will be used to determine if gender differences in elevated depression symptoms vary by denomination. This is plausible, given the previous findings regarding denomination and gender differences in religious engagement, cultural norms, and doctrinal messages about depression, all of which may affect the likelihood and strategy for addressing depressive symptomology (Bryant et al., 2014; McFarland, 2009; Payne, 2008). An interaction between denomination and gender will be included to detect the moderating effect of gender.

Control Variables

The following demographic characteristics were used as control variables: age (in years), ethnicity (i.e., African American and Afro-Caribbean), education (i.e., 0–11, 12, 13–15, or greater than 15 years), income (in 10,000 increments with a maximum of 200,000), work status (i.e., employed, unemployed, and not in labor force), and region (i.e., Northeast, Midwest, South, and West) as each of these variables has been linked to depression (Chatters et al., 1999; Lincoln et al., 2010). The analysis accounts for the effects of being Afro-Caribbean which controls for an overrepresentation among the Catholic denomination (Chatters et al., 2009) and potential differences in depression levels (Williams et al., 2007). Subjective health was used as an additional control due to its potential to affect depressive symptoms (Assari, 2014). Health was measured using a self-reported single-item rating of present overall physical health on a 5-point scale ranging from excellent to poor.

The R/S participation variables of church attendance and nonorganizational religious involvement also were used as controls. Participants were asked how frequently they typically attend religious services and the six response options ranged from never to nearly every day. Those who never attend services were combined with those who only attend approximately once per year, reducing responses to five categories. Moreover, participants were asked five questions about how often they participate in nonorganizational religious activities such as reading religious text or praying. The scores of these 6-point Likert-type items were combined to create a scale with an alpha of 0.68 and scores ranging from 0 to 25. Frequencies of church attendance and nonorganizational religious activities have both been studied in the past to assess their relationship to depression (Schnittker, 2001; Sternthal et al., 2010; Taylor et al., 2013).

Statistical Analysis

Due to NSAL’s multistage sampling design, the analysis was conducted using subpopulation survey commands in Stata 16.1. Primary analyses consisted of three logistic regression models. The initial model included denomination as a predictor for the odds of having elevated depressive symptoms, adjusting for control variables. The second analysis added gender to the previous model. The final model included an interaction between denomination and gender. Whenever applicable, Methodist and women were designated as reference groups. Tables include descriptive statistics, odds ratios, and 95% confidence intervals (CI). Odds ratios are not considered significant if 1.00 is present in the respective CI range.

Results

The final sample size for the models included 4,116 participants, after excluding some respondents due to missing data. Ages ranged 18–94, with an average age of 43.45. Women outnumbered men (66% vs. 34%), and Baptist (48.6%) was the largest denomination. Nearly 25% of the sample reported elevated depressive symptoms (i.e., CES-D score of 10 or greater). The distribution of demographic characteristics for the sample is shown in Table 1. Logistic regression analyses and odds ratios for independent and control variables are shown in Table 2.

Table 1.

Sociodemographic Characteristics for Total Sample, Women, and Men

Sociodemographic variable	Total (n = 4,334), n (%)	Women (n = 2,852), n (%)	Men (n = 1,492), n (%)
Denomination
Methodist	276 (6.35)	176 (6.17)	100 (6.70)
Baptist	2,119 (48.78)	1,424 (49.93)	695 (46.58)
Catholic	505 (11.63)	318 (11.15)	187 (12.53)
Pentecostal	282 (6.49)	215 (7.54)	67 (4.49)
Presbyterian	41 (0.94)	31 (1.09)	10 (0.67)
Protestant (other)	1,121 (25.81)	688 (24.12)	433 (29.02)
Education years
0–11	1,026 (23.62)	674 (23.63)	352 (23.59)
12	1,548 (35.64)	1,000 (35.06)	548 (36.73)
13–15	1,054 (24.26)	697 (24.44)	357 (23.93)
16 or greater	716 (16.48)	481 (16.87)	235 (15.75)
Employment status
Employed	2,923 (67.43)	1,857 (65.20)	1,066 (71.69)
Unemployed	415 (9.57)	300 (10.53)	115 (7.73)
Not in labor force	997 (23.00)	691 (24.26)	306 (20.58)
Region
Midwest	453 (11.00)	305 (11.28)	148 (10.48)
Northeast	1,103 (26.79)	735 (27.17)	368 (26.06)
South	2,353 (57.15)	1,535 (56.75)	818 (57.93)
West	208 (5.05)	130 (4.81)	78 (5.52)
Ethnicity
African American	3,134 (72.15)	2,077 (72.83)	1,057 (70.84)
Afro-Caribbean	1,210 (27.85)	775 (27.17)	435 (29.16)
Subjective health
Excellent	732 (16.85)	436 (15.29)	296 (19.84)
Very good	1,373 (31.61)	874 (30.65)	499 (33.45)
Good	1,236 (28.45)	847 (29.70)	389 (26.07)
Fair	684 (15.75)	487 (17.08)	197 (13.20)
Poor	180 (4.14)	122 (4.28)	58 (3.89)
Religious service attendance
Once/year or fewer	569 (13.10)	308 (10.80)	261 (17.49)
A few times/year	821 (18.90)	461 (16.16)	360 (24.13)
A few times/month	1,062 (24.45)	705 (24.72)	357 (23.93)
At least once/week	1,626 (37.43)	1,183 (41.48)	443 (29.69)
Nearly every day	266 (6.12)	195 (6.84)	71 (4.76)
Depressive symptoms
Normal range	3,083 (74.90)	1,971 (72.81)	1,118 (78.90)
Elevated	1,034 (25.10)	736 (27.19)	299 (21.10)
	M (SD)	M (SD)	M (SD)
Age	43.45 (16.16)	42.90 (15.90)	43.40 (16.10)
Household income	34,258.09 (29,957.20)	26,939.00 (28,428.10)	36,536.80 (33,845.90)
Nonorganizational religion	16.72 (5.31)	17.40 (4.90)	15.30 (5.70)

Note. SD = standard deviation; Elevated Depressive Symptoms: CES-D ≥ 10.

Table 2.

Logistic Regression Analysis Predicting Elevated Depression Symptoms

	Model 1		Model 2		Model 3
	OR	95% CI	OR	95% CI	OR	95% CI
Denominationa
Baptist	1.03	[0.68–1.56]	1.03	[0.67–1.56]	0.78	[0.49–1.24]
Catholic	1.10	[0.55–2.21]	1.12	[0.56–2.23]	0.70	[0.31–1.59]
Pentecostal	0.80	[0.49–1.31]	0.79	[0.48–1.30]	0.73	[0.46–1.18]
Presbyterian	0.37	[0.07–1.87]	0.37	[0.07–1.84]	0.14*	[0.03–0.64]
Protestant (other)	1.11	[0.71–1.72]	1.13	[0.73–1.74]	0.91	[0.57–1.46]
Education yearsb
0–11	1.53**	[1.16–2.02]	1.53**	[1.16–2.02]	1.51**	[1.14–2.00]
13–15	0.58***	[0.42–0.78]	0.57***	[0.42–0.77]	0.57***	[0.42–0.77]
16 or greater	0.61*	[0.41–0.91]	0.60*	[0.40–0.89]	0.60*	[0.40–0.89]
Employment statusc
Unemployed	2.01***	[1.39–2.91]	1.98***	[1.37–2.87]	2.01***	[1.39–2.90]
Not in labor force	1.38*	[1.07–1.78]	1.36*	[1.06–1.75]	1.36*	[1.06–1.75]
Regiond
Northeast	0.81	[0.60–1.08]	0.80	[0.60–1.06]	0.80	[0.61–1.06]
South	0.73*	[0.56–0.96]	0.74*	[0.56–0.97]	0.74*	[0.57–0.97]
West	0.82	[0.56–1.20]	0.82	[0.56–1.20]	0.84	[0.57–1.23]
Ethnicitye
Afro-Caribbean	0.88	[0.64–1.20]	0.89	[0.65–1.22]	0.89	[0.66–1.21]
Subjective healthf
Very good	1.04	[0.72–1.51]	1.03	[0.72–1.50]	1.05	[0.72–1.52]
Good	1.69**	[1.21–2.38]	1.66**	[1.17–2.34]	1.67**	[1.19–2.35]
Fair	3.19***	[2.30–4.42]	3.12***	[2.24–4.33]	3.13***	[2.25–4.36]
Poor	5.13***	[2.99–8.80]	5.08***	[2.96–8.72]	5.06***	[2.98–8.62]
Religious service attendanceg
A few times/year	0.93	[0.68–1.28]	0.94	[0.68–1.28]	0.92	[0.67–1.27]
1–3 times/month	0.78	[0.59–1.03]	0.77	[0.59–1.02]	0.77	[0.59–1.01]
1–3 times/week	0.71	[0.48–1.03]	0.69	[0.47–1.00]	0.68*	[0.47–0.99]
Nearly every day	0.81	[0.50–1.32]	0.79	[0.49–1.28]	0.78	[0.48–1.28]
Age	0.97***	[0.96–0.98]	0.97***	[0.96–0.98]	0.97***	[0.96–0.98]
Household income	0.94	[0.89–1.00]	0.95	[0.90–1.00]	0.95	[0.90–1.00]
Nonorganizational religion	1.04**	[1.01–1.06]	1.04**	[1.01–1.06]	1.04**	[1.01–1.06]
Genderh
Men			0.80	[0.64–1.00]	0.39**	[0.20–0.76]
Denomination × Gender
Baptist × Men					2.24*	[1.07–4.69]
Catholic × Men					3.27*	[1.22–8.72]
Pentecostal × Men					0.83	[0.29–2.40]
Presbyterian × Men					9.55	[0.38–240.60]
Protestant (other) × Men					1.89	[0.89–4.03]

Note. OR = odds ratio; CI = confidence interval. n = 4116 for all models.

^aMethodist.

^b12 years.

^cEmployed.

^dMidwest.

^eAfrican American.

^fExcellent Health.

^gOnce/year or fewer.

^hWomen.

^*p < .05.

^**p < .01.

^***p < .001.

Models 1 and 2: Main Effect of Denomination and Gender

According to the first two models, neither current denominational affiliation nor gender was associated with the odds of self-reporting elevated depressive symptoms, when accounting for the effects of the sociodemographic variables. Age, education, region, employment status, subjective health, and nonorganizational religious participation were all linked to differences in depressive symptoms in both models (see Table 2). Participants had lower odds of reporting elevated depression symptoms if they were older, more educated, from the South, employed, or more subjectively healthy when compared to participants in the respective reference groups. Those who scored lower on the measure of nonorganizational religious involvement also had lower odds. There were no significant differences in the odds of reporting elevated depression symptoms based on the frequency of religious service attendance, household income, or between African Americans and Black Caribbeans. Thus, there was no main effect for denomination or gender, as elevated depressive symptomology appeared to be similar across both genders and all denominations.

Model 3: Denomination and Gender Interaction

After including an interaction between denomination and gender to Model 2, the analysis suggests that the relationship between denomination and depression varies based on gender (see Table 2). After controlling for demographic variables and including the interaction effects, Presbyterian women [OR = 0.14, 95% CI = (0.03, 0.64)] had 86% lower odds of elevated depression symptoms than Methodist women.

Baptist men [OR = 2.24, 95% CI = (1.07, 4.69)] and Catholic men [OR = 3.27, 95% CI = (1.22, 8.72)] were more likely to report similar rates of elevated depressive symptomology as the women in their respective denominations than Methodist men. [Note: Due to the fact that the reference comparison of Methodist men to Methodist women was less than one (OR = 0.39), an interaction odds ratio of greater than one implies that the reference ratio is the smaller of the two ratios that are being compared.] This means that the gender gap in prevalence of elevated depression symptoms was approximately 2.24 and 3.27 times larger between Methodist men and women, than it was for Baptists and Catholics, respectively. The gender comparisons for each of the remaining denominational groups were not significantly different from the Methodists.

It is noteworthy that the CI for the interaction between Presbyterian and gender has a wide CI. This is likely due to the small sample size of this denomination; hence, there might not be enough power to detect the effect of the relationship. Those who attended religious services between one and three times per week had 32% lower odds than those who attended once per year or fewer. Excluding service attendance, the relationships between the demographic controls and elevated depressive symptoms remained consistent, as all categories with significant odds ratios in the previous models remained significant.

Discussion

The primary goal was to determine whether denominational differences in the prevalence of high depressive symptoms are moderated by gender for Black adults. Models 1 and 2 showed demographic dissimilarities, yet there were no significant denominational or gender differences in depression. This differs from the findings of Sternthal et al. (2010) who found that being Catholic was associated with fewer depressive symptoms than those with no religion. In the current study, being Catholic was unrelated to depressive symptoms.

In Sternthal’s study, researchers combined Protestants into a single category. By separating Protestants into distinct categories and testing the interaction between denomination and gender, analysis in Model 3 uncovered both gender and denominational differences that were not present in Models 1 and 2 or in the article by Sternthal et al. (2010). Testing the interaction showed divergent denominational patterns in the odds of reporting elevated depression symptoms based on gender and showed disparities within a group that often is treated as sharing a single, collective experience (Sternthal et al., 2010; Sullivan, 2010).

After controlling for relevant demographic characteristics, Presbyterian women had significantly lower odds of reporting elevated depressive symptoms than Methodist women. However, the gender differences in elevated depression symptoms between Presbyterian men and women compared similarly to that of the Methodists. In addition, although neither Baptist nor Catholic women were more likely to report elevated symptoms compared to Methodist women, Baptists and Catholics reported significantly smaller gender differences in prevalence rates than Methodists. These findings contrast with those of Maselko and Kubzansky (2006) who excluded Black Protestants but found that Catholic women and men differed in their psychological distress levels, whereas men and women from the Protestant denominations were similarly distressed.

The fact that the gender gaps for Baptists and Catholics were significantly smaller than for Methodists may signify that these groups make similarly gender-neutral distinctions about the accept-ability of experiencing depression. That is, these groups might communicate that these symptoms are universally suitable or unsuitable for men and women. However, these results do not imply that Baptist and Catholic men are similar in their levels of symptomology, only that their rates are more similar to the rates of women from their respective groups than are the rates of Methodist men and women.

It is noteworthy that Baptists and Catholics both were different from Methodists. In the past, Baptists and Methodists who are Black appeared to have no significant differences in subjective religiosity, church attendance, or nonorganizational religious participation (Chatters et al., 1999). Previous research comparing Black Baptists to Catholics has highlighted many group differences. Aside from the obvious distinctions in rituals and doctrine, they also differ in time spent participating in organizational and nonorganizational religious practices, as well as providing and receiving church-based social support (Chatters et al., 1999; Taylor et al., 2011). Regardless, detecting similar findings across groups that appear to be dissimilar on so many R/S measures is further evidence that there is more to learn about the mechanisms by which R/S influences mental health outcomes in individual denominations.

Although there were only significant gender differences among some denominations, the absence of differences for others does not necessarily mean that men and women in those groups experience the same level of depressive symptoms. Black men and women of faith each may experience distinct sociocultural pressures to mask their symptoms to avoid being perceived as weak, lacking faith, or unable to handle the demands associated with their assigned roles (Bryant et al., 2014; Hunn & Craig, 2009; Wharton et al., 2018). Both groups may be socialized to believe that unhealthy levels depressive symptoms are a normal part of life.

For Black women in particular, Walker-Barnes (2014) discussed how suppressing emotional pain and trauma may be seen as part of what has been called Strong Black Woman Syndrome. She posited that the appearance of overcoming emotional trauma is a demonstration of religious faith and sacrifice for Black women in the church (Walker-Barnes, 2014). Women who embrace this identity might avoid help-seeking and feel compelled to support others in difficult situations. Forgoing treatment might have a disproportionally greater impact on women due to their heavier involvement in church-based social support networks, which may place a greater emotional burden on them (Nguyen et al., 2016; Taylor et al., 2017). Participating in these networks may encourage Black women to overextend themselves, leaving few resources to assess their own feelings (Hunn & Craig, 2009). The internalization of denominational social pressures to hide depressive symptoms or display strength by eschewing assistance may have contributed to the present findings by increasing the likelihood that members of certain groups have more underreported or untreated depression.

Previous researchers have suggested possible rationales that might provide context for this study’s findings including the following: church stigma (Hankerson et al., 2013), gender roles (Wharton et al., 2018), misdiagnosis of symptoms (Bryant et al., 2014), and differential access to support (Payne, 2008; Wharton et al., 2018). Determining the particular combination of psychological and sociocultural factors that may have led to these results is beyond the scope of this secondary analysis. Nevertheless, the fact that men and women from three major denominations (i.e., Presbyterian, Baptist, and Catholic) compared differently to their respective reference groups highlights the importance of stratifying Black people when studying mental health, to better understand how gender roles and experiences may operate differently within this group.

Many demographic controls yielded similar results to those found in other studies. In accord with previous findings with Black samples, being older, having more years of education, being employed, and having higher subjective health were linked to lower odds of reporting elevated depressive symptoms (Assari et al., 2016; Lincoln et al., 2010; Williams et al., 2007). Household income and being Afro-Caribbean were unrelated to the likelihood of reporting elevated symptoms. This was surprising as Williams et al. (2007) used the NSAL data to find that African American and Afro-Caribbean people had comparable rates of 12-month MDD. However, they also found that the prevalence of lifetime MDD was higher for Black Caribbeans than for African Americans (Williams et al., 2007). Therefore, depression might have differed by ethnicity, had the symptoms been measured over a longer time period.

The relationship between church attendance and depression was not linear (Jagers et al., 2007), but also was not curvilinear (Taylor et al., 2012, 2013). The current analyses controlled for different demographic variables than Taylor’s articles, and found that people who attended services between one and three times per week were the only group who had lower odds of reporting elevated symptoms than those who attend once per year or fewer. This is an additional evidence that attending church is related to better mental health outcomes for those who attend regularly, but not too much. Relatedly, those who had higher scores on the nonorganizational religion measure had greater odds of elevated depressive symptoms. The association between high religious engagement and poorer mental health outcomes has been observed previously (Jagers et al., 2007; Taylor et al., 2012), and may be a sign that these participants were using these personal religious behaviors to cope with difficulties (Salmoirago-Blotcher et al., 2013).

Limitations and Future Directions

There are numerous strengths in this study, but there are also some limitations. Rather than using the assessment and diagnosis of trained professionals, self-reported data were used to measure depressive symptomology. Black people who are religious may be prone to underreporting depressive symptoms due to cultural stigma (Bryant et al., 2014; Hankerson et al., 2013). This study may have identified greater odds that people from some denominations underreport or refuse to seek assistance for their depressive symptoms, rather than determining that people from some denominations are more depressed. Determining whether the findings result from certain groups underreporting is beyond the scope of this study. However, detecting the presence of gendered denominational support for or stigma against reporting depression symptoms should be a goal of the future research. Interventionists should recognize that low depression rates in this group may be misleading and should not disqualify them from receiving mental health outreach (Bryant et al., 2014).

There also were limitations with the use of denominational categories. This study used secondary data and many of the participants’ denominational affiliations had been recategorized by the original researchers prior to sharing the data. Predetermined categorizations and small cell sizes also made it difficult to assess whether there were differences among participants from small Christian denominations and those who practice Islam, Buddhism, Judaism, or other non-Christian faiths. Also, denomination was only recorded at one time point, so the data did not account for whether participants were influenced by the past affiliations.

The present study focused on differences among Black Christian denominations, but the analyses could not account for differences in the experiences of those who attended predominately or historically Black churches and those who attended non-Black or diverse churches. Attending a predominately Black Baptist church likely creates a sociocultural experience that is qualitatively different from attending a Baptist church with a mostly White congregation (Steensland et al., 2000). Perhaps, additional church demographic information would have allowed for a more nuanced exploration of between-group differences in Black Christians’ experiences with depression. Despite these denominational limitations, this analysis uncovered depression disparities simply by acknowledging that this group is multifaceted.

Although there were no significant differences between African Americans and Caribbean Blacks, other projects have found that ethnicity is related to both religious denomination and depression prevalence (Assari, 2014; Chatters et al., 2009; Williams et al., 2007). Two people who similarly identify as Black but have different ethnicities can have very different contextual experiences. Future research should explore how ethnic and cultural differences as well as other social identities within the Black community contribute to depression disparities.

Conclusions

Those who believe that churches are essential in health promotion should view these results as a call to review denominational culture, doctrine, and social support structures to ensure that mental health benefits extend to members of all groups. Within the R/S context, gender and denomination classifications can demarcate social boundaries that dictate exposure to harmful or health-promoting socialization. Findings from this study highlight the importance of examining how denomination and gender may work in tandem to shape one’s experience in ways that impact the mental health of Black Christians.

Although it is methodologically convenient for researchers to cluster all Black religious people, this study suggests that members of some subcultures might have a higher likelihood of experiencing elevated depression symptoms. Disaggregating this group enabled the detection of disparities that may have been muted in prior analyses, revealing that the existence and magnitude of gender gaps in depression may differ by denomination. Further research is essential to determine which psychosocial mechanisms contributed to these observed differences.

The current findings can be applied when addressing depression in clinical and community settings. Most therapists likely recognize the salience of R/S as a guide or coping device in the lives of many of their Black Christian clients. However, it is important for therapists to know that unique denominational and gender roles may have a disparate effect on the depression symptomology of clients who seem similarly engaged in religion. Working toward a better understanding of how to mitigate these within-group depression disparities also will be imperative when planning church-based health outreach programming. Reducing depression symptoms with more effectively designed interventions not only will improve the day-to-day psychological well-being of people in the Black community but also may positively impact their long-term cardiovascular health outcomes.