Shades of Health: Skin Color, Ethnicity, and Mental Health among Black Americans

Abstract

Skin color is an important predictor of health outcomes among Black Americans. Black Americans with darker complexions experience worse physical and psychological functioning than those with lighter complexions. However, most research on the health effects of colorism focuses solely on African Americans, omitting the experiences of other Black subpopulations. Using data from the National Survey of American Life (NSAL), we investigate the relationship between skin color and mental health among African Americans (N=3,393) and Caribbean Blacks (N=1,378). Findings from multivariate logistic regressions reveal that Black Americans with the lightest complexions—regardless of ethnicity—report worse psychological functioning. However, the shape of the association between skin tone and mental health varies significantly based on ethnicity and the specific psychiatric outcome under study. For Caribbean Blacks, the association between skin color and any mental disorders and mood disorders is linear, while the relationship for anxiety disorders is curvilinear. For African Americans, the relationship between skin color and mental health shows an elevated risk among only those with the lightest skin tones. These results illustrate the heterogeneity within the Black community and highlight the importance of recognizing ethnicity in health disparities research.

Mental health is a growing concern—in 2019, 6.5 million Black Americans reported a mental health disorder, an increase of over 10% compared to the previous year. Furthermore, 23% of those who reported a disorder, described their condition as serious enough to limit daily activities or impact major life events (Substance Abuse and Mental Health Services Administration 2020). Although these estimates describe broad trends, there are important intraracial differences among Black Americans that are underexplored.

A limited number of studies investigate disparate patterns in mental health across Black subpopulations. These findings reveal key distinctions based on nativity, whereby foreign-born and native-born Black Americans experience differential risks of developing major depressive disorder, anxiety disorders, alcohol use disorder, and serious psychological distress (Gibbs et al. 2013; Mouzon and McLean 2017; Williams, González, et al. 2007; Williams, Haile, et al. 2007). Even fewer studies examine mental health differences among Black Americans based on skin color, despite mounting evidence that skin color is a significant predictor of both mental and physical health (Cobb et al. 2016; Hargrove 2018; Perreira, Wassink, and Harris 2019). Given the growing diversity of the Black population in the United States, research should critically and carefully examine mental health disparities across important subgroups rather than primarily focus on Black/White differences. To this end, the current study asks if and how the association between skin color and mental health differs between Caribbean Blacks (CBs) and African Americans (AAs) in the U.S.

Interethnic Differences in Black Mental Health

CBs represent 50% of the total U.S. Black immigrant population, with more than 2 million Black immigrants reporting Caribbean ancestry in 2019 alone (U.S. Census Bureau 2019). Research concerning differences in psychiatric outcomes among CBs and AAs is somewhat limited and has produced mixed findings (Gibbs et al. 2013; Mouzon and McLean 2017; Williams, González, et al. 2007; Williams, Haile, et al. 2007). CBs tend to report better overall levels of psychological functioning and lower rates of anxiety and substance use disorders (Gibbs et al. 2013; Himle et al. 2009; Mouzon and McLean 2017). However, AAs are less likely to experience mood disorders, such as major depressive episodes (Williams, González, et al. 2007). These interethnic differences have important implications for health disparities research — namely that exploring the heterogeneity within the Black community is an essential step towards developing a comprehensive understanding of health disparities in the U.S. Moreover, these complex set of findings suggest that the psychiatric risk profile for a given Black subpopulation is likely to depend on the specific outcome under study.

Notable gaps remain in our understanding of mental health among CBs. First, it is not clear if skin color influences the relationship between race/ethnicity and mental health for CBs. Evidence suggests that skin color is an important factor shaping mental health among AAs (Louie 2020; Monk 2015); however, to our knowledge, there are no empirical studies investigating the relationship between skin color and mental health among CBs. Second, we do not know how the relationship between skin color and mental health differs for AAs and CBs. Given the growing diversity of the U.S. Black population (Anderson 2015) and the recent rise in psychological disorders across the U.S. (Reinert, Fritze, and Nguyen 2021), it is important to understand how mental health varies across key Black subpopulations. The current study addresses these gaps in our empirical knowledge by examining the association between skin color and mental health independently for AAs and CBs. Further, we compare how this association differs between these two groups.

Skin Tone, Ethnicity, and Health Disparities

Colorism—social stratification based on skin tone—is a common global phenomenon tied to White supremacy. Unlike racism, which divides between racial groups, colorism is often used to create and perpetuate social divisions within racial groups. In the United States, colorism can be traced back to the transatlantic slave trade, thus having an outsized impact on populations that were forced into bondage. During slavery, Whites used colorism as a tool to create social distinctions and (re)enforce a social hierarchy among Black people, providing those with lighter complexions access to more social and economic resources than their counterparts with darker complexions (Horowitz 1973; Monk 2021). This preferential treatment placed lighter skinned Black Americans at an advantage, thus further solidifying intragroup inequalities. Even today, colorism remains a consequential predictor of several socioeconomic outcomes among Black Americans including education, employment, marriage, and income (Devaraj, Quigley, and Patel 2018; Hunter 2005). Despite the broader societal advantages that lighter complected AAs receive, research finds that AAs with the lightest and darkest complexions report the most intragroup stigmatization, while those with medium complexions report the least (Landor and McNeil Smith 2019).

Colorism is also a critical factor in the production and maintenance of health disparities. Darker skin tones are associated with an array of negative sequelae among nonwhites including hypertension (Sweet et al. 2007), allostatic load (Cobb et al. 2016), cardiometabolic functioning (Wassink, Perreira, and Harris 2016), and cumulative biological risk (Hargrove 2018). Most research efforts reveal a linear association between skin tone and health, whereby nonwhite individuals with darker complexions experience worse outcomes than those with lighter skin tones (Cobb et al. 2016; Hargrove 2018; Perreira et al. 2019; Sweet et al. 2007). A shortcoming of the literature on colorism and health is that much of this research is based solely on the experiences of AAs. Although a limited number of studies examine how health is differentially patterned according to skin tone among other racial/ethnic groups, this small body of research focuses on Latinos and Asians and, until now, has not considered CBs (Cuevas, Dawson, and Williams 2016; Dixon and Telles 2017; Perreira et al. 2019).

Investigating the connection between colorism and mental health among CBs can help us understand racialization processes among immigrants acculturating into the U.S. While colorism is a global phenomenon, the importance of race or skin color as the primary signifier of social status is context dependent. In racially heterogenous countries such as the U.S., race is a primary signifier of social status, and this primacy influences how colorism operates among AAs. Within racially heterogeneous contexts, skin color becomes less important for broader social stratification patterns because race not color is the primary source of stigma (Harvey et al. 2005). However, in more racially homogenous Caribbean countries such as Jamaica and Haiti, skin color serves as a key indicator of social status and resources (Hall 2008; Wilder 2015). Because the importance of skin color is context dependent, examining the relationship between skin color and outcomes among CBs can reveal information about how this group acculturates into American society and become “racialized others” (Viruell-Fuentes, Miranda, and Abdulrahim 2012). If the association between skin color and mental health is similar among CBs and AAs, this may indicate that CBs experience similar race-based stigmatization as AAs, prompting less emphasis on skin color as a primary signifier of social status.

Interethnic Differences between AAs and CBs

We identify three possible sources of interethnic variation between CBs and AAs that may drive the disparate association between skin tone and mental health across these Black subgroups. The first two types of interethnic differences suggest color-based stratification in mental health outcomes will be more pronounced among AAs than CBs, while the third source of interethnic differences predicts the opposite.

First, many CBs living in the U.S. are immigrants whose experiences and social exposures likely differ from those of native-born AAs. Prior research consistently finds that immigrants, especially those who came to the U.S. recently, fare better than non-immigrants across a range of health outcomes including infant birth weight (Buekens et al. 2000; Hamilton, Teitler, and Reichman 2011), all-cause and cause-specific mortality (Abraído-Lanza, Chao, and Flórez 2005; Hummer and Amir 2000), and psychiatric disorders such as substance use, depression, and anxiety (Escobar 1998; Ikonte et al. 2020; Williams, Haile, et al. 2007). Scholars typically attribute these differences to cultural and structural factors including strong social ties, high levels of social support, better health behaviors, and residing in immigrant enclaves which can protect them from health eroding race-related stressors (Abraído-Lanza et al. 1999, 2005; Viruell-Fuentes 2007; Viruell-Fuentes et al. 2012). However, the protective health effects of immigrant status tends to wane over time, with some studies showing a congruence of health outcomes among foreign- and native-born populations occurring in as little as 5 years (Antecol and Bedard 2006).

Second, CBs and AAs tend to have distinct social and cultural identities that emerge from critical differences in ethnicity, national origin, and lived experience (Agyemang 2005; Jones and Erving 2015; Waters 1999). Given the important shared meanings we attach to racial categories in the U.S. as well as the differential distribution of social and economic resources along racial lines, these distinctions are likely to have important implications for mental health. While members of both ethnic groups tend to identify as Black, this racial identity typically holds more salience for AAs than CBs, particularly among more recent immigrants, due to AAs’ lengthy history of race-based social exclusion and economic subordination in the United States (Dawson 1995; Wimmer 2008). Conversely, CBs often adopt either a regional ethnic identity (i.e. West Indian) or a national ethnic identity (i.e. Jamaican) as their primary group identification, further distinguishing themselves from AAs (Benson 2006; Rogers 2001; Vickerman 1999).

We expect that embracing a different or additional ethnic identity based on country or region of origin would protect individuals from the negative impacts of race-related stressors as well as colorism. Because being “Black” is a core, salient identity for AAs, stigmatization from other ingroup members may be particularly health harmful as it may be perceived as an attack on one’s key social identity (Basáñez et al. 2014; Uzogara and Jackson 2016). Lighter complected AAs report facing questions of authenticity and legitimacy related to their Blackness (i.e., being “Black enough”) (Hunter 2008; Landor and McNeil Smith 2019). Darker completed AAs often report being stigmatized by other Blacks as well for being “too dark” (Hunter 2005; Russell, Wilson, and Hall 2013). Both the immigrant health advantage and differences in how CBs and AAs self-identify suggest that the association between skin tone and mental health will be more pronounced for AAs compared to CBs.

Third, CBs migrate to the U.S. from countries with significantly different sociopolitical contexts. In many Caribbean countries, race and class are not intrinsically tied to one another to the same extent because Black people are the demographic majority (Vickerman 1999; Waters 1999). The disconnect between race and class in Caribbean countries is crucial because of the strong positive association between socioeconomic status and health (Elo 2009; Link and Phelan 1995). Research suggests that skin color plays a larger role than race in perpetuating inequalities in Caribbean countries (Kelly 2020). In nations as diverse as Jamaica and Trinidad, having lighter complexions is associated with more access to social and economic resources (Kelly 2020; Khan 2009; Wilder and Cain 2011). This stands in stark contrast to the United States, where access to social and economic resources is more closely tied to race than skin color (Williams, Priest, and Anderson 2016). These interethnic distinctions in the sociopolitical contexts of one’s country of origin suggests that the association between skin color and mental health will be more pronounced among CBs than among AAs.

Colorism and Racial Disparities in Mental Health

Despite the rapidly expanding body of evidence linking skin tone to racial disparities in physical health, far fewer studies extend this research focus to include mental health outcomes. Among those that do, results have been inconsistent and difficult to replicate (Borrell et al. 2006; Keith et al. 2010; Monk 2015; Perreira et al. 2019). Some investigators find no significant association between skin tone and overall psychological functioning or specific disorders such as depression (Borrell et al. 2006; Keith et al. 2010), while others show that individuals with darker skin tones experience worse overall psychological wellbeing and more depressive symptoms (Louie 2020; Perreira et al. 2019).

While most research linking skin tone to mental health finds that darker skin is associated with poorer outcomes, some findings suggest the relationship between skin color and mental health is more nuanced. Unlike previous studies, Monk (2015) finds that respondents with medium complexions report better self-rated mental health and fewer depressive symptoms than those with the darkest and lightest complexions. Given that prior research has revealed a consistent relationship between skin tone and physical health among Black Americans (Cobb et al. 2016; Hargrove 2018; Perreira et al. 2019; Sweet et al. 2007), the more complex patterning for mental health is somewhat surprising. One possible explanation is that the literature on colorism and health disparities tends to treat Black Americans as a single, monothilic group rather than further stratifying along other key social identities and divisions such as ethnicity or nativity (Borrell et al. 2006; Keith et al. 2010). In doing this, scholars run the risk of underestimating or obscuring intragroup differences because of the substantial heterogeneity within the Black community which has been shown to pattern health outcomes (Colen, Krueger, and Boettner 2018; Williams and Sternthal 2010). Acknowledging this heterogeneity, the current study highlights variation among Black Americans by investigating interethnic differences between two of the largest Black subpopulations in the U.S.

We use nationally representative data to address the following research questions pertaining to skin color and mental health among these two Black subpopulations:

1. Does the association between skin tone and mental health differ for AAs and CBs?

2. Is the relationship between skin tone and mental health more pronounced for AAs or CBs?

3. Does the shape of the gradient between skin tone and mental health follow similar patterning for AAs and CBs?

Data and Methods

Description of the Data

This study utilizes data from the National Survey of American Life (NSAL), a nationally representative survey of Black adults in the U.S. Data was collected via in-person interviews from 2000-2003. The NSAL contains oversamples of certain Black subpopulations, including immigrants. It is the first, and still one of the very few, nationally representative datasets to include a probability sample of CBs large enough to enable comparisons between them and AAs (Jackson et al. 2004). The NSAL features a sample of 3,570 AAs, 1,438 CBs, and 1,074 Whites (N=6,082). A total of 5,008 participants self-identified as either AA or CB. We excluded participants who had missing values on any of the key independent (ethnicity or skin tone), dependent (mood, anxiety, substance use, or any disorders), or covariate variables. These exclusion criteria resulted in a final analytic sample of 4,771 respondents, 3,393 of whom self-identified as AAs and 1,378 identified as CBs.

Description of the Measures

We use four dependent variables to capture mental health among NSAL respondents. We independently assess three specific indicators of life-time psychological functioning including mood, anxiety, and substance use disorders. We also include a measure of general mental health (any mental disorders), which captures the prevalence of experiencing any of the three aforementioned disorders. These outcomes were based on criteria from the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV), which is widely considered to be the “gold standard” for identifying psychiatric disorders based on underlying symptomatic expression. All outcome measures were coded dichotomously, with 1 indicating the presence of a psychological disorder and 0 indicating its absence.

The primary independent variables of interest were ethnicity, and skin color. Self-reported race was assessed using the following question: “Which do you feel best describes your racial background?” Respondents’ national origin was determined based on responses to the following question about their parents’ and grandparents’ nativity. We defined CB ethnicity to include Black respondents whose parents or grandparents originally immigrated from the Spanish Caribbean (i.e. Cuba, Puerto Rico), Haiti, Jamaica, and Trinidad and Tobago, among others. Black participants whose parents and grandparents were born in the U.S. were classified as AA. Ethnicity is coded as 0 for AAs and 1 for CBs.

Skin color was captured using answers to the following question, “Compared to most Black people, what shade of skin color do you have?” Response categories included very light brown, light brown, medium brown, dark brown, and very dark brown. Although there are likely differences in how an individual identifies their own skin color vs. how others might characterize it, we believe relying on a self-reported measure of skin color is particularly useful because it reflects individuals’ self-perceptions of this racialized trait and social statuses that follow from it. We use the category medium brown as the reference group for the skin tone variable in all analyses. This approach was taken given recent evidence that medium complexioned Blacks report better mental health outcomes when compared to those with lighter or darker complexions (Monk 2015).

We controlled for several covariates in multivariate regression analyses that could confound the association between skin tone and mental health among our sample. These included age, sex, marital status, household income (capped at $200,000), and educational attainment. Respondents who identified as male, were married, or did not graduate from high school were used as the reference categories in all analyses. We conducted sensitivity analyses which included generational status as a control variable. The inclusion of this additional covariate did not qualitatively change our results; therefore, in the interest of parsimony, we deleted it from our final regression models.

Analytic Approach

First, we generated descriptive statistics for the combined sample and stratified by ethnicity (AAs vs. CBs). These included means and standard deviations for continuous variables and frequency distributions for categorical variables. We also tested the extent to which each variable significantly differed for AAs vs. CBs by estimating bivariate logistic regression models. Finally, we ran a series of multivariate logistic regression models to estimate the association between skin tone and mental health for AAs and CBs. We calculated robust standard errors using the Huber/White correction method. All statistical analyses were conducted using probability weights to account for the NSAL’s complex multistage sampling design (Jackson et al. 2004).

Results

Descriptive findings are presented in Table 1. Overall, 31% of NSAL respondents experienced some type of mental health disorder (any mental disorder) in their lifetime. The prevalence of disorders is significantly higher among AAs than CBs—31% of AAs report having any mental disorder, compared to 26% of CBs. Regardless of ethnicity, the most common disorder is an anxiety disorder. Nearly 20% of NSAL participants report having an anxiety disorder at some point in their lifetime. Fewer CBs (16%) than AAs (19%) ever experienced an anxiety disorder. A similar pattern emerges for substance use disorders. Overall, 11% of the study sample report symptoms consistent with this psychiatric condition, this includes 12% of AAs but only 9% of CBs. Although AAs are more likely than CBs to suffer from anxiety or substance abuse, the opposite is true for mood disorders. 14% of CBs report experiencing a mood disorder, compared to only 11% of AAs.

Table 1.

Descriptive Statistics among African American and Caribbean Black NSAL Respondents

	Combined Sample	African Americans	Caribbean Blacks	p-value
Mental Health Outcomes
Any Disorder	30.70	31.00	26.40	0.12
Mood Disorder	11.40	11.30	13.50	0.38
Anxiety Disorder	19.20	19.40	16.10	0.16
Substance Use Disorder	11.40	11.50	9.29	0.47
Skin Color
Very Light	4.90	4.93	4.52	0.66
Light	15.50	15.50	14.70	0.28
Medium	48.50	49.10	38.40	Ref
Dark	25.00	24.60	31.60	<0.01
Very Dark	6.12	5.83	10.70	0.01
Ethnicity
African American	94.00
Caribbean Black	5.98
Mean Age	41.8	41.9	40.2	0.07
	(0.5)	(0.5)	(0.8)
Sex
Male	44.20	43.80	50.10	Ref
Female	55.80	56.20	49.90	0.02
Marital Status
Currently Married	42.20	41.80	49.00	Ref
Formerly Married	26.10	26.50	19.30	<0.01
Never Married	31.70	31.70	31.70	0.13
Education
Some High School	23.80	24.10	19.90	<0.01
High School Graduate	37.20	37.60	31.20
Some College	24.30	24.00	28.60
College Graduate	14.60	14.30	20.30
Median Household Income	29000	33828	39000	0.02
	(1,260)	(1,323)	(3,098)
N	4,771	3,393	1,378

p < 0.001

p < 0.01

p < 0.05

p < 0.10

Source: National Survey of American Life (NSAL)

Notes: P-values for bivariate differences were calculated using OLS, logistic, ordinal, & multinomial regression models where appropriate.

Among NSAL respondents, skin tones vary considerably with most respondents falling within the middle of the skin color spectrum. Almost half (49%) of the combined sample describes their skin tone as medium brown, while another 25% and 16% self-identify as having dark brown or light brown skin, respectively. Only 6% of NSAL participants described themselves as having very dark skin, while another 5% described their skin color as very light brown. We also note that the distribution of skin tone differs by ethnicity. 49% of AAs but only 38% of CBs self-identify as having medium brown skin. These interethnic distinctions continue when examining the next largest group of respondents. Only 25% of AAs characterize their skin color as dark brown compared to 38% of CBs.

There are notable interethnic differences among NSAL respondents for several sociodemographic controls. Generally, CBs tend to be better off than their AA counterparts. One-third (34%) of AAs are currently married compared to almost half (49%) of CBs. Relatedly, 27% of AAs but only 19% of CBs report being previously married. In terms of SES, 14% of AAs graduated with at least a bachelor’s degree compared to 20% of CBs. Median household incomes for AAs and CBs are $33,828 and $39,000, respectively.

Results from logistic regression models estimating the association between skin color and any disorder are presented in Table 2. Findings for the combined sample reveal that respondents with the lightest skin tones have the greatest risk of experiencing any type of psychiatric illness. Compared to Black Americans with medium brown skin, those with very light skin face a 93% increase in the odds of having any mental disorder. However, when we examine the association between skin color and this measure of general psychological functioning by ethnicity, distinct trends emerge.

Table 2.

Regression Results Predicting Lifetime Mental Health Disorders by Skin Tone for Pooled Sample, African Americans, & Caribbean Blacks

	Any Disorder
	Combined		African Americans		Caribbean Blacks
	OR	95% CI	OR	95% CI	OR	95% CI
Skin Color
Very Light	1.93**	1.31 - 2.83	1.86**	1.24 - 2.78	3.17*	1.02 - 9.83
Light	0.94	0.75 - 1.18	0.93	0.73 - 1.18	1.20	0.56 - 2.59
Medium	1.00	Ref	1.00	Ref	1.00	Ref
Dark	1.02	0.85 - 1.23	1.04	0.86 - 1.26	0.84	0.35 - 2.04
Very Dark	1.00	0.72 - 1.39	1.04	0.73 - 1.49	0.65	0.30 - 1.41
Ethnicity
African American	1.00	Ref
Caribbean Black	0.83	0.63 - 1.11
Age	0.98***	0.98 - 0.99	0.98***	0.98 - 0.99	0.99	0.97 - 1.00
Sex
Male	1.00	Ref	1.00	Ref	1.00	Ref
Female	0.98	0.82 - 1.17	1.00	0.83 - 1.22	0.64	0.33 - 1.26
Marital Status
Married	1.00	Ref	1.00	Ref	1.00	Ref
Formerly Married	1.67***	1.36 - 2.04	1.68***	1.35 - 2.08	1.35	0.86 - 2.11
Never Married	1.00	0.78 - 1.28	0.96	0.73 - 1.24	1.90*	1.07 - 3.37
Household Income (Ln)	0.88**	0.81 - 0.97	0.88*	0.80 - 0.97	0.89	0.63 - 1.25
Education
Some High School	1.00	Ref	1.00	Ref	1.00	Ref
High School Graduate	0.61***	0.48 - 0.77	0.61***	0.48 - 0.77	0.72	0.19 - 2.65
Some College	0.74*	0.55 - 0.99	0.75+	0.55 - 1.01	0.73	0.19 - 2.79
College Graduate	0.79	0.59 - 1.07	0.78	0.57 - 1.07	1.14	0.38 - 3.41
N	4,771		3,393		1,378

^***p < 0.001

^**p < 0.01

^*p < 0.05

⁺p < 0.10

Source: National Survey of American Life (NSAL)

First, CBs with the lightest skin tones seem to be at a greater risk of experiencing any disorder than AAs. CBs who fall into this category face more than a three-fold increase in the odds of having any type of psychiatric illness compared to coethnics with medium brown skin. AAs with very light skin only have an 86% increase in the odds of having any mental disorder. Results from postestimation testing suggest that the difference between light skinned CBs and AAs is not statistically significant (p > 0.05). However, this is likely due to a lack of precision caused by small cell sizes rather than an absence of effect.

Second, a linear gradient between skin color and general psychological functioning emerges among CBs but not AAs. This can be seen by comparing the magnitude of the odds ratios across skin color categories within each ethnic group. We focus on the magnitude of slope coefficients rather than tests of statistical significance due to limited samples sizes in the most extreme response categories. For CBs, those with the lightest skin tones face the greatest risk of experiencing any mental health disorder, and the magnitude of the odds ratios decreases monotonically as skin color becomes darker. For AAs, only respondents with the lightest skin tones experience an increased risk of having any mental disorder. The magnitude of the odds ratios for the other skin color groups remains very close to 1.00.

We used results from regression models to generate predicted probabilities of having any mental disorder separately for AAs and CBs. These are presented in Figures 1a and 1b. These graphs clearly illustrate how the association between skin color and general psychological functioning differs for these two subpopulations. For CBs, the probability of having any psychiatric condition is highest for those with the lightest skin tones (52%) but decreases as skin color darkens—30% for light brown CBs, 26% for medium brown, 23% for dark brown, and 19% for very dark brown (Figure 1a). However, among AAs, only those with the lightest complexions have an increased risk (44%) of experiencing psychiatric disorders (Figure 1b). AAs with other complexions have comparable probabilities of developing any disorder, ranging from 29-31%, with no significant differences between skin color groups.

Figure 1a.

Predicted Probabilities of Any Disorder: Caribbean Blacks

Figure 1b.

Predicted Probabilities of Any Disorder: African Americans

Next, we generated logistic regression results for cause-specific mental disorders (mood, anxiety, and substance use). These are presented in Table 3. Findings for mood disorders are similar to those for any disorder, with a few notable distinctions. Results for the combined sample indicate that only Black Americans with the lightest skin experience a significant increase (82%) in the odds of experiencing mood disorders compared to those with medium complexions. When examining findings separately by ethnicity, results for AAs mirror those of the combined sample. AAs with the lightest skin color experience a 79% increase in the odds of having a mood disorder compared to those with medium skin, while AAs with other skin tones experience similar odds of having a mood disorder. However, CBs with the lightest skin have the highest risk of experiencing a mood disorder and this decreases monotonically as complexion darkens. These interethnic differences in the association between skin color and the probability of having a mood disorder are apparent when examining Figures 2a and 2b. Again, to make comparison across Black subpopulations, we focus on the magnitude of the odds ratios rather than solely interpreting results based on statistical significance. This is an appropriate – some might argue preferable – approach when dealing with smaller cell sizes in certain (i.e. more extreme) response categories (Sullivan and Feinn 2012).

Table 3.

Regression Results Predicting Specific Mental Health Disorders by Skin Tone for African American & Caribbean Black NSAL Respondents

	Mood Disorder						Anxiety Disorder						Substance Use Disorder
	Combined		African Americans		Caribbean Blacks		Combined		African Americans		Caribbean Blacks		Combined		African Americans		Caribbean Blacks
	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI
Skin Color
Very Light	1.82*	1.13 - 2.95	1.79*	1.09 - 2.92	1.87	0.14 - 24.3	1.56*	1.07 - 2.27	1.49*	1.01 - 2.20	2.48	0.42 - 14.7	1.51+	0.96 - 2.38	1.27	0.81 - 1.98	10.2**	2.77 - 37.9
Light	0.91	0.69 - 1.21	0.90	0.67 - 1.19	1.25	0.35 - 4.46	0.83	0.65 - 1.07	0.80	0.61 - 1.05	1.71*	1.08 - 2.68	1.04	0.66 - 1.65	1.09	0.68 - 1.76	0.31*	0.13 - 0.75
Medium	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref
Dark	0.91	0.69 - 1.20	0.95	0.71 - 1.26	0.62	0.21 - 1.87	0.98	0.80 - 1.20	0.98	0.79 - 1.22	1.27	0.64 - 2.51	0.99	0.74 - 1.33	1.00	0.74 - 1.36	0.82	0.19 - 3.60
Very Dark	1.17	0.78 - 1.76	1.24	0.80 - 1.92	0.73	0.15 - 3.57	0.89	0.60 - 1.34	0.88	0.56 - 1.38	1.22	0.60 - 2.50	1.04	0.60 - 1.78	1.12	0.63 - 1.97	0.26*	0.079 - 0.84
Ethnicity
African American	1.00	Ref					1.00	Ref					1.00	Ref
Caribbean Black	1.29	0.82 - 2.03					0.88	0.61 - 1.26					0.78	0.42 - 1.44
Age	0.98***	0.97 - 0.99	0.98***	0.97 - 0.99	0.99	0.96 - 1.02	0.98***	0.98 - 0.99	0.98***	0.98 - 0.99	1.01	0.99 - 1.02	1.00	0.99 - 1.00	1.00	0.99 - 1.00	0.97	0.95 - 1.00
Sex
Male	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref
Female	1.58**	1.20 - 2.08	1.70***	1.27 - 2.28	0.71	0.29 - 1.75	1.84***	1.44 - 2.35	1.92***	1.48 - 2.49	0.96	0.40 - 2.31	0.25***	0.19 - 0.32	0.25***	0.19 - 0.33	0.16**	0.047 - 0.51
Marital Status
Married	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref
Formerly Married	1.78**	1.22 - 2.61	1.80**	1.20 - 2.70	1.63	0.66 - 4.02	1.38**	1.13 - 1.67	1.41**	1.15 - 1.72	0.91	0.32 - 2.60	1.52**	1.19 - 1.94	1.53**	1.19 - 1.98	0.75	0.38 - 1.51
Never Married	1.15	0.85 - 1.54	1.11	0.80 - 1.53	1.75*	1.07 - 2.87	1.00	0.74 - 1.35	0.96	0.69 - 1.32	2.24*	1.19 - 4.24	0.94	0.65 - 1.35	0.91	0.62 - 1.32	0.83	0.35 - 1.97
Household Income (Ln)	0.97	0.84 - 1.12	0.97	0.84 - 1.13	0.96	0.58 - 1.59	0.86*	0.77 - 0.97	0.87*	0.76 - 0.99	0.79*	0.65 - 0.95	0.78***	0.69 - 0.87	0.76***	0.68 - 0.86	0.98	0.67 - 1.42
Education
Some High School	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref	1.00	Ref
High School Graduate	0.88	0.62 - 1.25	0.91	0.64 - 1.31	0.57	0.13 - 2.48	0.65***	0.51 - 0.83	0.64***	0.49 - 0.82	1.14	0.34 - 3.80	0.47***	0.34 - 0.65	0.49***	0.35 - 0.68	0.20	0.037 - 1.11
Some College	0.98	0.63 - 1.50	1.00	0.63 - 1.57	0.84	0.19 - 3.71	0.72*	0.54 - 0.97	0.72*	0.53 - 0.97	1.12	0.33 - 3.79	0.66*	0.48 - 0.91	0.70*	0.51 - 0.96	0.34	0.066 - 1.78
College Graduate	1.27	0.77 - 2.08	1.26	0.74 - 2.15	1.44	0.34 - 6.17	0.81	0.57 - 1.17	0.82	0.56 - 1.20	0.97	0.34 - 2.81	0.33**	0.15 - 0.72	0.31*	0.13 - 0.75	0.44	0.12 - 1.59
N	4,771		3,393		1,378		4,771		3,393		1,378		4,771		3,393		1,378

^***p < 0.001

^**p < 0.01

^*p < 0.05

⁺p < 0.10

Source: National Survey of American Life (NSAL)

Figure 2a.

Predicted Probabilities of Mood Disorder: Caribbean Blacks

Figure 2b.

Predicted Probabilities of Mood Disorder: African Americans

Stratified regression results for anxiety disorders are similar to those for any and mood disorders. Only AAs with the lightest complexions experience a higher probability of having an anxiety disorder, with the odds of this particular outcome increasing by 49% for AAs with very light skin compared to coethnics with medium brown skin (Table 3, Model 2). Among CBs, the odds of experiencing an anxiety disorder are almost 2.5 times higher for those with very light skin compared to those with medium brown skin. This risk decreases in a stepwise fashion as skin color becomes darker for CBs with light brown and medium brown skin, before increasing as complexion darkens. CBs with dark brown and very dark brown skin experience a 27% and 22% increase in the odds of having an anxiety disorder compared to CBs with medium brown skin. Thus, the association between skin tone and anxiety appears to be approaching curvilinearity among this Black subgroup. Once again, we used findings from Model 2 Table 3 to generate predicted probabilities of having an anxiety disorder by skin color separately for CBs and AAs. These are presented in Figures 3a and 3b.

Figure 3a.

Predicted Probabilities of Anxiety Disorder: Caribbean Blacks

Figure 3b.

Predicted Probabilities of Anxiety Disorder: African Americans

Regression results for substance use disorders are shown in Model 3, Table 3. These findings are the most divergent compared to those for any, mood, or anxiety disorders. However, these differences might, at least in part, be driven by small cell sizes especially among CBs that result in large standard errors and less stable estimates. AAs with the lightest complexions face a 27% increase in the odds of having a substance use disorder compared to coethnics with medium brown skin; however, this difference fails to reach statistical significance (p > 0.05). AAs with all other skin tones have similar odds of experiencing this particular disorder. The pattern that emerges for CBs is much different. CBs with very light brown skin are more than ten times as likely as CBs with medium brown skin to report a substance use disorder, while coethnics with all other complexions face smaller probabilities of having a substance use disorder. CBs with light brown skin and very dark brown skin face a 69% and 74%, decrease in the odds of this outcome compared to CBs with medium brown skin. We present predicted probabilities that were generated using these results in Figures 3a and b.

Discussion

This study investigates how skin color influences mental health outcomes among different segments of Black Americans. First, we ask does the association between colorism and mental health differ for CBs compared to AAs. Second, is this relationship more or less pronounced for CBs than AAs? Third, does the shape of the skin tone/mental health gradient follow a similar pattern for these two Black subgroups? Our results suggest a more nuanced “story” worthy of further theorization and empirical investigation. We further elucidate our findings and place them within a broader context below.

Regarding the first research question, we find that Black Americans with the lightest skin tones, regardless of ethnicity, experience the lowest levels of psychological functioning. This result stands in contrast to much of the literature on colorism and health that typically reveals Black Americans with darker skin to have worse health outcomes than their counterparts with lighter skin (Cobb et al. 2016; Hargrove 2018; Sweet et al. 2007). We consider two possible explanations for this finding. First, Black Americans with the lightest complexions may experience more psychosocial stress and face higher risks of developing a psychological disorder due to the compounded effects of interpersonal discrimination from both White and Black Americans. Research on colorism suggests that Black Americans tend to view those with medium complexions more favorably than those on the ends of the color spectrum (i.e. very light or very dark) (Landor and McNeil Smith 2019). Moreover, because intragroup rejection is negatively associated with self-esteem, mastery, and depression (Basáñez et al. 2014; Uzogara and Jackson 2016), these intragroup divisions may have a stronger impact on mental health than physical health.

Second, our finding that Black Americans with the lightest complexions tend to have the worst mental health outcomes could be driven by a sense of relative deprivation, whereby this subset of Black Americans anticipate social benefits because of their proximity to whiteness. Due to the pervasive, entrenched nature of racism in contemporary U.S. society, however, their skin color does not bring the anticipated social resources known to be health protective (Link and Phelan 2015). Although Black Americans with the lightest complexions often receive preferential treatment in the labor and marriage markets as well as in educational settings (Devaraj et al. 2018; Hunter 2005), they are not immune to the health eroding effects of discrimination and unfair treatment (Williams 2012). To this end, Colen et al. (2018) demonstrate that upwardly mobile Black Americans do not experience a reduction in exposure to acute or chronic interpersonal discrimination as they experience upward socioeconomic mobility over time.

Although relative deprivation theory and proximity to whiteness are often invoked to gain insight into racialization processes, they have played an outsized role in explaining the unequal distribution of health across various segments of society. Conceptualizations that emphasize the importance of relative, as opposed to absolute, social processes and resources have long been used to explain social inequalities in health (Subramanyam et al. 2009; Sun et al 2012). For example, some scholars argue that income inequality drives suboptimal health outcomes among lower SES groups rather than the amount of material socioeconomic resources available to them (Marmot 2004; Pickett and Wilkinson 2015). Relative deprivation should not be viewed solely as the result of psychosocial processes. Instead, it arises when an individual’s position within the social hierarchy is decidedly different from those around them, thus invoking structural origins and explanations. This type of approach, which highlights the dynamic interplay among structural factors, individual status, and psychosocial mechanisms is similar to other frameworks specifically designed to explain how racial inequalities become embodied and “get under the skin” (Geronimus 1994).

Regarding the second and third research questions, our results provide preliminary evidence that the association between skin color and mental health is different for CBs than AAs. Only AAs with the lightest skin color face an increased risk of psychiatric illness. The relationship between skin color and mental health among CBs appears to adhere to a different trajectory. For any disorder and mood disorders, as skin color becomes darker, the likelihood of experiencing these outcomes decreases monotonically. For anxiety disorders, the association between skin color and this particular endpoint appears to be curvilinear, whereby individuals with the lightest and darkest skin tones face the greatest risk of having an anxiety disorder and those with medium brown tones face the lowest risk. Thus, not only does the shape of the relationship between skin color and mental health among Black Americans depend on ethnicity, it also depends on the particular outcome under study.

The linear gradient between skin color and any disorders as well as mood disorders among CBs is notable and underscores the importance of considering interethnic distinctions among Black Americans when investigating the health effects of colorism. Unfortunately, due to data limitations, we are not able to further explore specific reasons behind the finding that CBs with darker skin tones experience lower levels of psychological wellbeing than coethnics with lighter skin tones; however, we can hypothesize about potential explanations.

In Caribbean countries, individuals with lighter complexions are typically afforded a higher social status within society (Kelly 2020; Wilder and Cain 2011), and we expected this social standing to translate into health benefits. However, given the uniquely different social context that CBs find once they move to the U.S. and the ubiquitous nature of structural racism and racial stigmatization, those with lighter complexions might not experience the same social benefits of having lighter complexions. Resultingly, the transition into a “racialized other” may be particularly difficult for this subgroup because they occupied a privileged social position in their home countries. We would expect the health impacts of these racialization processes to be most pronounced among recent immigrants and weaken as time in the U.S. increases, particularly among members of subsequent generations.

Alternatively, it could be that CBs with lighter complexions face questions related to authenticity from others due to their relatively lighter complexions. Blacks with lighter skin report feeling excluded or discriminated against by other members of their racial group for not being “Black enough” (Landor and McNeil Smith 2019). While these studies focus on AAs, it is possible that CBs with lighter complexions have similar experiences. Research examining colorism amongst CBs is scarce—further investigation is necessary to fully understand the role of race, ethnicity, and skin tone, particularly as it relates to the patterning of population mental health.

The curvilinear relationship we found between skin color and anxiety disorders for CBs stands in contrast to the other disorders, indicating that the association between skin color and mental health depends on the specific outcome under study. This variation is likely attributable to the non-specific effects of social stress (Selye 2013). While social stress is associated with various psychological outcomes, the effects of social stress on specific indicators of mental health are not congruent (Aneshensel, Rutter, and Lachenbruch 1991). This is consistent with psychiatric research which finds that while disorders such as anxiety and depression may be related, they are independent ailments with numerous causes and should be treated as such (Tiller 2013). Research investigating mental health outcomes should take care to differentiate and compare between various types of disorders to avoid obscuring important differences across outcomes.

In sum, the inconsistencies between our findings and those from previous studies that examine the association between skin color and health among Black populations (Cobb et al. 2016; Hargrove 2018; Monk 2015) suggests that this relationship is likely to be more complicated than originally assumed. In particular, we expect to find further health disparities across Black subpopulations based on important sociological indicators such as skin tone, SES, and immigrant status. Thus, the results from the current study underscore the need for more careful and rigorous research that critically examines mental health disparities within rather than just across different racial/ethnic groups.

Limitations

Although our findings provide important insights, they should be interpreted in light of some shortcomings. First, NSAL data are cross-sectional. Thus, we were unable to assess how skin color predicts changes in psychological functioning over time. Since our independent variables of interest – skin color and ethnicity – are likely to be time invariant, findings presented here are not likely biased due to reverse causality. However, since self-perception is known to change, it is possible that perceived skin color varies over time. We look to future research to more fully examine the relationship between skin color and mental health once newly collected longitudinal data that oversamples key Black subpopulations becomes available.

Second, few CBs in the sample have very light brown or very dark brown complexions. Because of this, regression results for respondents with the lightest skin tone, especially among CBs, tend to have large confidence intervals. To be clear, a lack of statistical power among certain subgroups will not bias point estimates but will produce less precise estimates. Future studies should collect data from a larger sample of CBs with skin tones that range from the lightest to the darkest shades. This purposeful approach to participant recruitment will ensure adequate cell sizes, smaller standard errors, tighter confidence intervals, and more stable regression results.

Finally, NSAL data were collected nearly 20 years ago. Therefore, we do not know how generalizable our findings are to more recent cohorts of Black Americans. Since the NSAL was carried out, the U.S. has suffered not one, but two, of the most devastating economic upheavals in modern history (i.e. the Great Recession and the Covid-19 economic downturn), which negatively impacted the health and wellbeing of Black Americans more than their White counterparts (Hardy and Logan 2020; Reynolds 2014; Snowden and Bruckner 2014). In 2008, the United States elected its first Black president. While the economic downturn may have had a negative impact on mental health among Black Americans, some studies suggest Barack Obama’s election had the opposite effect (Brooks 2012).

Despite these limitations, this study highlights important disparities among Black subpopulations. As Black immigration to the U.S. continues to increase, it is vital to better our understanding of how Black immigrant health shapes health disparities in the United States. This study contributes to the existing knowledge on health disparities by examining understudied factors such as skin tone that may be contributing to both Black-White health disparities and disparities within the Black population.

Figure 4a.

Predicted Probabilities of Substance Use Disorder: Caribbean Blacks

Figure 4b.

Predicted Probabilities of Substance Disorder: African Americans

• The lightest Black Americans have the highest odds of mental health disorders

• The association between skin tone and mental health differs by ethnicity

• The association between skin tone and mental health differs by disorder type