A Nationwide Study of Discrimination and Chronic Health Conditions Among Asian Americans

Abstract

Objectives. We examined whether self-reported everyday discrimination was associated with chronic health conditions among a nationally representative sample of Asian Americans.

Methods. Data were from the Asian American subsample (n = 2095) of the National Latino and Asian American Study conducted in 2002 and 2003. Regression techniques (negative binomial and logistic) were used to examine the association between discrimination and chronic health conditions. Analyses were conducted for the entire sample and 3 Asian subgroups (Chinese, Vietnamese, and Filipino).

Results. Reports of everyday discrimination were associated with many chronic conditions, after we controlled for age, gender, region, per capita income, education, employment, and social desirability bias. Discrimination was also associated with indicators of heart disease, pain, and respiratory illnesses. There were some differences by Asian subgroup.

Conclusions. Everyday discrimination may contribute to stress experienced by racial/ethnic minorities and could lead to chronic illness.

Although nearly everyone encounters some unfair treatment in the course of a lifetime, members of marginalized groups are more likely to experience and report discrimination than are members of groups with more power and privilege.1 In extreme cases, discrimination may take the form of a hate crime and cause immediate harm. Chronic exposure to more subtle discrimination, however, might be equally important for health.2^,3 Over the long term, discrimination—biased actions against members of socially marginalized groups by individuals and institutions—may bring about socioeconomic and other disadvantages, contribute to exposure to environmental hazards such as pollution and job stress, result in inadequate health care, and lead to an accumulation of stressors over the life course.4^–9

These multiple pathways suggest that discrimination may be an important determinant of numerous outcomes. Not surprisingly, reports of discrimination have been linked to many issues, including high blood pressure,10^,11 respiratory problems,12 somatic complaints,13 self-rated health,14^,15 mental health,16^,17 and chronic health conditions.18^,19

Asian Americans provide an interesting test of the hypothesis that discrimination contributes to morbidity because of their image as a “model minority.” According to this image, Asian Americans are assumed to hold a unique status as a “successful” minority, one free from discrimination.20^,21

Challenging that assumption are studies documenting contemporary discrimination against Asian Americans.22^,23 For example, a recent national housing audit found that 1 in 5 Asian Americans experience systematic discrimination in home buying (the same level as African Americans).24 The general US population has a dual view of Asian Americans. Although Asian Americans are seen as hardworking and economically prosperous, they are also viewed as untrustworthy and inscrutable.25 One in 4 US residents report that they believe that Chinese Americans are taking away “American” jobs.26 The increasing prominence of Asian nations as global competitors to the United States may also increase anti-Asian sentiment.27^,28 Additionally, Asian Americans are one of the fastest-growing groups in the United States, projected to increase from 10.7 million in 2000 to 33.4 million in 2050.29 Although increasing diversity may increase tolerance, it may also generate intergroup conflict.28^,30 Accordingly, it is important to establish a baseline of reports by Asian Americans of discrimination and examine how these reports may be associated with their health.

Several studies of discrimination among Asians and Pacific Islanders in other countries can be found,12^,17^,31^,32 but studies in the United States are mostly based on small convenience samples.33^–38 Studies using randomly selected, population-based samples have come primarily from 2 studies on the West Coast. The first was a household survey of 1503 Chinese Americans residing in Los Angeles in 1993 and 1994. Participants’ replies to 2 questions on racial/ethnic and language and accent discrimination were associated with health-related quality of life and depressive symptoms,20 and decreased use of mental health services.39 The second was a household survey of 2200 Filipino Americans living in San Francisco and Honolulu in 1998 and 1999. This study found that both episodic and chronic experiences with discrimination were associated with depressive symptoms,40 substance use,41 and chronic health conditions.19

Although these studies found an association between self-reported discrimination and health, it is unclear to what extent the findings can be generalized to Asian Americans nationwide. In this study, we investigate the association between self-reported everyday discrimination and chronic conditions among a nationally representative sample of Asian Americans. In addition to examining Asian Americans in the aggregate, we examine 3 of the largest Asian American ethnic groups: Chinese, Filipino, and Vietnamese Americans. Ongoing, routine discrimination may represent a type of chronic stressor that can be especially toxic to ethnic minorities.6^,19^,19^,42^–44 The accumulation of tolls related to discrimination on an everyday basis may contribute to allostatic load (the “wear and tear” of body systems resulting from an accumulation of stressors over time), erode protective resources such as wealth and social support, and lead to a variety of chronic illnesses.5^,45^,46 Accordingly, we examine chronic health conditions as the outcome.

The association between perceived discrimination and health conditions may be confounded by survey response factors. We investigated one such factor, social desirability bias, a bias caused by the tendency of people to respond to surveys in ways that present a favorable impression of themselves.47^–49 Individuals concerned about protecting their self-image or who do not want to accept a lower social status may underreport discrimination. Social desirability may also be related to a behavior that is especially relevant for Asian populations, the prevention of “loss of face” (shaming of oneself or one’s family).50^,51 We investigated whether the association between perceived discrimination and health conditions holds after controlling for social desirability bias and other sociodemographic factors.

METHODS

Sample

Data were from the National Latino and Asian American Study, a household survey conducted in 2002 and 2003. Our analyses focus on the Asian American respondents.

The sampling design has been detailed elsewhere.52^,53 Briefly, the design included 3 components: (1) a core sampling of primary sampling units (metropolitan statistical areas and counties) and secondary sampling units (from continuous groupings of census blocks) with probability proportional to size, from which housing units and household members were sampled; (2) high-density supplemental sampling of census block groups in which the targeted ethnic groups made up more than 5% of the population; and (3) second-respondent sampling to recruit participants from households where a primary respondent had already been interviewed. Sample weights were developed to account for joint probabilities of selection for these 3 components and designed to allow the sample estimates to be nationally representative.53

Respondents were at least 18 years of age and resided in the United States. Trained interviewers, with linguistic and cultural backgrounds similar to those of the respondent, administered the survey in the respondent’s chosen language: English, Cantonese, Mandarin, Tagalog, Vietnamese, or Spanish. Instruments were translated into these languages through standard techniques (translation of the instrument to a given language, followed by translation back to English for verification). Interviews were conducted face-to-face unless respondents requested a telephone interview. A total of 2095 adults (1611 primary and 484 secondary respondents) were included in our sample. The response rates (calculated using the American Association for Public Opinion Research, Response Rate Method 3—or AAPOR-RR3 method)53 for primary and secondary respondents were 69.3% and 73.7%, respectively. Our sample included 600 Chinese, 508 Filipino, 520 Vietnamese, and 467 others (141 Indians, 107 Japanese, 81 Koreans, 39 Pacific Islanders, and 99 members of small subgroups). These 467 “others” were included in our aggregate analysis but excluded from the subgroup analysis.

Measures

Chronic conditions were assessed with the World Mental Health Composite International Diagnostic Interview (WMH-CIDI), a self-reported checklist of lifetime physical and psychophysiological disorders. Developed by the World Health Organization, the WMH-CIDI was designed for cross-cultural epidemiological research.54 Individuals were asked if they had ever had any of the following: arthritis or rheumatism, chronic back or neck problems, frequent or severe headaches, other chronic pain, hay fever and other seasonal allergies, stroke, heart attack, heart disease, high blood pressure, asthma, tuberculosis, other chronic lung disease, diabetes or high blood sugar, ulcer in stomach or intestine, HIV/AIDS, epilepsy or seizure, and cancer. The number of chronic conditions in our sample ranged from 0 to 10.

Some conditions may be more closely linked to the stress process than others. To examine this potential heterogeneity, we divided chronic conditions into 4 major classifications: cardiovascular (heart attack, stroke, heart disease, high blood pressure), respiratory (hay fever, asthma, tuberculosis, and other chronic lung diseases, including emphysema and chronic obstructive pulmonary disease), pain (chronic back or neck problems, frequent or severe headaches, arthritis, ulcer, other chronic pain) and other (diabetes, HIV/AIDS, cancer, epilepsy or seizures). Each of these 4 categories was coded “1” if the respondent had any of the conditions and “0” if not.

Everyday discrimination was assessed using a 9-item scale adopted from the Detroit Area Study,44 which measures perceptions of chronic and routine unfair treatment.42 The items were as follows: (1) “You are treated with less courtesy than other people,” (2) “You are treated with less respect than other people,” (3) “You receive poorer service than other people at restaurants or stores,” (4) “People act as if they think you are not smart,” (5) “People act as if they are afraid of you,” (6) “People act as if they think you are dishonest,” (7) “People act as if you are not as good as they are,” (8) “You are called names or insulted,” and (9) “You are threatened or harassed.” As with prior studies,2^,55 an exploratory factor analysis found support for a 1-factor structure (eigenvalue = 4.87; factor loadings = 0.67–0.79); hence, we used this measure as a unidimensional scale. Items were summed and then divided by 9. The range of responses was 1 to 6, with 1 indicating that respondents never perceived experiencing unfair treatment and 6 indicating that respondents experienced it almost every day. The Cronbach α for our sample was 0.91. After this series of questions, respondents were asked what they considered to be the main reason for their experiences of everyday discrimination (e.g., their ethnicity); these reasons were mutually exclusive.

Social desirability bias was measured using the Marlowe–Crowne 10-item scale,47 in which the number of affirmative replies to the following items are summed: “I never met a person that I didn’t like,” “I always win at games,” “I have never been bored,” “I never get annoyed when people cut ahead of me in line,” “I never get lost, even in unfamiliar places,” “I have always told the truth,” “My table manners at home are as good as when I eat out,” “I have never lost anything,” “No matter how hot or cold it gets, I am always quite comfortable,” and “It doesn’t bother me if someone takes advantage of me.” The range in our sample was 0 to 10, with 0 indicating no social desirability bias.

We assessed education using 4 categories (less than a high school diploma, high school graduate, some college education, and college graduate or beyond). Current employment was contrasted with unemployment. Per capita income was measured as the total household income (with missing values imputed) divided by the number of people in the household.

English language proficiency was assessed by the question, “How well do you speak English?” Responses were coded “excellent/ good” or “fair/poor.”

Gender, age, marital status (married vs others), nativity (US-born vs other), and region (Northeast, Midwest, South, West) were also assessed.

Analyses

We computed weighted descriptive statistics for the entire sample and for Asian subgroups. Our analyses took into account sample design effects by using the svy commands of Stata 9.2 (Stata Corp, College Station, Tex), which allow for estimation of standard errors in the presence of stratification and clustering. To account for within-household clustering, we also included a covariate in our analyses to indicate a first or second respondent from a household.

We used negative binomial regression to perform multivariate analyses of the number of chronic health conditions. One outlier was removed from the final models. We used logistic regression to model the specific conditions (cardiovascular, respiratory, pain, and other). First we analyzed the total sample, which allowed us to make inferences about all Asian Americans nationwide, including those not oversampled in our study (e.g., Pakistanis). We then analyzed the 3 largest groups represented in our study (Vietnamese, Filipinos, and Chinese) to examine potential ethnic heterogeneity. We centered continuous measures at their means (i.e., each continuous variable was subtracted by the overall sample mean of that variable; thus, the intercept represented the predicted number of chronic conditions for the “average” person in our sample) to reduce multicollinearity and to facilitate the interpretation of the intercept.56 Our analyses focused on everyday discrimination, and we also included the attribution of discrimination to race/ethnicity, nationality, or skin color as a covariate in initial models.

RESULTS

Table 1 ▶ shows characteristics of the sample. On average, participants reported 1.3 chronic conditions. Half of the sample were women, and most were married, employed, aged in their early 40s, and foreign born. Per capita income averaged $39 700. There were ethnic differences. Vietnamese respondents had a large proportion of immigrants, reported the lowest levels of English proficiency, and had the lowest per capita income. Their employment rate was similar to those of the other groups, however, and they had the highest marriage rate. Filipino respondents reported the highest level of English proficiency, although most were immigrants. Chinese respondents reported the highest per capita income and the largest percentage of college graduates, although a fifth did not complete high school

TABLE 1—

Weighted Sample Characteristics of Asian and Pacific Islander Americans, by Ethnicity: National Latino and Asian American Study, 2002–2003

	All Asian and Pacific Islander Americans	Vietnamese Americans	Filipino Americans	Chinese Americans
Total sample, no.	2095	520	508	600
Weighted percentage	100.0	12.9	21.6	28.7
Chronic conditions
Total, mean (SD)	1.3 (0.1)	1.3 (0.1)	1.6 (0.1)	1.3 (0.1)
Cardiovascular conditions, %	19.5	20.1	27.8	17.0
Respiratory conditions, %	35.0	35.9	38.7	39.0
Pain-related conditions, %	32.6	25.2	36.5	33.0
Other, %	21.4	22.6	25.7	22.8
Discrimination
Everyday discrimination, mean (SD)a	1.81 (0.03)	1.45 (0.04)	1.93 (0.05)	1.75 (0.03)
Main reason for discrimination, %
Ancestry, national origin, race/ethnicity, or skin color	56.1	53.8	57.3	55.2
Gender or sex	4.2	3.7	4.6	5.0
Age	5.8	7.8	5.1	6.9
Height or weight	2.2	2.0	4.6	1.8
Sexual orientation	0.5	0.0	0.0	0.6
Income or education	8.2	12.6	7.6	8.7
Other	23.1	20.1	20.8	21.9
Social desirability bias, mean (SD)b	2.3 (0.1)	2.4 (0.2)	2.0 (0.1)	2.5 (0.2)
Demographic characteristics
Women, %	52.5	54.6	55.5	53.6
Age, y, mean (SD)	41.3 (0.9)	43.7 (0.8)	43.0 (1.1)	43.0 (1.1)
Current marital status, %
Married	65.4	72.5	60.8	68.0
Never married	25.1	19.4	26.7	20.9
Widowed, separated, or divorced	9.6	8.1	12.6	11.2
Region, %
Northeast	15.7	17.2	8.2	17.4
Midwest	8.9	2.2	6.7	8.2
South	7.8	18.0	7.9	3.9
West	67.6	62.6	77.2	70.5
Education
Less than high school graduate	14.3	31.8	10.9	17.4
High school graduate	17.9	21.0	20.3	16.2
Some college	25.2	23.5	32.0	20.8
College graduate or beyond	42.6	23.8	36.8	45.6
Currently employed, %	63.7	61.5	64.5	64.2
Household income
Mean, $	84 180	63 515	90 913	89 852
Range, $, %
0–14 999	14.3	19.2	9.6	18.3
15 000–34 999	12.0	24.8	10.1	11.5
35 000–74 999	28.5	27.7	29.5	23.2
≥ 75 000	45.2	28.3	50.8	47.0
Household size, mean (SD)	2.9 (0.1)	3.5 (0.1)	3.1 (0.1)	2.6 (0.1)
Per capita income, $, mean (SD)	39 751 (1526)	24 272 (2257)	40 396 (3186)	45 555 (3622)
Foreign-born, %	76.9	97.0	70.2	82.0
Self-rated spoken English good or excellent, %	66.2	27.9	80.7	51.8

^aDetermined using a 9-item survey, adopted from the Detroit Area Study, measuring perceptions of chronic and routine unfair treatment (scale = 1–6); see “Methods” section.

^bSocial desirability bias is a bias due to a tendency of people to respond to surveys in ways that present a favorable impression of themselves. It was measured with the Marlowe–Crowne 10-item scale; see “Methods” section.

Reports of everyday discrimination varied by group. Filipinos reported the highest level, followed by Chinese and then Vietnamese. When respondents were asked for the main reason for everyday discrimination, the predominant reply was race, ethnicity or skin color, followed by other, income or education, age, gender, height or weight, and sexual orientation. Patterns regarding perceived reasons for discrimination were generally similar between the Asian subgroups, although Vietnamese respondents were more apt to attribute discrimination to income or education and Filipinos were more likely to cite height or weight.

Multivariate analyses focused on everyday discrimination. Final models examined only everyday discrimination; they did not include the main reasons for discrimination because these were not significantly associated with health conditions above and beyond the level of everyday discrimination, a finding echoed in the literature.2

Reports of discrimination were weakly correlated with score on the social desirability bias scale (r = −0.24, −0.15, −0.12, and −0.14 for Chinese, Filipinos, Vietnamese, and all Asians, respectively).

Table 2 ▶ shows the relationship between everyday discrimination and total chronic conditions, adjusted for covariates. As hypothesized, everyday discrimination was positively associated (b = .27) with chronic conditions for the entire sample. The parameter estimate for Chinese (b = .19) was lower than those for Vietnamese (b = .28) and Filipinos (b = .26). However, a test of the interaction between Chinese ethnicity and discrimination was not statistically significant, so the strength of the association was equal across the ethnic groups.

TABLE 2—

Association Between Self-Reported Everyday Discrimination and Total Chronic Health Conditions, by Ethnicity: National Latino and Asian American Study, 2002–2003

	All Asian and Pacific Islander Americans (n = 2095)		Vietnamese Americans (n = 520)		Filipino Americans (n = 508)		Chinese Americans (n = 600)
	Model 1, b (SE)a,b	Model 2, b (SE)a,b	Model 1, b (SE)a	Model 2, b (SE)a	Model 1, b (SE)a	Model 2, b (SE)a	Model 1, b (SE)a	Model 2, b (SE)a
Everyday discrimination	0.28 (0.04)**	0.28 (0.04)**	0.28 (0.08)**	0.28 (0.08)**	0.27 (0.05)**	0.26 (0.05)**	0.18 (0.07)**	0.19 (0.07)**
Social desirability bias		0.00 (0.01)		0.02 (0.03)		−0.05 (0.02)*		0.03 (0.02)

Note. Everyday discrimination was determined using a 9-item survey, measuring perceptions of chronic and routine unfair treatment (scale = 1–6); see “Methods” section. Social desirability was measured with a 10-item scale indicating the tendency of people to respond in surveys in ways that present a favorable impression of themselves. Model 1 included only everyday discrimination, and model 2 included everyday discrimination and social desirability bias.

^aAnalyses controlled for age, gender, per capita income, education, employment, marital status, region, nativity, language proficiency, and same household. Analyses are weighted to be nationally representative.

^bAnalysis controlled for ethnicity, in addition to all the covariates noted in footnote a.

*P > .05; **P > .01.

Additionally, social desirability bias did not influence the association between discrimination and chronic conditions. For the total sample, social desirability bias was not significantly associated with chronic conditions (b=.0009) and did not reduce the parameter estimates for everyday discrimination (b=.27). Across the subgroups, social desirability bias was associated with chronic conditions only for Filipinos (b=−.05), but adding it to the model does not change our inference about everyday discrimination for Filipinos or any other group.

Table 3 ▶ shows the associations between everyday discrimination and the disaggregated chronic conditions. For cardiovascular conditions, discrimination was a significant predictor for the total sample (odds ratio [OR] = 1.7), Vietnamese (OR = 1.9), and Chinese (OR = 1.7). Discrimination was nonsignificant for Filipinos, but the parameter estimate was in the expected direction (OR = 1.2). For respiratory conditions, discrimination was a significant predictor for the total sample (OR = 1.3), Vietnamese (OR = 1.5), and Filipinos (OR = 2.0); it was nonsignificant, but the estimate was in the expected direction for Chinese (OR = 1.2). For pain conditions, discrimination was a significant predictor for the total sample (OR=1.7), Vietnamese (OR=2.1), and Filipinos (OR = 1.5); it was nonsignificant but the estimate was in the expected direction for Chinese (OR = 1.3). For other health conditions, discrimination was not a significant predictor.

TABLE 3—

Association Between Self-Reported Everyday Discrimination and Categories of Chronic Health Conditions, by Ethnicity: National Latino and Asian American Study, 2002–2003

	All Asian and Pacific Islander Americans (n = 2095)		Vietnamese Americans (n = 520)		Filipino Americans (n = 508)		Chinese Americans (n = 600)
	Model 1, OR (95% CI)a,b	Model 2, OR (95% CI)a,b	Model 1, OR (95% CI)a	Model 2, OR (95% CI)a	Model 1, OR (95% CI)a	Model 2, OR (95% CI)a	Model 1, OR (95% CI)a	Model 2, OR (95% CI)a
Cardiovascular conditions
Everyday discrimination	1.69 (1.20, 2.39)**	1.69 (1.21, 2.38)**	1.85 (1.07, 3.18)*	1.87 (1.03, 3.39)*	1.19 (0.88, 1.60)	1.16 (0.88, 1.52)	1.65 (1.03, 2.65)*	1.69 (1.00, 2.83)*
Social desirability bias		1.02 (0.09, 1.12)		1.07 (0.93, 1.23)		0.89 (0.76, 1.04)		1.07 (0.89, 1.27)
Respiratory conditions
Everyday discrimination	1.37 (1.18, 1.58)***	1.37 (1.18, 1.58)***	1.47 (1.04, 2.08)*	1.45 (1.01, 2.06)*	2.06 (1.63, 2.61)**	2.02 (1.61, 2.54)**	1.17 (0.84, 1.64)	1.21 (0.87, 1.67)
Social desirability bias		1.00 (0.93, 1.06)		1.06 (0.92, 1.22)		0.99 (0.90, 1.09)		1.07 (0.97, 1.18)
Pain conditions
Everyday discrimination	1.69 (1.45, 1.98)***	1.71 (1.46, 1.99)***	2.11 (1.34, 3.33)**	2.12 (1.33, 3.38)**	1.50 (1.12, 2.01)**	1.51 (1.15, 1.99)**	1.23 (0.90, 1.67)	1.25 (0.92, 1.70)
Social desirability bias		0.98 (0.92, 1.04)		0.96 (0.87, 1.06)		0.9 (0.80, 1.01)		1.04 (0.95, 1.14)
Other conditions
Everyday discrimination	0.88 (0.63, 1.24)	0.89 (0.63, 1.25)	1.12 (0.57, 2.21)	1.13 (0.57, 2.23)	1.18 (0.85, 1.63)	1.18 (0.86, 1.63)	0.57 (0.27, 1.20)	0.54 (0.25, 1.17)
Social desirability bias		0.98 (0.92, 1.04)		1.04 (0.88, 1.24)		1.07 (0.93, 1.22)		0.92 (0.72, 1.18)

Note. OR = odds ratio; CI = confidence interval. Everyday discrimination was determined using a 9-item survey, adopted from the Detroit Area Study, measuring perceptions of chronic and routine unfair treatment (scale = 1–6); see “Methods” section. Social desirability was measured with a 10-item scale indicating the tendency of people to respond to surveys in ways that present a favorable impression of themselves. Model 1 included only everyday discrimination, and model 2 included everyday discrimination and social desirability bias.

^aAnalyses controlled for age, gender, per capita income, education, employment, marital status, nativity, language proficiency, region, and same household. Discrimination was modeled as a continuous variable in all models. Analyses are weighted to be nationally representative.

^bAnalysis controlled for ethnicity, in addition to all covariates noted in footnote a.

*P > .05; **P > .01; ***P > .001.

We further examined individual chronic conditions (e.g., headaches and backaches separately). The results from these analyses were consistent with those reported for the 4 categories (Table 3 ▶). We show the results for the 4 categories (cardiovascular, pain, respiratory, and other) rather than for the individual conditions (e.g., stroke, asthma) to be more parsimonious. Social desirability bias was not associated with any health condition for any group and did not influence the association between discrimination and health in any analysis.

We performed additional analyses (available upon request from the authors) to examine the robustness of our models. First, we examined an alternative specification of discrimination using a 3-item measure derived from Vega et al.57 Models using that indicator were generally consistent with those reported here. Second, everyday discrimination was modeled continuously to be consistent with how other studies have used this measure; however, it is not a truly continuous scale. We estimated models that treated discrimination as a binary variable (never vs any) and obtained similar results. Third, supplemental analyses included percentage of Asians in a respondent’s census tract and urbanicity (rural vs urban), but neither factor influenced our findings.

DISCUSSION

To our knowledge, this is the first nationally representative study of self-reported discrimination and chronic health conditions among Asian Americans. Reports of everyday discrimination were associated with increased numbers of chronic health conditions, after we controlled for social desirability bias, age, gender, language proficiency, nativity, region, per capita income, education, employment, and marital status.

The chronic conditions were disaggregated into 4 major categories. Analyses of these categories were consistent with the notion that self-reported discrimination represents an important psychosocial stressor. The “psychobiological reactivity” perspective suggests that the cardiovascular, endocrine, and immune responses to stressors are closely interlinked.58^,59 Consistent with this perspective, we found that reports of discrimination were associated with reports of cardiovascular conditions. Similarly, other studies have found relationships between self-reported everyday discrimination and blood pressure, hypertension, and cardiovascular reactivity.11^,60^–63

Our data also indicated 2 associations that deserve further investigation. First, everyday discrimination was associated with greater odds of respiratory conditions. Karlsen and Nazroo12 also reported that minorities experiencing verbal abuse or physical attacks were at increased risk of experiencing respiratory illness. Since the work of Holmes in the 1950s, research has suggested that stressors lead to susceptibility to tuberculosis and reactivation of latent infection.64 More recently, Cohen and others have demonstrated how stressful events can increase susceptibility to upper respiratory illness.65^–67 Our findings, although preliminary, suggest that discrimination may be a potential factor that contributes to vulnerability or reactivation of respiratory problems.

Second, everyday discrimination was also associated with indicators of pain (chronic back or neck problems, frequent or severe headaches, chronic pain, ulcers). Bowen-Reid and Harrell reported that racist experiences were associated with somatic complaints (e.g., headaches, nausea).13 Gee et al. reported that discrimination was associated with use of analgesics and tranquilizers, drugs often used to treat pain.41 More generally, studies have suggested that stressors can lead to the onset and aggravation of chronic pain and headaches.68^–70

Remarkably, a scale designed to measure everyday discrimination among African Americans in Detroit was a useful tool for assessing discrimination and health among Asian Americans nationwide. Asian Americans in our study reported slightly lower levels of everyday discrimination (mean = 1.9 on a 1–6 scale) than did African Americans in Detroit (mean = 2.3).44 However, this measure may not fully capture the types of unfair experiences relevant to Asian Americans, such as discrimination because of one’s accent and immigration status.34^,39^,71 That said, our findings suggest some commonality in the reporting of everyday discrimination. Increasing awareness of these common bonds may be important in helping to build coalitions.

There were some group differences in the reporting of discrimination, in health conditions, and in the association between discrimination and health conditions. These differences would be expected, given the heterogeneous history of Asian Americans in the United States. For example, historical discrimination against Chinese Americans was mainly through exclusion using laws that forbade citizenship, land ownership, and restricted immigration. However, historical discrimination against Filipinos included forced inclusion through the colonization of the Philippines by the United States. Vietnamese Americans face potential discrimination because of their status as refugees and the legacy of the Vietnam War. Obviously, the experiences of these groups are much more complex than can be captured in our study, but excellent discussions can be found elsewhere.72^–74

In our study, Filipinos reported the highest levels of discrimination, followed by Chinese and then Vietnamese. Two other studies found that Filipinos reported more discrimination than Chinese.33^,75 Cabesas et al. suggested that the greater reporting of discrimination among Filipinos results from their darker skin color or increased familiarity with discrimination because of colonization.75 Consistent with this observation, Filipinos in our sample were more likely than the other Asian ethnic groups to state that the main reason for discrimination was race, ancestry, or skin color. Further, the higher rates among Filipinos might result from socioeconomic differences, as some evidence suggests greater wage discrimination among Filipinos than among Chinese.76 A full investigation of how groups report discrimination is beyond the scope of this study but worthy of future research.

Although groups were generally similar in their reporting of health conditions, there were some differences in the association between discrimination and specific conditions. Discrimination was significantly associated with cardiovascular conditions for Vietnamese and Chinese. Discrimination was nonsignificant, but the parameter estimate was in the expected direction for Filipinos. A prior study using a larger sample of Filipino Americans found an association between discrimination and cardiovascular conditions for Filipinos,19 suggesting that our null findings might result from chance or low power. Discrimination was significantly associated with pain and respiratory conditions for Vietnamese and Filipinos, but not Chinese Americans. The estimate for Chinese was in the expected direction, which also suggests low power or chance findings. However, these differential associations might also indicate cultural, structural, and historical differences in the reporting of discrimination and health.

Limitations

Reports of discrimination and health are subject to response factors, including social desirability bias. Social desirability bias might be associated with decreased reporting of discrimination and illness if participants wish to avoid expressing themselves as victims or to “save face.”51 Social desirability bias was correlated with decreased reporting of discrimination, but it was significantly associated with health in only 1 of 20 models, a finding likely due to chance alone. Most important, our inferences between self-reported discrimination and health conditions were not influenced by social desirability bias. However, there are no perfect measures of social desirability bias,77 and future research should include other personality factors (e.g., optimism) and response factors (e.g., memory). That said, ours is one of the few discrimination studies that have included social desirability bias.

There were a few other limitations. Health conditions were not verified by clinical diagnosis and do not represent incident cases. However, this measure of health conditions has been used in prior studies and correlated with clinical measures of morbidity.19^,54^,79 Further, the data are cross-sectional and cannot establish causal directions between measures. For example, although we presume that discrimination leads to illness, it might be that illness leads individuals to contact the health system and subsequently report discrimination. Several longitudinal studies have found that discrimination predicts health and not the converse,80^,81 but more longitudinal research is needed. On balance, our study has several strengths, including use of a timely and nationally representative sample, disaggregation of several Asian ethnic groups, and use of standard measures.

Conclusions

Although Asian Americans are sometimes portrayed as a “model minority” who are presumed to no longer experience discrimination, our study shows that they do report discrimination, and these reports are associated with health. Future research should investigate other specific Asian American groups (e.g., Thai), assess other forms of discrimination (e.g., institutional), and examine the practices (e.g., multicultural curriculum in elementary school) and policies (e.g., fair housing legislation) that prevent discrimination from occurring. Recent resolutions by the American Public Health Association to study and intervene on racism as a fundamental cause of ethnic disparities provide a firm ground from which to develop such policies.82