Objectives. We investigated whether the conventional Spanish translation of the self-rated health survey question helps explain why Latinos' self-rated health is worse than Whites' despite more objective health measures showing them to be as healthy as or healthier than are Whites.

Methods. We analyzed the relationship between language of interview and self-rated health in the Chicago Community Adult Health Study (2001–2003) and the 2003 Behavioral Risk Factor Surveillance System.

Results. Being interviewed in Spanish was associated with significantly higher odds of rating health as fair or poor in both data sets. Moreover, adjusting for language of interview substantially reduced the gap between Whites and Latinos. Spanish-language interviewees were more likely to rate their health as fair (regular in Spanish) than as any other choice, and this preference was strongest when compared with categories representing better health (good, very good, and excellent).

Conclusions. Our findings suggest that translation of the English word “fair” to regular induces Spanish-language respondents to report poorer health than they would in English. Self-rated health should be interpreted with caution, especially in racial/ethnic comparisons, and research should explore alternative translations.

For more than 2 decades, researchers have been intrigued by studies indicating that Latinos experience health outcomes that are equal to or better than those of Whites, despite having lower socioeconomic status.18 One important exception to this apparent Latino health advantage is self-rated health; that is, Latinos’ assessments of their overall health status are lower than those of Whites and below what would be expected from more objective measures.815 Understanding this puzzling side of Latino population health is critical to advancing research on racial/ethnic disparities in health and addressing their root causes.16

Some scholars hypothesize that Latinos’ relatively poor self-rated health reflects a cultural orientation to somatize psychosocial or emotional distress as physical health conditions or to avoid boasting about health.11,12,17,18 Accordingly, some studies rely on measures of language use to tap into the cultural and cognitive factors that might explain Latinos’ perception of their own health. Such studies show that Latino respondents with greater proficiency in Spanish report lower levels of self-rated health than do Latinos with greater English proficiency1113,17,1922 and that controlling for language use reduces Latino–White disparities in self-rated health.1113,19

The explanation for Latinos' relatively low levels of self-rated health may be that the usual Spanish translation of the response categories to the self-rated health question induces Spanish-language respondents to report poorer health than they would if they were responding to the question in English. Although this hypothesis does not preclude other explanations, it is a parsimonious and testable proposition. The response categories for the self-rated health question are conventionally translated into Spanish as excelente (excellent), muy buena (very good), buena (good), regular (fair), and mala (poor). Because the Spanish word regular connotes more positive meanings than fair does in English, the translation of this response option may downwardly bias estimates of Latino self-reported health status. Although other scholars have suggested similar explanations,1113,17 none, to our knowledge, have empirically examined how language of interview shapes answers to the fair response category.

We analyzed ethnic and linguistic differences in self-rated health with data from 2 surveys with complementary strengths: the Behavioral Risk Factor Surveillance System (BRFSS) and the Chicago Community Adult Health Study (CCAHS). The CCAHS collected a multilevel, stratified probability sample of 3105 adults aged 18 years or older living in Chicago, Illinois.23 The data were obtained between May 2001 and March 2003 via face-to-face interviews with 1 individual per household; interviews were conducted in English, Polish, or Spanish, with a response rate of 71.8%. The data were weighted to match the age, race/ethnicity, and gender distributions of the 2000 Census population estimates for the city of Chicago. The BRFSS is an ongoing cross-sectional health survey of noninstitutionalized individuals aged 18 years or older conducted at the state level via telephone, with random-digit-dialing sampling.24 We used the 2003 BRFSS data from the 20 states that offered the survey in English or Spanish: Arizona, California, Colorado, Connecticut, Florida, Illinois, Indiana, Kansas, Massachusetts, Nebraska, Nevada, New Jersey, New Mexico, New York, North Carolina, Oklahoma, Rhode Island, Texas, Utah, and Washington.

We chose these 2 data sets for substantive and practical reasons. We first tested our hypothesis in the CCAHS, which had a wider range of relevant controls, and then replicated the analysis on the BRFSS because it was a large, national data set used in previous studies of Latino self-rated health.20,22,24

We also analyzed the 2002 National Health Interview Survey (NHIS) but did not report the results because of concerns about its comparability with other data sets vis-à-vis language of interview and Latino self-rated health. Previous research found (and we confirmed) that Latinos’ assessments of their overall health were more positive in the NHIS than in the BRFSS and other data sets.25,26 For example, a comparison of the 1997 NHIS and BRFSS found that the unadjusted risk ratio of being in poor or fair health for Latinos compared with Whites was only 1.15 in the NHIS, far below the 1.80 risk ratio reported for the BRFSS.25 Thus, the problem that motivated our study—the relatively low assessments that Latinos give to their health—was much less evident in the NHIS, which, although puzzling and worthy of further investigation, made it less relevant for our research. Also, the NHIS interview protocol was unique in that respondents could decide for each question whether to answer in English or Spanish. Although we could identify which NHIS respondents selected a mixture of English- and Spanish-language questions, determining which respondents answered any particular question in Spanish or English was not possible.

Measures

We constructed a set of nearly equivalent measures from the CCAHS and BRFSS and used them in all models. We constructed 2 dummy variables for language of interview, 1 for Spanish and a second for being interviewed in another language. (In the CCAHS, the only other language available was Polish, but the BRFSS offered a wider range of languages, which varied from state to state and were not explicitly documented in the data set.) Race/ethnicity was measured comparably in the 2 data sets with 4 categories: Latino, non-Latino White, non-Latino Black, and non-Latino other race. We also controlled in all models for comparable measures of gender, age, and household income. Educational attainment was measured as years of education completed in the CCAHS, grouped in relation to attainment of major degrees (high school and college), and as highest degree completed in the BRFSS.

In some models, we controlled for an additional set of health risk factors and conditions that were nearly comparable across the 2 data sets: reports of being unemployed, without health insurance, a smoker, or a heavy drinker (≥ 5 drinks on at least 1 occasion during the past 30 days). These models also controlled for respondents' reports of ever being diagnosed with asthma, hypertension, diabetes, or arthritis. We controlled for categories of body mass index, calculated from height and weight measurements that were collected anthropometrically in the CCAHS but via self-report in the BRFSS. When a variable had missing data, we treated missing as a category (all of our covariates were categorical) and included a dummy variable, but we did not report these coefficients in the tables.

In supplemental analyses, we controlled for immigration-related factors measured in the CCAHS but not in the BRFSS: nativity status (foreign born), age at migration (whether a foreign-born respondent immigrated when younger or older than 17 years), and a dichotomous measure of language use (indicating whether the respondent spoke and read only in Spanish, spoke only Spanish at home, and spoke only Spanish with friends). We also included indicators of Latino subgroups (Mexicans, Puerto Ricans, and other Latinos).

Analytical Strategy

Our analysis proceeded in 2 stages. First, we replicated and extended previous work on Latino self-rated health disparities by comparing Latinos and other racial/ethnic groups in logistic regressions on their odds of reporting poor or fair health. Our first model examined racial/ethnic disparities in self-rated health, with adjustments for gender, age, education, and income. We next controlled for language of interview, to see how much it changed estimates of Latino–White disparities in self-rated health, and then added controls for additional risk factors and conditions that could confound the association between language of interview and self-reports.

In the second stage of analysis, we tested the hypothesis that persons taking the interview in Spanish are more likely to rate their health as fair (regular), all else being equal; we used multinomial logistic regression to predict the odds of reporting health as regular compared with other categories. All of our analyses were weighted to account for each survey's complex design and were performed with Stata version 10/SE (StataCorp LP, College Station, TX).

Weighted descriptive statistics on the CCAHS and BRFSS are shown in Table 1 for all respondents and for the Latino subsamples from both data sets. The distribution of self-rated health across response categories was fairly similar in the 2 data sets for both the full sample and the Latino sample. For example, the unadjusted odds ratio (OR; not shown) of being in poor or fair health for Latinos compared with Whites was 2.95 (SE = 0.52; P < .001) in the CCAHS and 2.24 (SE = 0.09; P < .001) in the BRFSS. The proportion of Latinos who were administered the questionnaire in Spanish was almost identical in the 2 data sets (0.44 in the CCHAS and 0.43 in the BRFSS). The CCAHS had higher proportions of Latinos and Blacks and a smaller proportion of Whites than did the BRFSS, and CCAHS respondents had lower levels of education than did those in the BRFSS, but otherwise, the 2 data sets were fairly comparable on most variables.

Table

TABLE 1 Characteristics of Study Participants: Chicago Community Adult Health Study, 2001–2003, and Behavioral Risk Factor Surveillance System, 2003

TABLE 1 Characteristics of Study Participants: Chicago Community Adult Health Study, 2001–2003, and Behavioral Risk Factor Surveillance System, 2003

CCAHS
BRFSS
Full Sample (Unweighted n = 3105), Weighted Proportion (SE)Latinos (Unweighted n = 804), Weighted Proportion (SE)Full Sample (Unweighted n = 124 885), Weighted Proportion (SE)Latinos (Unweighted n = 12 025), Weighted Proportion (SE)
Self-rated health
    Poor0.03 (0.00)0.02 (0.01)0.04 (0.00)0.05 (0.00)
    Fair0.13 (0.01)0.20 (0.02)0.12 (0.00)0.20 (0.01)
    Good0.30 (0.01)0.32 (0.02)0.30 (0.00)0.36 (0.01)
    Very good0.36 (0.01)0.28 (0.02)0.32 (0.00)0.22 (0.01)
    Excellent0.18 (0.01)0.18 (0.02)0.22 (0.00)0.16 (0.01)
Male0.47 (0.01)0.49 (0.02)0.49 (0.00)0.49 (0.01)
Age, y
    18–290.28 (0.01)0.34 (0.02)0.22 (0.00)0.33 (0.01)
    30–390.23 (0.01)0.28 (0.02)0.19 (0.00)0.26 (0.01)
    40–490.19 (0.01)0.16 (0.01)0.20 (0.00)0.19 (0.01)
    50–590.13 (0.01)0.10 (0.01)0.16 (0.00)0.12 (0.01)
    60–690.09 (0.01)0.06 (0.01)0.10 (0.00)0.05 (0.00)
    ≥ 700.09 (0.01)0.05 (0.01)0.12 (0.00)0.05 (0.00)
Race/ethnicity
    Latino0.26 (0.02)0.19 (0.00)
    White0.38 (0.02)0.65 (0.00)
    Black0.32 (0.02)0.08 (0.00)
    Othera0.04 (0.01)0.07 (0.00)
Educationb
    < High school0.23 (0.01)0.45 (0.02)0.13 (0.00)0.36 (0.01)
    High school0.24 (0.01)0.25 (0.02)0.28 (0.00)0.29 (0.01)
    Some college0.25 (0.01)0.20 (0.02)0.27 (0.00)0.20 (0.01)
    ≥ College degree0.28 (0.02)0.10 (0.01)0.32 (0.00)0.14 (0.01)
Income, $
    <15 0000.19 (0.01)0.22 (0.02)0.10 (0.00)0.20 (0.01)
    15 000–24 9990.12 (0.01)0.15 (0.02)0.15 (0.00)0.24 (0.01)
    25 000–34 9990.10 (0.01)0.11 (0.01)0.12 (0.00)0.14 (0.01)
    35 000–49 9990.13 (0.01)0.16 (0.02)0.14 (0.00)0.11 (0.01)
    ≥ 50 0000.26 (0.01)0.18 (0.02)0.36 (0.00)0.17 (0.01)
    Missing0.19 (0.01)0.18 (0.02)0.13 (0.00)0.14 (0.01)
Language of interview
    Spanish0.12 (0.01)0.44 (0.03)0.08 (0.00)0.43 (0.01)
    Otherc0.04 (0.01)0.00 (0.00)0.00 (0.00)
Other economic factors
    Uninsured0.20 (0.01)0.32 (0.02)0.17 (0.00)0.36 (0.01)
    Unemployed0.36 (0.01)0.34 (0.02)0.39 (0.00)0.37 (0.01)
BMI, kg/m2
    < 250.35 (0.01)0.23 (0.02)0.39 (0.00)0.31 (0.01)
    25–< 300.32 (0.01)0.38 (0.02)0.21 (0.00)0.23 (0.01)
    ≥ 300.33 (0.01)0.38 (0.02)0.35 (0.00)0.35 (0.01)
    Missing0.05 (0.00)0.11 (0.00)
Health behaviors
    Heavy drinkerd0.22 (0.01)0.20 (0.02)0.16 (0.00)0.18 (0.01)
    Current smoker0.25 (0.01)0.18 (0.02)0.21 (0.00)0.18 (0.01)
Health conditions, ever diagnosed
    Asthma0.12 (0.01)0.10 (0.01)0.12 (0.00)0.09 (0.00)
    Hypertension0.26 (0.01)0.21 (0.02)0.26 (0.00)0.20 (0.01)
    Diabetes0.08 (0.01)0.10 (0.01)0.08 (0.00)0.09 (0.00)
    Arthritis0.18 (0.01)0.13 (0.01)0.24 (0.00)0.14 (0.01)

Note. BMI = body mass index; BRFSS = Behavioral Risk Factor Surveillance System; CCAHS = Chicago Community Adult Health Study. BRFSS sample was restricted to persons in the 20 states that conducted interviews in English and Spanish and with valid self-rated health data. Ellipses indicate not applicable.

aIncluded multiple race.

bEducation categories shown are for the BRFSS; equivalent categories for the CCAHS were in years: < 12, 12, 13–15, and ≥ 16.

cIn the CCAHS, Polish was the other language of interview.

dDefined as having had 5 or more drinks on at least 1 occasion during the past 30 days.

We also compared the weighted Latino samples in the BRSS and CCAHS to the 2002 national Latino population (results not shown) and found them to be very similar in gender and age composition.27,28 The share of Mexicans in the CCAHS Latino sample (67.0%) was comparable to that in the national population (66.9%), but the CCAHS had a greater share of Puerto Ricans (15.6% vs 8.6%) and foreign-born Latinos (55.9% vs 40.2%).29

Table 2 shows logistic regression models predicting poor or fair self-rated health. Model 1 results indicated that the odds of self-reporting health as poor or fair were 86% higher for Latinos than Whites in the CCAHS and 66% higher in the BRFSS, after adjustment for gender, age, education, and income. In model 2, being interviewed in Spanish was still associated with higher odds of reporting poor or fair health; this adjustment for language of interview narrowed the estimated Latino–White gap considerably, to a point where the coefficients were almost identical in the 2 data sets (and the CCAHS coefficient was no longer significant). In model 3 we added controls for other health risks and conditions that could confound the association between language of interview and self-rated health. Interestingly, we found that the Spanish-language effect became larger in both data sets with adjustment for these risk factors, and supplemental analysis indicated that the main reason for this increase was that Spanish interviewees reported fewer chronic health conditions.

Table

TABLE 2 Odds Ratios of Poor or Fair Self-Rated Health: Chicago Community Adult Health Study, 2001–2003, and Behavioral Risk Factor Surveillance System, 2003

TABLE 2 Odds Ratios of Poor or Fair Self-Rated Health: Chicago Community Adult Health Study, 2001–2003, and Behavioral Risk Factor Surveillance System, 2003

CCAHS
BRFSS
CovariatesModel 1, OR (SE)Model 2, OR (SE)Model 3, OR (SE)Model 1, OR (SE)Model 2, OR (SE)Model 3, OR (SE)
Race/ethnicity
    White (Ref)1.001.001.001.001.001.00
    Latino1.86** (0.34)1.30 (0.30)1.20 (0.29)1.66** (0.08)1.31** (0.09)1.40** (0.10)
    Black1.48* (0.25)1.60** (0.29)1.30 (0.24)1.45** (0.09)1.48** (0.09)1.22** (0.08)
 Othera0.65 (0.34)0.69 (0.37)0.81 (0.43)1.37** (0.11)1.38** (0.11)1.35** (0.11)
Male0.71** (0.09)0.69** (0.09)0.89 (0.12)0.99 (0.03)0.98** (0.03)1.19** (0.04)
Age, y
    18–29 (Ref)1.001.001.001.001.001.00
    30–392.16** (0.43)2.05** (0.41)1.79** (0.38)1.43** (0.10)1.39** (0.10)1.24** (0.09)
    40–493.06** (0.65)2.91** (0.63)1.96** (0.45)2.40** (0.16)2.38** (0.16)1.80** (0.13)
    50–594.07** (0.90)3.93** (0.88)1.76* (0.45)3.85** (0.26)3.84** (0.26)2.16** (0.16)
    60–693.83** (0.95)3.89** (0.95)1.27 (0.35)4.12** (0.28)4.15** (0.28)1.71** (0.13)
    ≥ 703.33** (0.85)3.42** (0.87)0.97 (0.28)5.03** (0.33)5.13** (0.34)2.10** (0.17)
Educationb
    < High school5.27** (1.29)4.54** (1.13)3.06** (0.82)3.68** (0.21)3.43** (0.20)2.54** (0.16)
    High school2.76** (0.65)2.70** (0.64)2.23** (0.55)1.87** (0.09)1.89** (0.09)1.53** (0.08)
    Some college2.50** (0.59)2.51** (0.60)2.00** (0.49)1.60** (0.08)1.62** (0.08)1.35** (0.07)
    ≥ College degree (Ref)1.001.001.001.001.001.00
Income, $
    < 15 0003.75** (0.92)3.62** (0.89)2.38** (0.62)4.89** (0.31)4.72** (0.30)2.11** (0.15)
    15 000–24 9993.12** (0.79)2.96** (0.75)2.16** (0.58)3.13** (0.18)3.03** (0.17)1.60** (0.10)
    25 000–34 9992.11** (0.59)1.95* (0.54)1.62 (0.47)2.12** (0.13)2.09** (0.13)1.20** (0.08)
    35 000–49 9991.76* (0.46)1.69* (0.45)1.59 (0.43)1.59** (0.10)1.59** (0.10)0.73** (0.05)
    ≥ 50 000 (Ref)1.001.001.001.001.001.00
    Missing2.35** (0.52)2.20** (0.49)1.85* (0.45)2.46** (0.14)2.39** (0.14)1.37** (0.09)
Language of interview
    English (Ref)1.001.001.001.00
    Spanish2.35** (0.51)3.52** (0.80)1.69** (0.14)2.39** (0.22)
    Otherc1.42 (0.48)1.92 (0.64)1.27** (0.48)1.84 (0.74)
Other economic factors
    Uninsured1.18 (0.19)1.20** (0.06)
    Unemployed2.02** (0.30)1.85** (0.08)
BMI, kg/m2
    < 25 (Ref)1.001.00
    25–< 300.77 (0.14)0.96 (0.04)
    ≥ 301.02 (0.19)1.45** (0.07)
Health behaviorsd
    Heavy drinker0.83 (0.15)1.09 (0.07)
    Current smoker1.49* (0.23)1.57** (0.07)
Health conditions, ever diagnosed
    Asthma1.99** (0.37)1.91** (0.09)
    Hypertension2.73** (0.46)1.78** (0.07)
    Diabetes2.31** (0.45)3.04** (0.16)
    Arthritis1.95** (0.33)2.25** (0.09)

Note. BMI = body mass index; BRFSS = Behavioral Risk Factor Surveillance System; CCAHS = Chicago Community Adult Health Study; OR = odds ratio. BRFSS sample restricted to persons in the 20 states that conducted interviews in English and Spanish and with valid self-rated health data. Ellipses indicate not available.

aIncluded multiple race.

bEducation categories shown are for the BRFSS; equivalent categories for the CCAHS were in years: < 12, 12, 13–15, and ≥ 16.

cIn the CCAHS, Polish was the other language of interview.

dDefined as having had 5 or more drinks on at least 1 occasion during the past 30 days.

*P < .05; **P < .01.

The results from Table 2 generally replicated findings from previous studies regarding the role of language of interview in reducing self-rated health disparities between Latinos and Whites. However, most previous studies did not investigate whether the translation of the response categories from English to Spanish resulted in a clustering of responses in the regular category among Spanish speakers after multiple controls. To test this hypothesis, we used multinomial logistic regression to estimate the log odds of reporting fair (regular) health compared with each of the other categories. Table 3 presents the relative risk ratios associated with interviewing in Spanish on the odds of reporting health as regular compared with the other options. Only the Spanish-language coefficients are displayed in Table 3. Model 1 adjusted for the core set of covariates (Table 2, model 2), and model 2 added controls for the additional health risk factors and conditions (Table 2, model 3).

Table

TABLE 3 Multinomial Logistic Regression Models of Effects of Language of Interview on Response Categories of Self-Rated Health Question: Chicago Community Adult Health Study, 2001–2003, and Behavioral Risk Factor Surveillance System, 2003

TABLE 3 Multinomial Logistic Regression Models of Effects of Language of Interview on Response Categories of Self-Rated Health Question: Chicago Community Adult Health Study, 2001–2003, and Behavioral Risk Factor Surveillance System, 2003

CCAHS
BRFSS
Self-Rated Health Response ComparisonModel 1 RRR (SE)Model 2 RRR (SE)Model 1 RRR (SE)Model 2 RRR (SE)
Fair versus poor health2.49 (1.26)1.72 (0.86)2.48** (0.43)1.77** (0.32)
Fair versus good health2.63** (0.68)3.41** (0.89)1.49** (0.14)1.82** (0.18)
Fair versus very good health2.88** (0.80)4.10** (1.21)3.65** (0.44)4.76** (0.60)
Fair versus excellent health2.04* (0.61)3.13** (1.00)2.65** (0.34)3.93** (0.55)

Note. BRFSS = Behavioral Risk Factor Surveillance System; CCAHS = Chicago Community Adult Health Study; RRR = relative risk ratio. For each contrasting set of self-rated health response categories, the odds ratio compares the odds of being in fair health for Spanish-language respondents to English-language respondents (the reference group). RRRs in Model 1 were generated from multinomial logistic regression models that controlled for age, gender, race/ethnicity, education, income, and whether the survey was administered in a language other than English or Spanish. Model 2 also controlled for employment status, health insurance status, body mass index, health behaviors, and chronic health conditions.

*P < .05; **P < .01.

The results were consistent with the translation hypothesis. Respondents who interviewed in Spanish were more likely to rate their health as regular than any other category, even after adjustment for the full set of covariates in model 2, and in all but 1 case (fair vs poor health in the CCAHS), this association was significant. The tendency for Spanish interviewees to rate their health as fair (regular) was most pronounced in comparison with the odds of selecting categories that represent better health (good, very good, or excellent). In addition, the Latino–White gap in the odds of reporting fair health was reduced by controlling for language of interview (results not shown).

Supplemental analysis (not shown) assessed whether the association between being interviewed in Spanish and reporting health as regular was confounded by immigrant status and other language use measures available in the CCAHS. Controlling for nativity status, age at migration, and language use did not alter our results, and none of these measures were significantly associated with self-rated health after controlling for the covariates in model 2, Table 3. We also examined whether the language-of-interview effect on self-rated health varied by nativity, age at migration, or Latino subgroup by including appropriate interaction terms in our models, but we found none to be statistically significant.

To help interpret the results from the multinomial models, we used the coefficients from model 1 in Table 3 to calculate the predicted probability of responding to each self-rated health category and graphed these probabilities in Figure 1, highlighting the comparison between Latinos who responded to the interview in English and those who used Spanish. The predicted probabilities that Latino–English and Latino–Spanish interviewees would rate their health as fair (regular) were comparable across the 2 data sets: the probability of selecting fair for Latino–English interviewees was 0.10 in the CCAHS and 0.12 in the BRFSS; the probability of selecting regular for Latino–Spanish interviewees was 0.22 in the CCAHS and 0.23 in the BRFSS. In both data sets, these differences were statistically significant (P < .05). The predicted probabilities for the other categories did not match as closely across the 2 data sets; the main difference was that the Latino–Spanish respondents in the CCAHS were more likely than were their counterparts in the BRFSS to rate their health as excellent or very good and less likely to rate it as good.

Our analysis of the CCAHS and the 2003 BRFSS supports the hypothesis that language of interview is a critical source of variation in self-rated health among Latinos. Consistent with previous studies,1113,17,1922 we found that Latinos who interviewed in Spanish reported their health as worse than did Latinos who interviewed in English and that controlling for language of interview significantly reduced the differences in self-rated health between Latinos and Whites in both data sets.1113,19

Moving beyond previous studies, we examined whether the translation of 1 response category—fair or regular—downwardly biased the distribution of self-rated health among Latinos who responded to the survey in Spanish. Our multinomial analysis demonstrated that in both samples, people who interviewed in Spanish had higher odds of reporting regular health than had those interviewed in English of reporting fair health, after adjustment for a host of possible confounders, including health status, education, income, and immigration-related characteristics. The propensity for Spanish-language interviewees to select regular was most pronounced when it was compared with categories that reflected better levels of health—a finding consistent with the hypothesis that translation problems induce some Spanish-language respondents to answer regular when they actually mean to rate their health in more positive terms than conveyed by the English term fair.

Although our findings regarding the effect of language of interview on self-rated health are compelling, they do not preclude other explanations for self-reports of poorer health among Latinos than Whites. Some scholars have suggested that a traditional cultural orientation among Latinos leads them to downgrade their health assessments or somatize psychosocial symptoms, either of which could lead first-generation immigrants and Spanish speakers to report poorer health on surveys.11,12,17,18 Other scholars have suggested that immigrants may be more likely to downgrade their ratings because their health has suffered from the stress of migration and barriers to social integration in the receiving country or because immigrants’ reference groups shift through the process of migration and settlement.14,21

Although we cannot rule out these explanations, in supplemental analysis of the CCAHS, we took several steps to ensure that the language-of-interview effect was not confounded by other differences between immigrants and nonimmigrants. We controlled for nativity status, age at migration, and language use and found that that the language-of-interview effect remained strong and significant. However, language of interview was highly correlated with both nativity and our measure of language use, making it difficult to disentangle the effects of survey translation from other social factors that may differentiate Spanish- from English-language interviewees.

Another explanation that has been offered for Latinos’ poorer health self-assessments is that self-rated health may be driven by limited access to life opportunities and resources, as reflected in their lower levels of education.12,3032 Indeed, education had strong and significant effects on the odds of being in poor or fair health. Moreover, supplemental analysis revealed that controlling for education reduced the Latino–White gap in the odds of reporting poor or fair health almost as much as language of interview in both data sets.

Limitations

We were unable to replicate our findings in our analysis of the 2002 NHIS. In our supplemental analysis of the NHIS, we found that the Latino–White self-rated health disparity was much smaller (OR = 1.17; SE = 0.09, in a model comparable to model 1 from Table 2), and we detected no effect from being interviewed in Spanish (OR = 0.98; SE = 0.11, in a model comparable to model 2 from Table 2). We also compared descriptive statistics on Latinos from the NHIS to those from the BRFSS and CCAHS to detect any differences that might explain why the distribution of Latino self-rated health was higher in the NHIS than in the other data sets. The main differences we found were that the NHIS had a lower weighted proportion of Latinos in general (0.10 [SE = 0.00] in the NHIS; 0.19 [SE = 0.00] in the BRFSS; and 0.26 [SE = 0.02] in the CCAHS) and of Latinos who were interviewed only in Spanish (0.25 [SE = 0.01] in the NHIS, 0.43 [SE = 0.01] in the BRFSS, and 0.44 [SE = 0.03] in the CCAHS). An additional 12% of the Latinos in the NHIS were interviewed in both English and Spanish, but it was impossible to determine which language was used for the self-rated health question.

Thus, one reason that Latinos in the NHIS gave more positive assessments of their health could be that fewer of them were interviewed in Spanish. However, when we replicated the multinomial logit models from Table 3 on the NHIS, we did not find any significant effects of interviewing in Spanish, so it is not clear that increasing the number of NHIS respondents who were interviewed in Spanish would change the distribution of Latino self-rated health. Still, with a smaller proportion of NHIS Latinos interviewing only in Spanish, it is likely that NHIS Spanish-language interviewees were selectively different from their counterparts in the other data sets, and this is a major reason that we did not report these results alongside those from the BRFSS and CCAHS.

Another reason that the distribution of self-rated health may vary across surveys is that it could be affected by the ordering of the health question vis-à-vis questions on specific health conditions. One study found that Spanish-language interviewees reported worse self-rated health when this question was asked before inquiring about chronic health conditions.33 In the BRFSS and NHIS, the self-rated health question was asked before questions on specific health conditions; the order was reversed in the CCAHS.

A final noteworthy limitation is that we did not disaggregate our estimates of the Latino–White self-rated health gap by Latino subgroup because such data were not available in the BRFSS, and some subgroups were small in the CCAHS. The dominant Latino subgroup in the CCAHS was Mexican, and all of the main findings held up for Mexicans when we disaggregated our analysis. In this sense, our CCAHS sample was comparable to those of other studies of self-rated health that report Latino subgroup characteristics, in which Mexicans tend to be the majority.9,11,12,17,21

Conclusions

The most important implication of our findings is that great caution must be taken in analyzing self-rated health data when the standard translation of the question from English to Spanish is used. In the future, researchers might consider alternative Spanish translations of the English word fair, such as más o menos, which have more negative connotations than regular. Interestingly, this translation problem may not be limited to Spanish: our multinomial models also revealed that respondents who interviewed in other languages were more likely (albeit not significantly) to report being in fair health. Perhaps the English adjective fair (in the sense of so-so) is complicated to translate into most other languages, especially if done somewhat literally.13,34

Further research—preferably with an experimental component—is necessary to determine whether differing interpretations of the self-rated health question in Spanish (or other languages) versus English are attributable to its translation, to culturally mediated understanding of this item, to immigrant social integration, to access to life opportunities, or to a combination of these factors. One useful experiment would randomly assign a group of respondents interviewed in Spanish to be asked questions with an alternative translation of fair and compare their responses with those of participants who were given the conventional translation.

Qualitative research is also needed on how Latinos interpret the meaning of the self-rated health question. In particular, cognitive interview techniques—asking respondents to think aloud about each question and their process for arriving at an answer—would be useful in shedding light on the role of language in shaping interpretation of self-rated health response options.35

Because this measure is an important indicator of population health and is routinely relied on in predicting health outcomes, a better understanding of the factors causing disparities in self-rated health is needed. Until this is achieved, scholars drawing scientific, policy, or clinical implications from this measure—especially with regard to comparisons across ethnic groups—should do so with caution and awareness of this potential linguistic bias.

Acknowledgments

This work was supported by a grant from the Robert Wood Johnson Foundation via the Center for Society and Health, Harvard School of Public Health (to E. A. Viruell-Fuentes) and by the National Institute of Child Health and Human Development of (grants P50HD38986 and R01HD050467, to J. D. Morenoff and J. S. House).

We are grateful to Beth Taylor and Juan Albertorio at the National Center for Health Statistics for their assistance with the NHIS and to William P. Bartoli at the Centers for Disease Control for his assistance with the BRFSS.

Note. The contents of this article are the sole responsibility of the authors.

Human Participant Protection

The Chicago Community Adult Health Study was approved by the University of Michigan institutional review board. The analyses presented here were also approved by the institutional review boards at Harvard University and the University of Illinois at Urbana–Champaign.

References

1. Franzini L, Ribble JC, Keddie AM. Understanding the Hispanic paradox. Ethn Dis. 2001;11(3):496518. MedlineGoogle Scholar
2. Escarce JJ, Morales LS, Rumbaut RG. The health status and health behaviors of Hispanics. In: , Tienda M, Mitchell F, eds. Hispanics and the Future of America. Washington, DC: National Academies Press; 2006:362409. Google Scholar
3. Palloni A, Arias A. Paradox lost: explaining the Hispanic adult mortality advantage. Demography. 2004;41(3):385415. Crossref, MedlineGoogle Scholar
4. Palloni A, Morenoff JD. Interpreting the paradoxical in the Hispanic paradox: demographic and epidemiologic approaches. Ann N Y Acad Sci. 2001;954:140174. Crossref, MedlineGoogle Scholar
5. Markides KS, Eschbach K. Aging, migration, and mortality: current status of research on the Hispanic paradox. J Gerontol B Psychol Sci Soc Sci. 2005;60B(Spec issue 2):6875. CrossrefGoogle Scholar
6. Acevedo-García D, Bates L. Latino health paradoxes: empirical evidence, explanations, future research, and implications. In: , Rodriguez H, Saenz R, Menjivar C, eds. Latino/as in the United States: Changing the Face of America. New York, NY: Springer; 2008:101113. CrossrefGoogle Scholar
7. Liao Y, Cooper RS, Cao G, et al.. Mortality patterns among adult Hispanics: findings from the NHIS, 1986 to 1990. Am J Public Health. 1998;88(2):227232. LinkGoogle Scholar
8. Hummer RA, Benjamins MR, Rogers RG. Racial and ethnic disparities in health and mortality among the U.S. elderly population. In: , Anderson NB, Bulatao RA, Cohen B, eds. Critical Perspectives on Racial and Ethnic Differences in Health in Late Life. Washington, DC: National Academies Press; 2004:5394. Google Scholar
9. Cho Y, Frisbie WP, Hummer RA, Rogers RG. Nativity, duration of residence, and the health of Hispanic adults in the United States. Int Migr Rev. 2004;38(1):184211. CrossrefGoogle Scholar
10. Borrell LN, Dallo FJ. Self-rated health and race among Hispanic and non-Hispanic adults. J Immigr Minor Health. 2008;10(3):229238. Crossref, MedlineGoogle Scholar
11. Shetterly SM, Baxter J, Mason LD, Hamman RF. Self-rated health among Hispanic vs non-Hispanic White adults: the San Luis Valley Health and Aging Study. Am J Public Health. 1996;86(12):17981801. LinkGoogle Scholar
12. Bzostek S, Goldman N, Pebley A. Why do Hispanics in the USA report poor health? Soc Sci Med. 2007;65(5):9901003. Crossref, MedlineGoogle Scholar
13. Kandula NR, Lauderdale DS, Baker DW. Differences in self-rated health among Asians, Latinos, and non-Hispanic Whites: the role of language and nativity. Ann Epidemiol. 2007;17(3):191198. Crossref, MedlineGoogle Scholar
14. Williams DR, Mohammed SA. Poverty, migration, and health. In: , Lin AC, Harris DR, eds. The Colors of Poverty: Why Racial and Ethnic Disparities Persist. New York, NY: Russell Sage; 2008:135169. Google Scholar
15. Dey AN, Lucas JWPhysical and Mental Health Characteristics of US- and Foreign-Born Adults: United States, 1998–2003. Hyattsville, MD: National Center for Health Statistics; 2006. Advance Data From Vital and Health Statistics, 369. Google Scholar
16. Anderson NB, Bulatao RA, Cohen B, eds. Critical Perspectives on Racial and Ethnic Differences in Health in Late Life. Washington, DC: National Academies Press; 2004. Google Scholar
17. Angel R, Guarnaccia PJ. Mind, body, and culture: somatization among Hispanics. Soc Sci Med. 1989;28(12):12291238. Crossref, MedlineGoogle Scholar
18. Finch BK, Hummer RA, Reindl M, Vega WA. Validity of self-rated health among Latino(a)s. Am J Epidemiol. 2002;155(8):755759. Crossref, MedlineGoogle Scholar
19. Angel JL, Buckley CJ, Finch BK. Nativity and self-assessed health among pre-retirement age Hispanics and non-Hispanic Whites. Int Migr Rev. 2001;35(3):784803. CrossrefGoogle Scholar
20. Phillips LJ, Hammock RL, Blanton JM. Predictors of self-rated health status among Texas residents. Prev Chronic Dis. 2005;2(4):19. Google Scholar
21. Franzini L, Fernandez-Esquer ME. Socioeconomic, cultural, and personal influences on health outcomes in low income Mexican-origin individuals in Texas. Soc Sci Med. 2004;59(8):16291646. Crossref, MedlineGoogle Scholar
22. DuBard CA, Gizlice Z. Language spoken and differences in health status, access to care, and receipt of preventive services among US Hispanics. Am J Public Health. 2008;98(11):20212028. LinkGoogle Scholar
23. Morenoff JD, House JS, Hansen BB, Williams DR, Kaplan GA, Hunte HE. Understanding social disparities in hypertension prevalence, awareness, treatment, and control: the role of neighborhood context. Soc Sci Med. 2007;65(9):18531866. Crossref, MedlineGoogle Scholar
24. Centers for Disease Control and Prevention. Health risks in the United States: Behavioral Risk Factor Surveillance System: at a glance 2009. Available at: http://www.cdc.gov/NCCDPHP/publications/aag/brfss.htm. Accessed December 15, 2009. Google Scholar
25. Nelson DE, Powell-Greiner E, Town M, Kovar MG. A comparison of national estimates from the National Health Interview Survey and the Behavioral Risk Factor Surveillance System. Am J Public Health. 2003;93(8):13351341. LinkGoogle Scholar
26. Salomon JA, Nordhagen S, Oza S, Murray CJL. Are Americans feeling less healthy? The puzzle of trends in self-rated health. Am J Epidemiol. 2009;170(3):343351. Crossref, MedlineGoogle Scholar
27. The 2003 Statistical Abstract of the United States. Available at: http://www.census.gov/prod/2004pubs/03statab/pop.pdf. Accessed March 7, 2010. Google Scholar
28. The Hispanic Population in the United States: March 2002. Washington, DC: US Census Bureau; 2003:Table 1.1. Current Population Survey, March 2002, PPL-165. Available at: http://www.census.gov/population/socdemo/hispanic/ppl-165/tab01-1.pdf. Accessed March 7, 2010. Google Scholar
29. Ramirez RR, de la Cruz GP. The Hispanic Population in the United States: March 2002. Washington, DC: US Census Bureau; 2002. Current Population Reports, P20–545. Google Scholar
30. Ren XS, Amick BC. Race and self assessed health status: the role of socioeconomic factors in the USA. J Epidemiol Community Health. 1996;50(3):269273. Crossref, MedlineGoogle Scholar
31. Ostrove JM, Adler NE, Kuppermann M, Washington AE. Objective and subjective assessments of socioeconomic status and their relationship to self-rated health in an ethnically diverse sample of pregnant women. Health Psychol. 2000;19(6):613618. Crossref, MedlineGoogle Scholar
32. Acevedo-García D, Bates LM, Osypuk TL, McArdle N. The effect of immigrant generation and duration on self-rated health among US adults 2003–2007. Soc Sci Med. 2010;71(6):11611172. Crossref, MedlineGoogle Scholar
33. Lee S, Grant D. The effect of question order on self-rated general health status in a multilingual survey context. Am J Epidemiol. 2009;169(12):15251530. Crossref, MedlineGoogle Scholar
34. Abdulrahim S, Baker W. Differences in self-rated health by immigrant status and language of preference among Arab Americans in the Detroit metropolitan area. Soc Sci Med. 2009;68(12):20972103. Crossref, MedlineGoogle Scholar
35. Willis GB, Royston P, Bercini D. The use of verbal report methods in the development and testing of survey questionnaires. Appl Cogn Psychol. 1991;5(3):251267. CrossrefGoogle Scholar

Related

No related items

TOOLS

SHARE

ARTICLE CITATION

Edna A. Viruell-Fuentes, PhD, MPH, Jeffrey D. Morenoff, PhD, David R. Williams, PhD, MPH, and James S. House, PhDEdna A. Viruell-Fuentes is with the Department of Latina/Latino Studies, University of Illinois at Urbana–Champaign. Jeffrey D. Morenoff and James S. House are with the Department of Sociology and the Survey Research Center, University of Michigan, Ann Arbor. David R. Williams is with the Department of Sociology and the Department of Society, Human Development, and Health, Harvard University, Boston, MA. “Language of Interview, Self-Rated Health, and the Other Latino Health Puzzle”, American Journal of Public Health 101, no. 7 (July 1, 2011): pp. 1306-1313.

https://doi.org/10.2105/AJPH.2009.175455

PMID: 21164101