SEX DIFFERENCES AND DEPRESSION IN PUERTO RICO

Psychology of Women Quarterly, 1987, 11, 443-459. Printed in the United States of America.

SEX DIFFERENCES AND DEPRESSION IN PUERTO RlCO

Glorisa J. Canino and Maritza Rubio-Stipec University of Puerto Rico Patrick Shrout

Columbia University Milagros Bravo and Robert Stolberg University of Puerto Rico Hector R. Bird N e w York Psychiatric Institute

Sex differences in rates of depressive disorders and depressive symptomatology, as measured by the Diagnostic Interview Schedule, are examined for an island-wide probability sample of Puerto Rico. Consistent with previous research, depression is significantly more prevalent in Puerto Rican women than men. Risk factors associated with depressive symptomatology are examined from a sex-role perspective. The results of multiple regression analyses show that even after demographic, health and marital and employment status variables are controlled, women continue to be at higher risk of depressive syrnptomatology than men. These results are interpreted within a cultural and sex-role perspective.

Sex differences in the rate of specific mental disorders among Hispanics have rarely been assessed using probability samples of the populations studied. Epidemiologic studies of clinical depression in generaj population surveys have been difficult to interpret because of variations in case definiThis research was supported by the Division of Biometry and Epidemiology of the National Institute of Mental Health, l R O l MH 36230-01. The authors would like to acknowledge Mr. Tomas Matos, our Data Analyst, Dr. Miguel Valencia, our sampling expert, Dr. Myrna Sesman, co-investigator, Dr. Lee Robins, and Dr. John Helzer from the Washington University in S t . Louis €or their invaluable collaboration with this project. Requests for reprints can be sent to Glorisa J. Canino, Ph.D., Department of Psychiatry, School of Medicine, G.P.O. Box 5067, San Juan, 00936, Puerto Rico. Published by Cambridge University Press. 0361-6843187 $5.00

+ .OO.

443

444

CANINO,RUBIO-STIPEC,SHROUT, BRAVO,STOLBERC, & BIRD

tions, diagnostic procedures and methodology (Boyd & Weissman, 1981). These difficulties are greater in studies among Hispanics because of the use of instruments that have not been standardized or validated for those populations. Studies with both Anglo and Hispanic clinical or community samples have consistently reported a higher rate of depressive symptoms and depressive disorders in women than in men. Weissman and Klerman (1977) reviewed over 40 community studies from 30 countries and found few exceptions to the observed female: male ratio on depression rates (about 2 : 1 to 3 : 1). Recently three of the Epidemiologic Catchment Area Study (ECA) sites have also found female predominance in both six-month and lifetime prevalence of major depression and dysthymia (Myers et al., 1984; Robins et al., 1984). Reports on sex differences in rates of depressive symptoms and diagnoses on Hispanic populations are similar (e.g., Burnam et al., in press; Vernon & Roberts, 1982). The rate of depressive symptoms in women is higher, regardless of their ethnicity or acculturation level (Frerichs, Aneshensel, & Clark, 1981; Quesada, Spears, & Ramos, 1978; Vega, Warheit, Auth, & Meinhardt, 1984; Warheit, Vega, Auth & Meinhardt, 1984). The prevalence of depression has not been studied in the general population of Puerto Rico. However, studies on mainland Puerto Ricans done in outpatient psychiatric clinics (Arce & Torres-Matrullo, 1978), university settings (Torres-Matrullo, 1976) or low-income urban areas (Hurst & Zambrana, 1980) have also shown that rates of depressive symptomatology or suicide attempts among mainland Puerto Rican women are higher than those of men. As Boyd and Weissman (1981) have illustrated, high levels of depressive symptoms are not synonymous with the diagnosis of depressive disorder. The explanation of the increased risk of depression in women, whether disorder or symptoms, is controversial. Some researchers question whether the reported findings are real or are reporting artifacts, i.e. that women are more likely to report symptoms (Phillips & Segal, 1969; Verbugge, 1976) or recall effects (Angst & Dobler-Mikota, 1983). The latter explanation has been challenged by Weissman, Leaf, Holzer, Myers, and Tischler (1984) who examined the cumulative rates of major depression by sex for different time periods. Past episodes were less likely to be remembered by either sex. However, in all the time periods assessed, the sex ratios were consistently higher for women, suggesting no differential recall among the sexes. Scholars offering other explanations for this female preponderance of depression consider this finding a real phenomenon and attribute the results to either female biological susceptibility or social causes (Weissman & Klerman, 1977). In this article we hypothesize social causation based on sex roles to explain the gender differences found in the prevalence of depression.

Sex Differences and Depression

445

Various researchers have stressed the importance of interactions between gender and social roles (such as marriage, parenting, employment) in the etiology of psychological disorders. According to this perspective, depression is expected to differ for men and women because of differing societal expectations (Aneshensel, Frerichs, & Clark, 1981; Gove & Geerken, 1977; Gove & Tudor, 1973; Radloff, 1975). Many of women’s traditional roles, particularly homemaking, are given low societal value and the roles may themselves be unrewarding. For women, work outside the home usually does not provide relief from homemaking responsibilities. In many instances, women’s assumption of nontraditional roles conflicts with societal expectations thereby generating greater stress in women than in men (Aneshensel et al., 1981). Thus from a sex role perspective, employed married men and women should differ since the interaction of gender with social role is what may lead to symptomatology rather than either marriage or employment alone. Research has supported this perspective, indicating significantly higher levels of depressive symptoms among employed married women than employed married men (Cleary & Mechanic, 1983; Newberry, Weissman, & Myers, 1979; Pearlin, 1975; Radloff, 1975). However, Gore and Mangione’s (1983) as well as Roberts and Roberts’ (1982) findings suggest that examining only employment is insufficient. They found that employed married women did not significantly differ in level of depression from comparable men and were significantly less depressed than their housewife counterparts. Gore and Mangione postulate that this discrepancy in findings may be due to the changing labor force composition or to geographical sample differences. Unfortunately much of this work fails to distinguish between type of occupation (professionaUnonprofessiona1; traditionaUnontraditional), and whether or not the women who went to work were forced to do so by economic circumstances. More specifically, it is hypothesized that males and females have a different vulnerability to depressive disorder because of the way they are socialized (Rosenfield, 1980). Several aspects of the traditional female socialization process have been associated with depression. While the overt manifestation of hostility or aggression is accepted in men, in women such expression is strongly discouraged (Bardwick, 1971; Block, 1973; Chesler, 1971). According to some sex-role theorists, women are more likely to introject anger or hostility when frustrated or in conflict, as well as when they experience a sense of loss, real or imagined, of a love object or aspiration. Psychodynamic theories hold that this sense of loss may result in intense feelings of rage and hostility, which are turned against the self and result in a depressive reaction (Bibring, 1953; Jacobson, 1971). Furthermore, imposed dependence lowers women’s self-esteem and increases their vulnerability to depression (Bart, 1971; Rosenfield, 1980). Radloff (1975) argues that depression is more prevalent among women because women have less controI over their environments than men and learn to feel help-

446

CANINO, RUBIO-STIPEC,SHROUT, BRAVO, STOLBERC,& BIRD

less. Depression results from learning that one’s actions may have unpredictable consequences. The traditional Puerto Rican family structure has been described as male dominant with a double standard relative to the expression of overt aggression and sexual behavior, and one in which women are expected to be homemakers and assume a passive, submissive role (Bird & Canino, 1982; Fernandez, Maldonado-Sierra, & Trent, 1958; Mufioz, 1979; Steward, 1956; Wolf, 1952). Although important changes in this traditional family structure have been recently observed (Rodriguez, Gonzalez, Mufioz, 1978) marked sex role differentiation in cultural expectations still prevails. One would expect that the effects of sex-role differentiation would apply to a greater degree to Puerto Rican and other Hispanic women since their socialization process strongly stresses that differentiation (Bird & Canino, 1982; Garcia Bahne, 1977). The present paper examines the association of gender with the prevalence of both depressive disorders and symptomatology in a probability sample of the adult Puerto Rican population. The data employed in our analyses are part of a larger research project on the epidemiology of mental disorders in Puerto Rico (Canino et al., 1987a). The epidemiologic survey made use of the Diagnostic Interview Schedule (Robins, Helzer, Croughan, & Ratcliff, 1981) after adapting it to the Puerto Rican population (Bravo, Canino, & Bird, 1987a; Canino et al., 1987b). The present report examines demographic, social, and health variables with the goal of establishing factors associated with observed gender differences in the prevalence of depressive conditions in Puerto Rico. The specific social role variables studied are employment status and marital status. In order to obtain a clearer picture as to how these roles may influence the prevalence of depression we control for known risk factors associated with the condition, such as age, education, area of residence, and health status. Most research on Anglos and Hispanics has reported an inverse relationship between the prevalence of depressive symptoms or current major depression and educational level (Comstock & Helsing, 1976; Craig & Van Natta, 1976; Dohrenwend & Dohrenwend, 1969; Vega, Kolody, & Warheit, 1985; Weissman & Myers, 1978). Differential rates of depressive symptoms and disorders have also been associated with different age patterns (Burnam et al., in press; Craig & Van Natta, 1979; Robins et al., 1984; Vega et al., 1984; Vega, Kolody, & Warheit, 1985; Weissman et al., 1984). Differences in prevalence rates by residence have been associated with depressive symptoms as well as major depressive disorders (Blazer et al., 1985; Myers et al., 1984; Robins et al., 1984; Weissman & Klerman, 1978). Depression has been found to be more prevalent in urban areas possibly because of the environmental stress associated with this type of residence. Poor health, as well as threatening disorders and conditions such as strokes, have been linked to depressive symptomatology (Robins, 1976;

447

Sex Differences and Depression

Robins & Guze, 1972). Link and Dohrenwend (1980) claimed that physical health problems are an important antecedent to a cluster of symptoms found in nonclinical depression, which they described as demoralization, elaborating on Frank’s (1973) construct. Recent evidence from an Anglo and Mexican-American sample showed a strong association between severe health complaints and depressive symptoms (Kolody, Vega, Meinhardt, & Bensussen, 1986). This association remained strong even after controlling for social and demographic variables. The association of depressive disorders with the presence of other mental disorders has also been corroborated in past research (Boyd et al., 1984). This study constitutes an important step in cross-cultural research of sex differences in depression since it is the first systematic application of the Diagnostic and Statistical Manual of Mental Disorders Criteria (DSM-111) through structured interview methods to the population of Puerto Rico. It is also the first time in which various demographic, social, and health variables are examined in relation to rates of depression for both men and women in Puerto Rico. METHOD

The NIMH Diagnostic Interview Schedule (DIS)

The DIS is a structured diagnostic interview designed to be administered by lay interviewers to respondents sampled from the general population (Robins et al. , 1981). Having been used in the National Institute of Mental Health (NIMH) Epidemiologic Catchment Area (ECA) program, the instrument has been discussed extensively elsewhere (Anthony et al., 1985; Eaton & Kessler, 1985; Helzer et al., 1985; Robins et al., 1982). The Spanish translation of the DIS developed for the ECA program (Karno, Burnam, Escobar, Hough, & Eaton, 1983) was adapted for use in Puerto Rico (Bravo, Canino, & Bird, 1987). The Spanish in this version differed slightly from the original translation, and the instrument was shortened to contain 14 DIS/DSM-I11 diagnoses. The agreement between results from this version when administered by psychiatrists and when administered by lay interviewers, as well as its comparison with psychiatric diagnoses, (Canino et al., 1987b) were comparable to similar studies conducted in the United States (Robins et al. , 1984). The present study reports data obtained on the prevalence of depressive disorders and symptomatology as measured by the DIS according to DSMI11 criteria. Lifetime prevalence refers to the proportion of persons who have met criteria for the disorder up to the date of assessment using retrospectively obtained data. Depressive disorders include specific DIS/DSMI11 disorders of major depressive episodes and dysthymia. Depressive

448

CANINO,RUBIO-STIPEC,SHROUT, BRAVO,STOLBERC,& BIRD

symptomatology refers to lifetime symptoms associated with major depressive episode andlor dysthymia. As reported by Canino et al. (1987a), the DIS was administered to persons selected from the residential population of Puerto Rico using a multistage probability sampling design. Noteworthy about this survey is its high completion rate: 91 % (n = 1551) of those selected were successfully interviewed. Besides being restricted to residents, the study population was limited to those aged 17-64. Statistical Models

This paper focuses on the interaction between gender and social roles to explain sex differences in depressive symptomatology. Multiple regression analysis is used to estimate the percentage of the variance that can be accounted for by sex, marital status, employment status, and a set of health variables, while controlling for age, area of residence, and education. The focus of our analysis is the importance of sex as a predictor of depressive symptoms even after accounting for other demographic characteristics, health conditions, and work and family roles. If, as proposed, depressive symptoms are more frequent among women irrespective of their health conditions or marital or employment status (role variables) , then the regression coefficient associated with sex should remain significant and positive after entering health and role variables into the equation. Variable definitions. The dependent variable in our model is the total number of depressive symptoms in either the major depression and/or dysthymia schedules of the DIS. The explanatory variables were dummy or binary coded as zero or one. Sex was coded as one for female, and zero for male. Age was divided into four dummy variables, one each for the following ranges: 17-24, 25-34, 45-54, 55-64. Each was coded as one for persons within one of the age intervals, and zero for the remaining ages. The 35-44 age served as the reference group in the regression equation. Area of residence was coded as one for those living in urban areas and zero for rural areas. Education level was divided into three classes: those who completed 11 years of education or less, those with high school education (12 years of schooling), and those with 13 years of school or more. Two dummy variables were constructed with the reference group being those with a high school diploma. Three levels of marital status were represented with two dummy variables, married and formerly married (separated, divorced, or widowed). The reference group was composed of those never married. Employment status was analyzed with two dummy coded variables: employed full-time and staying at home (coded as housewives, unemployed, and handicapped -physically or mentally). The reference group for employment status in the regression equation were those involved in an activi-

449

Sex Differences and Depression

ty that took them outside the home but who were not working full-time (working part-time, self-employed, doing volunteer work or students). An index of physical health was constructed as the presence or absence of a medically explained condition reported during the DIS interview. Mental status was defined as the presence or absence of a nonaffective disorder. Regression analyses. The explanatory variables of the model were entered in blocks in a hierarchical order. Background variables were entered first to partial out the effect of sample composition (in terms of sex, age, education, and geographical area). The set of variables referring to marital and employment status were entered in a second step. In the third step, variables reflecting the mental and physical health of the person were entered, under the assumption that an unhealthy condition increases the vulnerability of the individual to depressive symptoms (Robins, 1976; Robins & Guze, 1972). Finally, a regression analysis was used to examine interactions between sex and employment status. The main effects discussed above (age, sex, education, area of residence) and the interactions were entered simultaneously. RESULTS

Table 1 shows the distribution of the explanatory variables according to gender. There are no major differences between the demographic characteristics of men and women in this sample. Both are evenly distributed in urban and rural areas, have the same median age, and education levels. Although both sexes appear to have a similar mental health status as previously defined, women tend to report more DIS symptoms attributable to medical causes. In addition, men and women differ in the social roles they assume. More women are formerly married whereas more men have never married. Also, fewer women work full-time outside the home. Depressive disorders in Puerto Rico are more than twice as prevalent in women than men (10.7% vs. 4 . 9 % ) ,and the prevalence of dysthymia is almost four times greater for women than men. Table 2 shows the average number of lifetime depressive symptoms and the prevalence of affective disorders in men and women. The differences between sexes are not significant for major depression or depressive symptoms. Sex, age, area of residence, and education were all significantly related to depressive symptoms. Table 3 shows the results of the regression analysis with the main effects. Lifetime depressive symptoms are more frequent among women; the regression coefficient associated with sex, shown in the first column, is significant and positive. In addition, young people tend to have fewer symptoms than 35- to 49-year-old women; the coefficient of the group between 17-24 years is negative. At the other age intervals the

CANINO,RUBIO-STIPEC, SHROUT, BRAVO,STOLBERC,& BIRD

450

Table 1 Sample composition classified by sex Female Explanatory Variables of the Model

n

%

Mule a

n

%

Demogruphics

Age (yrs.) 17-24 25-34 35-44 45-54 55-64

Education (yrs.) 0-11 12 13 or more* Area Urban Rural

161 257 200 110 148

25.2 26.5 20.4 15.8 12.9

172 173 138 100 92

26.6 25.9 20.2 15.7 12.6

390 223 263

44.8 25.4 33.2

333 183 158

49.0 27.7 23.7

588 288

66.2 33.8

429 246

61.0 38.9

689 187

76.3 23.7

443 232

65.0 35.0

415 461

43.8 56.2

290 385

42.7 57.3

196 484 196

27.5 57.5 15.3

233 373 69

36.3 56.2 07.9

224 524

24.3 56.4

285 236

41.4 36.0

125

19.4

151

23.0

Health

Medically explained symptoms Present * Absent Nonaffective DIS Diagnosis Present Absent Social Roles Martial status Never married** Married Formerly married** Employment status Working full-time** At home (housewives, disabled, unemployed) Not at home but not working full-time (working part-time, self-employed,volunteer work, or student)

Note. F-tests were used to determine if differences between the sexes are statistically significant (n = 1551). a % =weighted percentage. * p < .05; * * p < ,025.

coefficient changes in sign and significance, suggesting a nonlinear association between age and depressive symptoms. Individuals living in urban areas tended to have more depressive symptoms than those in rural areas; a positive and significant coefficient was associated with area of residence. Individuals with low education tended

45 1

Sex Differences and Depression

to have more depressive symptoms; the coefficient associated with the lowest level of education was positive and significant. In the second hierarchical regression shown in the second column in Table 4, being formerly married, as compared with never married, was found to increase the vulnerability to depressive symptoms; the coefficient associated with formerly married is significant, positive and significantly different from that associated with being currently married. Fewer depressive symptoms are found among those classified as working full-time; the coefficient associated with holding a full-time job is negative. The coefficient of sex remains positive and significant after controlling for both marital and employment status. The inclusion of health-related variables in the third equation does not change the significance of the sex coefficient. Although women report more health problems and nonaffective disorders, these are related to depressive symptomatology. The effect of health and nonaffective disorders on depression does not account for the sex effect. Table 4 shows the results of the regression analysis with sex and employment status, which interact in complex ways. Women who stay at home are less likely to be depressed than men staying at home. The two-way interaction between working full-time and sex was not significant, but it combined married and unmarried women. However, the mean number of depressive symptoms in married women working full-time is significantly higher than in married men working full-time ( M = 1.4, SD = .6 vs. M = 0.8, SD = .5). There is no significant difference in the mean number of

Table 2 Lifetime prevalence rates of DIS/DSM-I11affective disorders and mean number of depressive symptoms by sex in Puerto Rico, 1984 ( % with lifetime history) Femaleb Disorders Affective Disorders Manic Episode Major Depressive Episode

Dysthymia Depressive Symptoms (mean)

Male

%

SEa

%

SE

10.7" 0.4 5.8 7.5* 2.4

1.1 0.2 0.9 0.9 2.7

4.9 0.6 3.9 1.9 1.6

0.9 0.4 0.8 0.6 2.3

Note. DIS indicates Diagnostic Interview Schedule; F-tests were used to

establish if differences are statistically significant in the breakdown. The sample ranged in age from 17-64 years. aSE =standard errors; bn = 876; cn = 675. * p < .01.

CANINO,RUBIO-STIPEC,SHROUT, BRAVO,STOLBERC, & BIRD

452

Table 3 Regression of demographic, social role and health variables on number of depressive symptoms Step 1 Explanatory Variables Sex (female) Age (yrs.1 17-24 25-34 45-54 55-64

Area (urban) Education (yrs.) 0-11 13 or more

ba 0.78

Step 2

b

t

6.06***

Step 3 t

Demographics 0.59 4.23***

b 0.52

-0.55 -2.59* - 0.21 -0.44 -2.19* -0.36 - 1.88 - 0.25 -0.32 - 1.69 0.08 0.29 0.02 0.07 - 0.06 0.08 -0.31 -0.28 - 1.09 - 0.33 0.63 4.22*** 0.61 4.25*** 0.56 0.31 2.20* -0.17 -0.93 I

Marital status Formerly married Married Employment status Working full-time At home

1.11 0.17 -0.06 -0.32

0.09 - 0.05

t 3.84**

- 1.00 - 1.37 - 0.27 - 1.39

4.38* * * 0.63

- 0.27

Social Roles 0.69 2.59* -0.23 -0.14 -0.35 - 1.87 1.11 0.21

0.49

1.95*

- 0.03

- 0.17

- 0.40

- 2.22* - 0.21

- 0.04

Health Presence of medically explained symptoms Presence of non-

0.94

affective disorder R2

1.40 0.05

0.07

7.08** * 10.08*** .18

Note. Steps 1, 2, and 3 indicate the sequence in which the sets of variables were entered in the hierarchical regression. Reference groups for explanatory variables are: male for sex, 35-44 years of age, rural for area, 12 years of schooling for education, never married for marital status and not at home or working full-time for employment status. ab = beta coefficient; t = t-statistic. * p < .05; " p < .01; * * * p < ,001.

depressive symptoms between married employed women and housewives ( M = 1.4, SD = .6 vs. M = 1.6, SD = .4).Apparently, the interaction of marital status with employment status varies by gender. However, we were not able to test this directly due to insufficient degrees of freedom when introducing a third variable to the interaction. The small R2 (. 18) suggests that there are other explanatory variables which have not been included in the model. For example, Brown and Harris (1978) have found that women who had experienced a severe event

453

Sex Differences and Depression

or major difficulty in their lives were at high risk for depression if they had lost their mothers prior to age 11, or had three or more children under age 14 and no intimate relationship. For example, Weissman and Klerman (1978) reported that marital discord is associated with depression in women. We did not control for whether or not women were employed by choice, nor did we control for type of occupation or number of children. While we did not have measures of these and other social variables that may be relevant to depressive symptoms, we did check other interactions of variables in Table 4.We found no significant interaction of sex and age, sex and marital status, and sex and area of residence.

Table 4 Regression of explanatory variables, and interactions on number of depressive symptoms Explanatory Variables

ba

t

0.95

3.35

- 0.21

- 0.99

- 0.04 - 0.36

-0.17 - 1.51 4.41

Demographic Sex (female) Age (yrs.) 17-24 25-34 45-54 55-64 Area (urban) Education (yrs.) 0-11 13 or more

- 0.25 0.57 0.10

- 1.36

0.74

- 0.06

- 0.35

0.57 0.04

2.22 0.21

- 0.27

- 1.30

0.34

1.51

0.94 1.38

7.05 9.99

Social Roles Marital status Formerly married Married Employment status Working full-time At home

Health Presence of medically explained symptoms Presence of nonaffective disorder Interactions Working full-time by sex Staying at home by sex RZ

-0.29 -0.74

-0.88 -2.13* .18

Note. Reference groups are male for sex, 35-44 years for age, rural for area, 12 years of schooling for education, never married for marital status and not at home or working full-time for employment status. a b = beta coefficient; t = t-statistic. * p < .05.

454

CANINO,RUBIO-STIPEC,SHROUT, BRAVO,STOLBERC,& BIRD DISCUSSION

As previously noted, in Puerto Rico the socialization process strongly stresses sex-role differentiation (Canino, 1982). Men are expected to work outside the home, whereas full-time employment for married women does not relieve them of their main responsibilities as homemakers. We have hypothesized that this patriarchal social context will result in sex differences in depressive symptoms, and these differences will be greater than those found in previous research with Anglo populations. Consistent with our hypotheses, the lifetime diagnosis of DIS/DSM-I11 depressive disorders was found to be significantly more prevalent among women than men in Puerto Rico. While the diagnoses of major depressive episode appeared to be higher in women, the sex difference for this specific diagnosis was not statistically significant, consistent with results in three U.S. communities (Robins et al., 1984). Dysthymia, on the other hand, was found to be significantly more prevalent among women than men in Puerto Rico. Furthermore, in Puerto Rico dysthymia was found to be 4.75 times more frequent in females than males, whereas in St. Louis the fema1e:male ratio was 2.57, in Baltimore 2.42, in New Haven 1.42, and in the Hispanic subsample in Los Angeles 1.85 (Burnam et al., in press). The higher sex ratio in Puerto Rico is consistent with our th2oretical model which postulates that Puerto Rican males and females have a different vulnerability to depressive disorder because of the more patriarchal social context in which they are socialized. Consistent with previous research using Anglo or Hispanic samples, even when demographic, social, and health variables are controlled, sex continues to predict depression in our study. This suggests that a more extensive examination of the relationship between gender and sex roles and depressive symptomatology is a fruitful area of research. Separated, widowed or divorced individuals have consistently been found to be at higher risk for both affective disorders and depressive symptomatology than married individuals (Blumenthal, 1975; Comstock & Helsing, 1976; Craig & Vare Natta, 1979; Cove, 1972; Quesada et al., 1978; Radloff, 1975; Vega et al., 1985; Vega, Kolody, & Valle, 1986; Vega, Kolody, Valle, & Hough, 1986; Vega, Warheit, & Meinhardt, 1984; Weissman & Myers, 1978). Poor physical or mental health has also been associated with increased risk for depression (Boyd et al., 1984; Kolody et al., 1986; Robins, 1976). In addition, most of these investigations have found that working full-time is associated with lower rates of depression. Our findings are consistent with these results. Factors that predict depressive symptomatology in Anglo samples are similar for both Puerto Rican men and women. The two-way interaction of full-time employment status by sex was not found to be significant for depressive symptomatology. At first glance this result seems to run counter to what is expected in our sex-role model, since

Sex Differences and Depression

455

we hypothesized that differences should be found even when both sexes perform the same role. However, a comparison between full-time employed married men and full-time employed married women revealed significant differences in depressive symptomatology between the sexes. Employment status does represent a different social role for married men and women. These results are in agreement with most previous research with Anglo populations that found significant differences in depressive symptoms between married employed men and women (Cleary & Mechanic, 1983; Newberry, Weissman & Myers, 1979; Pearlin, 1975; Radloff, 1975). However, the data are not consistent with those reported by Roberts and Roberts (1982) and Gove and Mangione (1983). Our data did not reveal a significant difference in depressive symptomatology between employed married women and housewives. While some investigations have obtained similar results (Newberry et al., 1979; Pearlin, 1975; Radloff, 1975; Roberts & O’Keefe, 1982), others have found significant differences in depression between these two groups (Cleary & Mechanic, 1983; Gove & Mangione, 1983; Rosenfield, 1980). Our findings suggest the need for more research on Puerto Rico women’s roles as workers in the context of differing marital roles. In estimating lifetime prevalence of disorders using cross-sectional designs it is not possible to disentangle the potential bias introduced by the effects of birth cohort, differential mortality due to the condition studied, and faulty recall. Nor does the cross-sectional design of this study address the possibility of reverse causation. Only with a prospective design would it be possible to begin to specify the direction and magnitude of the effects of risk factors associated with affective disorders. Our results on sex differences associated with depression are consistent with most previous research. Puerto Rican women report significantly higher rates of depressive symptoms than men, even when controlling for demographic, social role, and health variables. These findings suggest that social investment in enhancing the roles and status of women could reduce the symptomatology of depression. Fundamental changes in Puerto Rican sex-role stereotypes, however, are necessary if women’s vulnerability to depression is to be reduced. This process will require premeditated social measures in affirmative action, social legislation, and education through the mass media.

REFERENCES

Aneshensel, C. S., Frerichs, R. R . , & Clark, V. A. (1981).Family roles and sex differences in depression. Journal of Health and Social Behavior, 22, 379-93. Angst, G . , & Dobler-Mikota, A. (1983, December). Do the diagnostic criteria determine the sex ratio in depression? Presented at the 22nd annual meeting of the American College of Neuropsycho-pharmacology, San Juan, Puerto Rico. Anthony, J., Folstein, et al. (1985). Comparison of the Lay Diagnostic Interview Schedule

45 6

CANINO,RUBIDSTIPEC, SHROUT, BRAVO,STOLBERG,& BIRD

and a standardized psychiatric diagnosis experience in Eastern Baltimore. Archioes of General Psychiatry, 42, 667-675. Arce, A., & Torres-Matrullo, C. (1978, May). Acculturation: Its impact on treatment strategies. Presented at the American Psychiatric convention, Atlanta. Bardwick, J. (1971). Psychology of women. New York: Harper & Row. Bart, P. C. (1971). Mother Portnoy’s complaint. Transuction, 8, 69-74. Bibring, E. (1953). The mechanism of depression. In P. Greenacre (Ed.), Affective disorders (pp. 13-48). New York: International Universities Press. Bird, H., & Canino, G. (1982). The Puerto Rican family: Cultural factors and Family intervention strategies. Journal of the American Academy of Psychoanalysis, 10,257-268. Blazer, D., George, L. K., Landerman, R., Pennybacker, M., Melville, M. L., Woodbury, M., Manton, K.G., Jordan, K . , & Lock, B. (1985). Psychiatric disorders: A rurallurban comparison. Archioes of General Psychiatry, 42, 651-656. Block, J. H. (1973). Conceptions of sex role: Some cross-cultural and longitudinal perspectives. American Psychologist, 28,512-526. Blumenthal, M. D. (1975). Measuring depressive symptomatology in a general population. Archioes of General Psychiatry, 32,971-978. Boyd, J. H., Burke, J. D., Gruenberg, E., Holzer, C. E., Rae, D. S., Karno, C. M., Stoltzman, R., McEvoy, L., & Nestadt, C. (1984). Exclusion critieria of DSM-111. Archioes of General Psychiatry, 41,983-989. Boyd, J. H., & Weissman, M. W. (1981). Epidemiology of affective disorders. Archives of General Psychiatry, 38, 1039-1046. Bravo, M., Canino, G., & Bird, H. (1987). El DIS en Espaiiol: Su traduccion y adaptaci6n a Puerto Rim. Acta Psiquiatrica y Psicoldgica de America Latina, 33. Brown, G. W., & Harris, T. (1978). Social origins of depression. New York: Free Press. Burnam, M. A., Hough, H. L . , Escobar, J. I., Karno, M., Timbers, D. M., Telles, C. A,, & Locke, B. Z. (in press). Six month prevalence of psychiatric disorders: Mexican Americans and Non-Hispanic whites in Los Angela. Archiues of General Psychiatry. Canino, G. (1982). The Hispanic woman: Sociocultural influences on diagnoses and treatment. In R. Becerra, M. Karno, & J. Escobar (Ed.), The Hispanic: Mental health issues andstrategies (pp. 117-138), New York: Grune & Stratton. Canino, G., Bird, H . , Rubio, M., Bravo, M., Martinez, R., & Sesman, M. (1987a). The prevalence of specific psychiatric disorders in Puerto Rico. Archioes of General Psychiat y , 44,727-735. Canino, G . , Bird, H., Shrout, I?, Rubio, M . , Bravo, M . , Sesman, M., Guevara, M., Costas, H., & Cuzman, A. (1987b). The Spanish DIS: Reliability and concordance with clinical diagnosis in Puerto Rico. Archioes of General Psychiatry, 44,720-726. Chesler, P. (1971). Patient and patriarch: Women in the psychotherapeutic relationship. In V. Gornick & B. K. Moran (Eds.), Women in sexist society (pp. 43-66). Stanford: Stanford University Press. Cleary, P. D., & Mechanic, D. (1983). Sex differences in psychological distress among married people. Journal of Health and Social Behaoior, 24, 111-121. Comstock, C. W., & Helsing, K . J. (1976). Symptoms of depression in two communities. Psychological Medicine, 6, 551-564. Craig, T., Pearl, A., Van Natta, M. A. (1979). Influence of demographic characteristics on two measures of depressive symptoms. Archives of General Psychiatry, 36, 149-154. Craig, T. J. & Van Natta, P. A. (1976). Presence and persistence of depressive symptoms in patient and community populations. American Journal of Psychiatry, 133, 1426-1429. Craig, T. J. & Van Natta, P. A. (1979). Influence of demographic characteristics on two measures of depressive symptoms. Archiues of General Psychiatry, 35, 149-154, Dohrenwend, B. P., Dohrenwend, B. S. (1969). Social status and psychological disorders (pp. 207). New York: Wiley. Eastwood, M., & Trevelyan, M. H. (1972). Relationships between physical and psychiatric disorders. Psychological Medicine, 2 , 363-372.

Sex Differences and Depression

457

Eaton, W. W., Kessler, L. G. (Eds.). (1985). Epidemiological field methods in psychiatry: The NlMH Epidemiological Catchment Area Program. Orlando, FL: Academic. Fernandez, M. R., Maldonado-Sierra, E., & Trent, R. (1958). Three basic themes in Mexican and Puerto Rican family values. Journal of Social Psychology, 48, 167-181. Fox, J. W. (1980).Gove’s specific sex-roletheory of mental illness: A research note. Journal of Health and Social Behavior, 21,260-267. Frank, J. (1973). Persuasion and healing. New York: Schocker Books. Frerichs, R. R., Aneshensel C. S., & Clark, V. A. (1981). Prevalence of depression in Los Angeles County. Journal of Epidemiology, 113,691-699. Garcia Bahne, B. (1977). La Chicana and the Chicano family. In R. Sanchez & R. MartinezCruz (Eds.), Essays on la mujer (pp. 150-168). Los Angeles: Chicano Studies Center Publications, University of California. Gove, S., & Mangione, T. W. (1983). Social roles, sex roles and psychological distress: Additive and interactive models of sex differences. Journal of Health and Social Behavior, 24, 300-312. Gove, W., & Geerken, M. (1977). The effect of children and employment on the mental health of married men and women. Social Forces, 55,66-76. Gove, W., & Tudor, J. (1973). Adult sex roles and mental illness. American Iournal of Sociology, 78, 812-835. Gove, W. R. (1972). The relationship between sex roles, marital status and mental health. Social Forces, 51,34-44. Helzer, J. E., Robins, L. N., McEvoy, L. T., Spitznagel, E. L., Stoltzman, R. K., Farmer, A , , & Brockington, I. F. (1985). A comparison of clinical and diagnostic interview schedule diagnoses: Physician reexamination of lay-interviewed cases in the general population. Archives of General Psychiatry, 42,657-666. Hidroglou, M. A., Fuller, W. A., & Heikman, R. (1980). Supercarp (6th ed.).Ames, IA: Survey Section, Iowa State University. Hurst, M., & Zambrana, R. E. (1980). The health careers of urban women: A study in East Harlem. Journal of Women in Culture and Society, 5 , 112-126. Jacobson, E. (1971). Depression. New York: International Universities Press. Karno, M., Burnam, A., Escobar, J., Hough, R., & Eaton, W. (1983). Development of the Spanish-language version of the National Institute of Mental Health Diagnostic Interview Schedule. Archives of General Psychiatry, 40, 1183-1188. Kessler, R., & McRae, J. (1981). Trends in the relationship between sex and psychological distress. American Sociological Review, 46, 443-452. Kolody, B., Vega, W., Meinhardt, K., & Bensussen, G. (1986). The correspondence of health complaints and depressive symptoms among Anglos and Mexican Americans. Journal of Nervous and Mental Disease, 174,221-228, Link, B., & Dohrenwend, B. (1980). Formulation of hypotheses about the true prevalence of demoralization in the United States. In B. P. Dohrenwend, M. Schwartz-Gould, R. Neugebauer, & R. Wunsch-Hitzig (Eds.), Mental illness in New York: Praeger. Mufioz, M. (1979). The effects of role expectationson the marital status of urban Puerto Rican women. In B. Acosta (Ed.), Females and males in Latin America (pp. 75-84). Pittsburgh: University of Pittsburgh Press. Myers, J. K., Weissman, M. M., Tischler, G. L., Holzer, C. E., Leaf, P. J., Orvaschel, H . , Anthony, J. C., Boyd, J., Burke, J., Kramer, M., & Stoltzman, R. (1984). Six month prevalence of psychiatric disorders in three communities. Archives of General Psychiaty, 10,959-970. Newberry, P. M., Weissman, M. M., & Myers, J. K. (1979). Working wives and housewives. Do they differ in mental status and social adjustment? American Journal of Orthopsychiatry, 49, 282-291. Pearlin, L. R. (1975). Sex roles and depression. In N. Datan and L. H. Ginsberg (Eds.), Life-span developmental psychology: Normatioe life M.isis (191-207). New York: Academic.

458

CANINO,RUBIO-STIPEC,SHROUT, BRAVO,STOLBERC,& BIRD

Phillips, D., & Segal, B. (1969). Sexual status and psychiatric symptoms. American Sociological Review, 34, 58-72. Quesada, G . M., Spears, W., & Ramos, P. (1978). Inter-racial depressive epidemiology in the Southwest. Journal of Health and Social Behavior, 19, 77-85. Radloff, L. (1975). Sex differences in depression. The effects of occupation and marital status. Sex Roles, 1 , 249-265. Roberts, R., & O’Keefe, S. J. (1981). Sex differences in depression reexamined. Journal of Health and Social Behavior, 22, 394-400. Roberts, R., & Roberts, R. R. (1982). Marriage, work and depressive symptoms among Mexican Americans. Hispanic Journal of Behavioral Sciences 4, 199-221. Robins, A. (1976). Are stroke patients more depressed than other disabled subjects? Journal of Chronic Disease, 29,479-482. Robins, L. N., & Guze, S. (1972). Classification of affective disorders: The primary, secondary, the endogenous-reactive, and the neurotic-psychotic dichotomies. In T. W. Williams, M. Katz, & J. Shields (Eds.), Recent advances in the psychobiology of depressive illness (pp. 283-293). Washington DC: U.S. Government Printing Office. Robins, L. N . , Helzer, J. E., Croughan, J., & Ratcliff, K. S. (1981). The NIMH diagnostic interview schedule: Its history, characteristics, and validity. Archives of General Psychiatry, 38, 38-389. Robins, L. N., Helzer, J. E., Ratcliff, K. S., et al. (1981). Validityof theDiagnostic Interview Schedule, Version 11: DSM-I11 Diagnoses. Psychological Medicine, 12, 657-666. Robins, L. N., Helzer, J. E., Weissman, M. M., Orvaschel, H., Gruenberg, E., Burke, J. D., & Reigier, D. A. (1984). Lifetime prevalence of specific psychiatric disorders in three sites. Archives of General Psychiatry, 10, 949-958. Rodriguez, L. V., Gonzalez, J. L., & Muiioz, R. A. (1978). Study of child rearing practices, parents information and child development outcomes in Puerto Rico. Hato Rey, Puerto Rico: Health and Social Studies. Rosenfield, S. (1980). Sex differences in depression: Do women always have higher rates? Journal of Health and Social Behavior, 21,33-42. Steward, J. (1956). The people of Puerto Rico: A study of social anthropology. Evanston: University of Illinois Press. Torres-Matrullo, C. (1976). Acculturation and psychopathology among Puerto Rican women in mainland United States. American Journal of Orthopsychiatry, 46, 710-719. Vega, W. A . , Kodoly, B., & Valle, J. R. (1986). The relationship of marital status, confidant support, and depression among Mexican immigrant women. Journal of Marriage and the Family, 48,597-605. Vega, W. A., Kolody, B., Valle, R . , & Hough, R. (1986). Depressive symptoms and their correlates among immigrant Mexican women in the United States. Social Science Medicine, 22, 645-652. Vega, W. A., Kolody, B., Warheit, C . (1985). Psychoneuroses among Mexican Americans and other whites: Prevalence and caseness. American Journal of Public Health, 75, 523-527. Vega, W. A , , Warheit, G. J . , Buhl-Auth, J., & Meinhardt, K. (1984). The prevalence of depressive symptoms among Mexican Americans and Anglos. American Journal of Epidemiology, 120, 592-607. Vega, W. A,, Warheit, G. J., Meinhardt, K. (1984). Marital disruption and the prevalence of depressive symptomatology among Anglos and Mexican Americans. Journal of Marriage rL. Family, 46,817-824. Verbugge, L. (1976). Females and illness: Recent trends in sex differences in the United States. Journal of Health and Social Behavior, 17, 387-403. Vernon, S. W., & Roberts, R. E. (1982). Use of the SADS-RDC in a tri-ethnic community survey. Archives of General Psychiatry, 39,47-52. Warheit, G. J., Vega, W. A., Auth, J., & Meenhardt, K. (1984). Mexican American immigration and mental health: A comparative analysis of psychosocial stress and dysfunction. In W. A . Vega & M. R. Miranda (Eds.). Stress and Hispanic Mental Health: Relating

Sex Differences and Depression

459

Research to Service Delivery (76-109). Rockville, MD: U.S. Department of Health and Human Services, National Institute of Mental Health. Weissman, M. M., & Klerman, G. L. (1977). Sex differences and the epidemiology of depression. Archives of General Psychiatry, 34, 98-1 11. Weissman, M. M., & Klerman, G. L. (1978). Epidemiology of mental disorder: Emerging trends in the U.S. Archives of General Psychiatry, 35, 705-712. Weissman, M. M., Leaf, P. J., Holzer, C . F., Myers, J. K., & Tischler, G. L. (1984). The epidemiology of depression: An update on sex differences in rates. Journal of Affective Disorders, 7, 179-188. Weissman, M. M., & Myers, J. K. (1978). Affective disorders in a U.S. urban community. Archives of General Psychiatry, 35, 1304-1311. Wolf, K. (1952). Growing up and its price in three Puerto Rican sub-cultures. Psychiatry, 15, 40 1-433.

First draft receiued: December 22, 1986 Final draft received: June 8, 1987