Gender Differences in Depression: The Role Played ...

Sex Roles, Vol. 38, Nos. 7/8, 1998

Gender Differences in Depression: The Role Played by Paternal Attitudes of Male Superiority and Maternal Modeling of Gender-Related Limitations^ Brett Silverstein- and Arthur D. Lynch City College of New York

A reanalysis of a large multi-generational, predominantly Caucasian sample of adults found a large gender difference in self-reported depression involving anxiety, appetite and sleep disturbance, and fatigue (''anxious somatic depression ) but not in ''pure depression" unaccompanied by many of these other symptoms, replicating earlier findings on high school and college samples. Anxious somatic depression was prevalent among women whose fathers reported attitudes of male superiority and among women whose mothers reported emphasizing the importance of professional success but feeling that the jobs they held were not respected by others. Anxious somatic depression, but not pure depression, was found to be prevalent among women who did not attend college and among those who attended college but felt that the jobs they held were not respected by others. The gender difference in depressive symptomatology has received a great deal of attention (Culbertson, 1997; McGrath, Keita, Strickland, & Russo, 1990.) Recent evidence suggests that the higher prevalence of clinical and subclinical depression among females results because one subtype of depression rooted in limitations placed upon women is much more prevalent among females. Another subtype of depression, however, appears to be 'The research reported here was based on the data set entitled A Longitudinal Study of Generations and Mental Health. These data were collected by Vern L. Bengtson and donated to the archive of the Henry A. Murray Research Center of Radcliffe College, Cambridge, Massachusetts (Producer and Distributor). ^To whom correspondence should be addressed at Department of Psychology, City College of New York, New York, NY 10031; e-mail: BRSCC(aCUNYVM.CUNY.EDU. 539 0360-0025/98/0400-0539$ 15.00/0 © 1998 Plenum Publishing Corporation

540

Silverstein and Lynch

uninfluenced by such forces and is equally prevalent among females and males. The first type, which has been termed "anxious somatic depression," involves depression accompanied by anxiety and somatic symptomatology such as sleep and appetite disturbance as well as fatigue. The second type, termed "pure depression," involves depression unaccompanied to any great extent by these other symptoms and may possibly be heavily rooted in genetic factors. Two bodies of evidence are consonant with the hypotheses that the gender difference in depression results from a difference in a subtype of depression involving anxiety and somatic symptoms and that this subtype is rooted in limitations placed upon women. The first body of evidence involves studies that measure the individual symptoms of anxious somatic depression but that do not combine them into a measure of a single syndrome. For example, many studies of depressive symptomatology, including several studies of large epidemiologic samples, have found large gender differences in the prevalence of anxiety, fatigue, sleep, and appetite disturbance but little or no gender differences in the prevalence of the other symptoms of depression in samples of depressed patients as well as in the general community (Weissman, Bruce, Leaf, Florio, & Holzer, 1991; Young, Fogg, Scheftner, Keller, & Fawcett, 1990; Young, Scheftner, Fawcett, & Klerman, 1990. For citation of other studies, see Silverstein, Caceres, Perdue, & Cimarolli, 1995.) Furthermore, many studies have found relationships between the individual symptoms of anxious somatic depression and women's reports of having been exposed to gender bias. Landrine, Klonoff, Gibbs, Manning and Lund (1995) found that the degree to which women reported having experienced gender discrimination significantly predicted scores on self-report inventories of depression, anxiety, and somatization. Martz, Handley and Eisler (1995, p. 506) found that appetite disturbance was particularly common among women who scored high on a scale measuring the extent of perceptions of "the negative aspects of the feminine role." Depression, anxiety, and somatization, including appetite disturbance, have also been found in women who have been subjected to sexual harassment (Koss et al., 1994), sexual assault (Winfield, George, Swartz, & Blazer, 1990), and incest (Pribor & Dinwiddie, 1992.) The second body of evidence in support of the hypotheses examines the symptoms of anxious somatic depression in combination with one another. Many studies have reported high intercorrelation or comorbidity between depression, anxiety, somatization, and appetite disturbance, particularly among females (e.g., Maser & Cloninger, 1990; Swartz, Blazer, Woodbury, George, & Landerman, 1986; Wilson & Eldredge, 1991). A series of studies used self-report inventories, such as the Center for Epidemiologic Studies Depression Scale, to divide respondents into those

Gender Differences in Depression

541

reporting low levels of depression, those reporting the combined symptoms of anxious somatic depression, and those reporting pure depression. These studies found females to exhibit a higher level of anxious somatic depression than males but not a higher level of pure depression in both a high school sample (Silverstein, Caceres et al., 1995) and a college sample (Silverstein, Clauson, Perdue, Carpman, & Cimarolli, 1998). In addition, measures of social-psychological variables related to limitations placed on female achievement were found in these studies to be highly related to female respondents' reports of anxious somatic depression, but unrelated to reports of pure depression. These variables include the beliefs reported by females in a sample of high school students and an independent sample of college students that their opportunities had been limited by social reactions to their gender (sample item: "More people would pay attention to my ideas if I were male") (Silverstein, Caceres et al., 1995); female high school and college students' behefs that their mothers' opportunities had been limited by reactions to their mothers' gender (sample item: "When you were growing up, how much did your mother feel limited because she was female?") (Silverstein, Caceres et al., 1995; Silverstein et al., 1998; Silverstein, Perlick, Clauson, & McKoy, 1993); and female high school and college students' beliefs that their fathers would have preferred them to be male (sample item: "When you were growing up, how much did you have the sense that your father would have been prouder of you if you were not a girl?") (Silverstein et al., 1993; Silverstein et al., 1998). In a recent study of female high school students and their mothers, mothers' self-reports of having been limited by reactions to their gender (sample item: "When your daughter was growing up, how much did you feel limited by being female?") were found to predict their daughters' reports of anxious somatic depression, but not daughters' reports of pure depression (Silverstein & Blumenthal, 1997). Mothers' self-reported depression was found to be unrelated to their daughters' reports of anxious somatic depression but related to daughters' reports of pure depression, suggesting the possibility that pure depression may be influenced by genetic factors. Additional evidence that anxious somatic depression may be influenced by social psychological forces comes from a series of studies that have found that several indices of the prevalence of anxious somatic depression among females share similar variation across birth cohorts. These indices include the gender difference in the prevalence of depression (Silverstein & Perlick, 1991), the gender difference in the prevalence of somatic symptomatology (Silverstein & Perlick, 1995), the incidence of anorexia among teenage females (Silverstein & Perlick, 1995 derived from Lucas, Beard, O'Fallon, & Kurland, 1991), and the extent of the relation-

542

Silversteio and Lynch

ship among females between self-reported depression and somatic symptomatology (Silverstein, Clauson, McKoy, Perdue, & Raban, 1995). All of these studies include measures demonstrating that the shared covariation is due not to the age of respondents but to the year in which they were born. In other words, these indices of anxious somatic depression were found to be similarly related to the period when respondents were born, which is further suggestive of a syndrome rooted in social-psychological factors that exhibit short-term historical change rather than genetic factors that do not. The mechanisms wherein limitations placed upon women may eventuate in the development of anxious somatic depression have been discussed in detail elsewhere (Silverstein & Perlick, 1995). Briefly, many bright, ambitious girls develop identities and senses of self-worth based upon their intelligence and their competence in school. But some of these girls find limitations placed on their achievements or on the respect they receive for their achievements because they are female, particularly after they reach adolescence. The message that living up to their aspirations and obtaining the respect for their achievements that they desire may be rendered more difficult by social responses to their gender can be communicated to these girls through several channels. They may receive less support, both financial and psychological, than boys for attending college. They may come up against gender biases in hiring and promotion practices that limit their career opportunities. Thus, we might expect anxious somatic depression to be particularly prevalent among women who were unable to attend college or to achieve respected careers. They may find that people whose opinions matter to them, such as their fathers, do not accord them as much respect for their achievements as they would if they were male. Thus, we might expect anxious somatic depression to be particularly prevalent among women whose fathers hold attitudes of male superiority. And by modelling themselves upon, and identifying with, mothers who themselves may have been unable to live up to their aspirations, some girls may come to associate being female with facing limitations on their possibilities of achievement. Thus, we might expect anxious somatic depression to be particularly prevalent among women whose mothers failed to live up to their aspirations for career success. The sense of failure many of these women may come to feel may result in depressed mood. The fear they may develop that they will never be allowed to live the lives they desired may result in anxiety. The combination of depression and anxiety may result in somatic symptoms. The study reported here reanalyzes data from a study (Bengtson & Dunham, 1986) in which the Center for Epidemiologic Studies Depression Scale (CES-D) was distributed to a large multi-generational sample of families. This database allows several further tests of the hypothesis that the


543

gender difference in depression results because females are more likely than males to report a type of depression accompanied by anxiety and somatic symptomatology rooted in factors related to limitations placed on female achievement, but not more likely to exhibit a type of depression unaccompanied by these other symptoms, possibly rooted more heavily in genetic factors. First, included in the study are many pairs of mothers and daughters, allowing replication of the finding that daughters' reports of anxious somatic depression are correlated with mothers' reports of limitations placed on their own achievement. Second, the inclusion in the study of many father-daughter pairs allows the first direct examination of the relationship between daughters' reports of anxious somatic depression and fathers' reports of attitudes regarding male superiority. This extends the previous finding that young women who believe that their fathers exhibited a preference for males are particularly likely to report anxious somatic depression. Third, the gender difference in anxious somatic depression but not pure depression and the relationship between anxious somatic depression and indices of limitations placed on females have previously been reported for several samples of high school and college students. This study reports similar analyses of a large multi-generational sample of adults. Fourth, although the study does not include controls that allow the effects of birth cohort to be disentangled from the effects of age, it does allow preliminary tests of the hypothesis that the prevalence of a social-psychologically rooted syndrome of anxious somatic depression should vary among different birth cohorts while the prevalence of a genetically rooted pure depression should not vary among people born during different time periods.

METHOD Sample

The study reported here was a reanalysis of data from a longitudinal study of generations and mental health done by Bengtson (Bengtson et al., 1986). In the original wave of the study, in 1971-72, data were collected from 2045 predominantly Caucasian subjects. The participants in the study were recruited through the first-generation male subjects, who were members of a large metropolitan health maintenance organization in southern California. Twelve percent of the men who were above age 55 and who had at least one dependent that were enrolled in the plan were sent a screening questionnaire. Five hundred and fifteen of these men were selected because they were members of multi-generation families and were willing to participate. These men and the members of their families (totalling 3,184 indi-

544


viduals) were mailed lengthy questionnaires. In 1985, the families were contacted for a second time. 2612 members of the original Time 1 pool were identified and sent questionnaires. Sixty two percent of the questionnaires were returned. The surveys mailed to these respondents in the second wave of the study, from which the data analyzed here were taken, included the CES-D, the depression inventory used in previously-pubhshed studies of anxious somatic depression. Fifty eight percent of the Time 2 respondents were females and 74% of the respondents were married. The 221 respondents belonging to the first generation (the original members of the HMO and their spouses) had a mean age at Time 2 of 78. The 556 respondents belonging to the second generation of the study (the daughters and sons of the first generation and their spouses) had a mean age of 57. The 554 third generation respondents had a mean age of 33.

Measures The CES-D is one of the self-report depression inventories most frequently used in epidemiologic research. It consists of 20 questions regarding symptomatology experienced during the previous week. Possible scores range from 0 through 60. In utilizing a self-report inventory to compare two types of depression, it is necessary to use a cutoff to define high levels of depression. Previous studies of non-adolescent groups have used the cutoff of 16 or greater recommended by the developers of the scale (Radloff & Locke, 1986) for measuring moderately-high levels of depression, but this cutoff results in high estimates of depression prevalence (Silverstein et al., 1997.). This is the first study of anxious somatic depression in which the sample size of adult respondents is large enough to use the cutoff of 28 or greater on the CES-D that is suggested by the developers of the scale for defining clinical depression because it results in prevalence estimates similar to the prevalence of major depression found in studies that have used research interviews. Other symptoms used to create an operational definition of anxious somatic depression that is as close as possible to that used in other studies were measured using individual items from the CES-D. This inventory asks respondents to indicate using a four-point scale ranging from "Rarely or none of the time" to "Most of the time" how often during the previous week they experienced each of a number of symptoms. Sleep disturbance was measured with the item "My sleep was restless." Appetite disturbance was measured with the item "I did not feel like eating; my appetite was poor." Fatigue was measured with the item "I could not get 'going'." Anxiety was measured with the item "I felt fearful."


545

Respondents who reported experiencing these problems "occasionally" or "most of the time" were categorized as scoring high on these items. The double use of items from a single questionnaire results because the currently accepted definition of depression does not distinguish between two subtypes of anxious somatic and pure depression and so includes somatic symptoms such as sleep and appetite disturbance and fatigue. As a result, the research instruments that are widely-used and well-validated include such symptoms as possible criteria in their definition of depression. At present, it is not possible to create operational definitions of any type of depression based on validated instruments that do not include these symptoms. But because all research measures also allow depression to be defined or diagnosed in the absence of these symptoms, it is possible to use such measures to define two subtypes of depression with and without much anxiety and somatic symptomatology. Categorizing Pure vs. Anxious Somatic Depression. Respondents were categorized as exhibiting anxious somatic depression if they scored above cutoff on the total CES-D and also reported at least two of the other four symptoms: sleep disturbance, appetite disturbance, fatigue, and anxiety. Respondents were categorized as exhibiting pure depression if they scored above cutoff on the CES-D but reported only zero or one of the other symptoms. (This operational definition of pure depression receives further attention in the discussion.) Respondents were categorized as exhibiting low depression if they scored below cutoff on the CES-D. Other Measures. Limitations possibly associated with gender were measured using items regarding respondents' education and their attitudes regarding jobs. Respondents answered the question "Do you feel the job you do is respected by others?" using a five-point scale ranging from "1 = never" to "5 = very often". Respondents who answered "fairly often" or "very often" were categorized as feeling that their jobs were respected. Respondents were asked to rank order the importance of nine "values in life" (e.g., patriotism, personal freedom, a world at peace). Those who ranked "achieving success in your job or profession" in the top five were categorized as placing emphasis on career success. Two items measuring attitudes about gender roles were included in the survey. Both were measured on four-point agree-disagree scales in which 1 = "strongly agree" and 2 = "somewhat agree". One item was "Some equality in marriage is a good thing, but by and large the husband ought to have the main say in family matters." The other item was "It goes against nature to place women in positions of authority over men." Respondents who agreed with either of these statements were categorized as exhibiting belief in male superiority.

546


For the cohort analyses, respondents were divided by their decade of birth (pre-1920, 1920-29, 1930-39, 1940-49, 1950-59. The 203 respondents born prior to 1920 were combined into one cohort in order to obtain an adequate number of respondents for the earliest cohort.) Because the measure of symptomatology was only administered to this sample at one point in time, it is not possible to disentangle birth cohort effects from age effects as has been done in previously published studies of cohort differences in the symptoms of anxious somatic depression.

RESULTS Perhaps the most intriguing findings reported here deal with the relationship between daughters' symptomatology and parental attitudes. In order to avoid confounds with the cohort effects reported below, examination of the relationships between the symptomatology reported by women and the attitudes of their parents were limited to daughters in one generation and their parents. Symptomatology reported by women in the third generation (the last generation of adults reported in the study), which was born after World War II, was related to attitudes reported by their parents. Mothers were divided into those who did not place emphasis on career success, those who placed emphasis on career success and felt that their jobs were respected, and those who placed emphasis on career success but did not feel that their jobs were respected. Although the number of mother-daughter pairs (141) included in the analysis was not small, given the relatively low prevalence of the two subtypes of depression and the relatively low percentage (13%, n = 19) of mothers who placed emphasis on careers but did not feel that their jobs were respected (it appears that some women without jobs did not respond to this item), a 3 (mothers' attitudes toward jobs) by 3 (daughters' depression category) analysis could not be validly performed due to small cell sizes. But by collapsing cells, it was possible to perform an analysis comparing the prevalence (21*^ = 4/19) of anxious somatic depression reported by daughters of mothers who placed high emphasis on jobs and did not feel that their jobs were respected with the prevalence {19c = 8/122) of anxious somatic depression reported by daughters of other mothers (Fishers exact test p = .05). (The daughters of the mothers in the two groups that were collapsed into the "other" group in this analysis exhibited 6.3% and 6.7% prevalence of anxious somatic depression.) It appears that mothers who place importance on, but do not attain, career success have daughters who are particularly likely to report anxious somatic depression. A similar analysis relating


547

daughters' anxious somatic depression to fathers' attitudes toward jobs did not even approach significance (Fishers exact test/? = .15), but this finding is difficult to interpret given that only 9% of fathers (n = 13) reported placing emphasis on career success but feeling that their own jobs were not respected. On the other hand, 60% (n = 88) of the 143 fathers for whom data regarding their daughters' symptomatology was available agreed with at least one of the two attitude items regarding male superiority. Whereas 11% of the daughters of these men reported anxious somatic depression, none of the daughters of the 57 fathers who did not agree with either of these two items reported anxious somatic depression (Fishers exact test two-tail p = .01). Of the mothers in this sample for whom daughter data was available, 42% agreed with at least one of the two items measuring beliefs in male dominance. In contrast with the analysis of the father-daughter data, however, the daughters of mothers who agreed that males should be dominant actually exhibited slightly (nonsignificantly) lower prevalence of anxious somatic depression (8%) than the daughters of mothers who did not agree with either of the two male-dominance items (11%). It appears that fathers, but not mothers, who believe in male superiority have daughters who exhibit anxious somatic depression. The importance of this subtype of depression that is associated with these parental attitudes is exhibited in the comparison of depression prevalence among females and males. In this sample, females were twice as likely as males to score 28 or greater on the CES-D (12% vs. 6%, x'(l,1232) = 13.64, p < .001). But, as shown in Table I, this (6%) gender difference in self-reported depression is due primarily to the (5%) difference in prevalence of anxious somatic depression (females—9%, males 4%), not to the small (1%) gender difference in prevalence of pure depression (overall X-(2,1231) = 13.71, p < .002). As in earlier studies of high school and college students, in this sample of adults, psychosocial variables were related to reports of anxious somatic Table I. Male and Female Respondents Reporting Low Depression, F^ire Depression, or Anxious Somatic Depression"

Males Females

Low Depression

Pure Depression

Anxious Somatic Depression

485 (94%) 634 (88%)

9 (2%) 24 (3%)

20 (4%) 61 (9%)

"Numbers in parentheses indicate the percentage of respondents in each row exhibiting each type of symptomatology.


548

depression. The proportion of respondents reporting anxious somatic depression was almost twice as high among those who had not attended college (9%) as among those who had attended college (5%) and almost three times as high (11%) among respondents who reported feeling that their jobs were unrespected as among those who reported feeling that their jobs were respected (4%). The effects of both of these limitations are most apparent in Table II, in which the female and male respondents were separately categorized into those who did not attend college, those who attended college but felt their jobs were unrespected, and those who attended college and felt their jobs were respected. Both a practical and a theoretical rationale existed for not dividing the respondents who did not attend college into two groups based on how respected they felt their jobs were. The practical rationale was the need to maximize cell sizes. The theoretical rationale was that for many years, not being allowed to attend college has been a major limitation

Table IL Female and Male Respondents Categorized Separately According to College Attendance and Feeling Unrespected in Their Jobs Who Report Low Depression, Pure Depression, or Anxious Somatic Depression" Women

No college Attended college but felt job was unrespected Attended college and felt job was respected

Low Depression

Pure Depression


271(86%)

12(4%)

32(10%)

315 (52% of women)

73 (83%)

3 (3%)

12 (14%)

88 (14% of women)

188 (91%)

9 (4%)

9 (4%)

206 (34% of women)

Total

Men

No college Attended college but felt job was unrespected Attended college and felt job was respected

Low Depression

Pure Depression


141 (90%)

5 (3%)

11 (7%)

157 (35% of men)

58 (91%)

2 (3%)

4 (6%)

64 (14% of men)

221 (97%)

2(1%)

4 (2%)

227 (50% of men)

Total

''Numbers in parentheses in first three columns indicate the percentage of respondents in each row exhibiting each type of symptomatology. Numbers in parentheses in fourth column indicate the percentage of female or male respondents in each college/job category.


549

placed upon women. Even some women who have eventually attained success have reported feeling distress over not having been able to pursue higher education (Silverstein & Perlick, 1995.) The prevalence of anxious somatic depression reported by women who attended college and felt their jobs were respected (4%) was much lower than the prevalence reported by women who had not attended college (10%) or by women who had attended college but felt their jobs were unrespected (14%). In fascinating contrast, very little difference in reported prevalence of pure depression was found between women who had attended college and felt their jobs were respected (4%), those who had not attended college (4%), and those who had attended college and felt their jobs were unrespected (3%) (overall x"(4,605) = 8.47, p < .08). The nonsignificance of the analysis results from the lack of difference between the three groups of women in the prevalence of pure depression. In order to study the relationship between the psychosocial measures and reports of anxious somatic depression, all women who did not meet criteria for anxious somatic depression were combined into one group. If the low depression and pure depression groups are collapsed into one group of women who do not exhibit anxious somatic depression, a 3 (college/job category) by 2 (respondents do or do not report anxious somatic depression) y} analysis results in a highly significant (p < .02) x^ of 8.41. This analysis indicates a strong relationship between women's reports of limitations in their educational or occupational achievement and their reports of anxious somatic depression. As depicted in the table, the lack of relationship between the college/job categories and reports of pure depression exhibited by females is not so clear among males, but cell sizes are too small to allow a 3 x 3 analysis. When the male respondents who did not exhibit anxious somatic depression are combined into one group, a significant (^-(2,448) = 7.03, p < .03) relationship between college/job category and prevalence of anxious somatic depression is found. It appears that the gender difference in anxious somatic depression, and thus the gender difference in depression, is due in part to the lower percentage of women (34%) than of men (51%) who belong to the group that attended college and felt that their jobs were respected that exhibits low prevalence of anxious somatic depression. In the cohort analysis presented in Table III, whereas the prevalence of anxious somatic depression was over twice as high in one cohort as it was in another, the prevalence of pure depression varied little by cohort. Although the cohort analyses were based on over 1200 respondents, statistical tests of inter-cohort variance must use the number of cohorts. Notwithstanding the small number of cohorts (5) used in the analysis, the inter-cohort variance in prevalence of anxious somatic depression (3.3) was

550


Table IIL Respondents in Each Birth Cohort Reporting Low Depression, Pure Depression, or Anxious Somatic Depression''

Years of Birth Prior to 1920 1920-1929 1930-1939 1940-1949 1950-1959

Low Depression 190 253 177 112 367

(94%) (93%) (87%) (92%) (90%)

Pure Depression 4 5 7 3 13

(2%) (2%) (3%) (3%) (3%)

Anxious Somatic Depression 9 15 19 7 26

(4%) (6%) (9%) (6%) (6%)

^Numbers in parentheses indicate the percentage of respondents in each row exhibiting each type of symptomatology.

significantly greater (over ten times as high) than the minuscule inter-cohort variance in prevalence of pure depression (0.3, t = 2.51, df = 4, onetailed p < .05). The cohort analyses had to be performed on the data of females and males combined because when the data of females and males were separately divided into five cohorts, the number of respondents reporting pure or anxious somatic depression in most cohorts was so small (in 9 of 20 cells, the n was less than four) as to render meaningless any comparison of prevalence between cohorts.

DISCUSSION This is the third sample in which the gender difference in the prevalence of self-reported depression accompanied by anxiety and somatic symptomatology was found to be large but the difference in depression unaccompanied by these other symptoms was found to be small or nonexistent. Much higher prevalence among females compared to males of anxious somatic depression but not pure depression has now been found among high school students, college students, and adults. It should be noted that all of these studies have relied on self-report measures of depression. This is the fourth sample in which issues related to achievement reported by females have been found to be correlated with self-reported prevalence of anxious somatic depression but not of pure depression. Previous studies designed to test these relationships found reports of anxious somatic depression but not pure depression to be related to respondents' reports of feeling that their opportunities had been limited by responses to their gender. Because the data reported here are based on reanalysis of a survey distributed by others, the explicit link made in previously-pub-


551

lished studies between achievement limitations and gender must be inferred here. We know that reported anxious somatic depression but not pure depression is high among women who did not attend college and among those who attended college but did not feel that their jobs were respected. The relationship between gender and both educational and occupational attainment must have been well known to most of the respondents in this study, two-thirds of whom were born prior to 1950. But using the data available from this survey, we cannot determine whether the high prevalence of anxious somatic depression reported by women who did not attend college or who felt their jobs were unrespected was explicitly related to connections they may have made between their gender and their opportunities for advancement, as found in previous studies. Furthermore, the items used here to measure perceived job respect, emphasis on career success, and attitudes of male superiority are of unknown validity and reliability. The same limitations apply to the mother-daughter data. A previous study of high school students found daughters' reports of anxious somatic depression to be related to mothers' beliefs that their own opportunities had been limited by their gender (Silverstein et al., 1997). In the analysis reported here, mothers who place importance on career success but do not feel that their jobs are respected have daughters who are likely to report anxious somatic depression, but we cannot determine the importance of the links made by either the mothers or the daughters between the mothers' gender and the lack of the mothers' career success. To our knowledge, this study is the first to directly link women's symptomatology to attitudes toward women reported by their fathers. Earlier studies reported a significant relationship between daughters' disordered eating and their beliefs that their fathers agreed that "a woman's place is in the home" (Silverstein, Perdue, Wolf, & Pizzolo, 1988) and a relationship between daughters anxious somatic depression and the daughters' perceptions that their fathers exhibited a preference for males (Silverstein et al., 1993; Silverstein et al., 1998), but none of these studies actually surveyed fathers. In the study reported here, fathers who agree that men should be dominant over women are more likely than fathers who do not agree to have daughters who exhibit anxious somatic depression. The comparison of the mother-daughter and father-daughter analyses reported here resembles findings from other studies. Mussen and Rutherford (1963) found that the sex-role preferences exhibited by girls were related to the amount of acceptance of femininity reported by their fathers and the level of self-acceptance reported by their mothers, but not to fathers' self-acceptance or mothers' attitudes toward femininity. Silverstein, Perdue, Wolf, & Pizzolo (1988) found the disordered eating reported by college women to be significantly related to the women's perceptions of

552


what their mothers thought about their own careers and to whether they felt that their fathers treated a boy as the most intelligent sibling in the family but not to their perceptions of what their fathers thought of their own careers or whether they felt that their mothers treated a boy as the most intelligent sibling in the family. In the analyses reported here, the amount of anxious somatic depression exhibited by women was found to be significantly related to reports made by their mothers of emphasizing career success but not feeling respected in their jobs but not comparable reports made by their fathers and to their fathers' attitudes of male superiority but not to comparable attitudes of mothers. (The small number of parents who reported both placing emphasis on career success and also feeling that their jobs were not respected renders confidence in the first conclusion reported above less certain than confidence in the second conclusion.) Many women appear to treat their mothers as role models and sources with whom to identify and to treat their fathers as sources whose opinions are influential. The analyses reported here, including the cohort analyses, add some support for the hypothesis that pure depression is not based in social-psychological factors. Much more work must be done before firm conclusions can be drawn regarding this conjecture. As mentioned above, because at the present time the distinction between pure and anxious somatic depression is not yet recognized in the fields of psychology and psychiatry, validated self-report and diagnostic measures of depression include the somatic symptoms of fatigue, appetite and sleep disturbance, making pure depression a residual category comprised of people who meet criteria for depression but not for anxious somatic depression. If research continues to support the distinction between pure and anxious somatic depression, it should eventually be possible to develop independent criteria for the two subtypes of depression, overlapping in only a few symptoms, such as dysphoria, but with mostly distinct symptoms. Much future research remains to be done to develop these two sets of criteria. Future research must also compare females and males using research interview measures of depression. Furthermore, the paternal attitudes of male superiority measured here among fathers of baby boomers may now be somewhat dated. Nowadays, fewer fathers may straightforwardly assert that men should be dominant over women. But many fathers may continue to exhibit to their daughters more subtle versions of these attitudes of male superiority, such as believing that it is more important for men than for women to succeed in careers. Research is needed to relate these more subtle contemporary paternal attitudes to the prevalence of anxious somatic depression reported by daughters born after the baby boom years.


553

Beginning at adolescence, women exhibit higher prevalence than men of several disorders, including, depression, anxiety, disordered eating, headaches, and fatigue. We hypothesize that most of these differences, which now tend to be treated as independent, may result from a gender difference in prevalence of a single disorder—anxious somatic depression. Furthermore, the higher prevalence of this disorder among women results from limitations placed on the ability of many women to achieve successful careers and to be accorded the respect for these achievements that might be given to men. In order to understand who is apt to suffer anxious somatic depression and why it is now so prevalent, we must focus on the psychology of how these limitations affect women. That is, the people most likely to feel inadequate are those who fail to live up to their standards for themselves. Women who do not aspire to career success, for example, are less hkely to feel limited by traditional gender roles than those who judge themselves, at least in part, on their ability to achieve in areas historically reserved for men. The past few decades have seen a great increase in the number of young women who grow up feeling that they have the ability to succeed in nondomestic roles, who develop identities centered in part around such success, and who do not view full time homemaking as being a complete success. The good news is that more women than ever before are now able to attain such success by attending college and achieving respected careers. The bad news is that the increase in the number of women who harbor these aspirations may have outpaced the increase in the number who have been allowed to live up to them. That is, many of the contemporary young women who aspire to attain respect for achievements in such areas as politics, professions, sports, and business also receive messages and treatment that undercut their sense of having succeeded in living up to these standards. These may take the form of glass ceilings, pressure to take "mommy track" jobs, and gender biases in hiring and promotion. They may also take the form of more subtle psychological biases such as media depiction of women as more helpless than men, teachers' beliefs that boys are naturally better than girls at math and science, community attitudes that working women, even those in good jobs, are not truly successful if they rely on others for childcare, and the attitudes of many fathers that the career achievements of their daughters are in the end less important than those of their sons. We suggest that feelings of inadequacy, and the symptoms of anxious somatic depression, are particularly likely to afflict women who exhibit the following three characteristics: 1. They aspire to achieve outside of the home; 2. They are exposed to mothers who model the notion that being female has hindered them from

554


feeling successful and respected; 3. They are exposed to several of the limitations placed on women attaining respect just described. Very many contemporary women appear to exhibit these three characteristics, which we believe goes a long way to explaining why women exhibit much more depression, anxiety, disordered eating, and somatic symptomatology than men.

REFERENCES Bengtson, V L, & Dunham, C. (1986). Conceptual and theoretical perspectives of generational relations. In N. Datan, A. Green, & H. Reef (Eds.), Life span developmental psychology: Intergenerational networks. Hillside, NJ: Eribaum. Culbertson, F. M. (1997). Depression and gender: An international review. American Psychologist, 52, 25-34. Koss, M. P, Goodman, L. A., Browne, A., Fitzgerald, L. F, Keita, G. P, & Russo, N. F. (1994). No safe haven: Male violence against women at home, at work, and in the community. Washington, DC: American Psychological Association. Landrine, H., Klonoff, E. A., Gibbs, J., Manning, V, & Lund, M. (1995). Physical and psychiatric correlates of gender discrimination: An application of the schedule of sexist events. Psychology of Women Quarterly, 19, 473-492. Lucas, A. R., Beard, M., O'Fallon, W M., & Kurland, L. T (1991). 50-year trends in the incidence of anorexia nervosa in Rochester, Minn.: A population-based study. American Joumal of Psychiatry, 148, 917-922. Martz, D. M., Handley, K. B., & Eisler, R. M. (1995). The relationship between feminine gender role stress, body image, and eating disorders. Psychology of Women Quarterly, 19, 493-508. Maser, J. D., & Cloninger, C. R. (1990). Comorbidity of anxiety and mood disorders: Introduction and overview. In J. D. Maser & C. R. Cloninger (Eds.), Comorbidity of mood and attxiety disorders. Washington, DC: American Psychiatric Press. McGrath, E., Keita, G. P, Strickland, B. R., & Russo, N. P (1990). Women and depression: Risk factors and treatment issues. Washington, DC: American Psychological Association. Mussen, P, & Rutherford, E. (1963). Parent-child relations and parental personality in relation to young children's sex-role preferences. Child Development, 34, 589-607. Pribor, E. F, & Dinwiddie, H. (1992). Psychiatric correlates of incest in childhood. American Joumal of Psychiatry, 149, 52-56. Radloff, L. S., & Locke, B. A. (1986). The Community Mental Health Assessment Survey and the CES-D Scale. In M. M. Weissman, J. K. Myers, & C. E. Ross (Eds.), Community sun'eys of psychiatric disorders. New Brunswick, NJ: Rutgers University Press. Silverstein, B., & Blumenthal, E. (1997). Depression mixed with anxiety, somatization, and disordered eating: Relationship with gender-role-related limitations experienced by females. Sex Roles, 36, 709-724. Silverstein, B., Caceres, J., Perdue, L., & Cimarolli, Y (1995). Gender differences in depressive symptomatology: The role played by "anxious somatic depression" associated with gender-related achievement concerns. Sex Roles, 33, 621-636. Silverstein, B., Clauson, J., McKoy, E., Perdue, L, & Raban, J. (1995). The correlation between depression and headache: The role played by generational changes in female achievement. Joumal of Applied Social Psychology, 25, 35-48. Silverstein, B., Clauson, J., Perdue, L., Carpman, S., & Cimarolli, V (1998). The association between female college students reports of depression and their perceptions of parental attitudes regarding gender. Joumal of Applied Social Psychology, 28, 537-549.


555

Silverstein, B., Perdue, L , Wolf, C , & Pizzolo, C. (1988). Bingeing, purging, and estimates of parental attitudes regarding female achievement. Sex Roles, 19, 723-733. Silverstein, B., & Perlick, D. (1991). Gender differences in depression: Historical changes. Acta Psychiatrica Scandinavica, 84, 327-331. Silverstein, B., & Perlick, D. (1995). The cost of competence. New York: Oxford University Press. Silverstein, B., Perlick, D., Clauson, J., & McKoy, E. (1993). Depression combined with somatic symptomatology among adolescent females who report concerns regarding maternal achievement. Sex Roles, 28, 637-653. Swartz, M., Blazer, D., Woodbury, M., George, L , & Landerman, R. (1986). Somatization disorder in a U.S. Southern community: Use of a new procedure for analysis of medical classification. Psychological Medicine, 16, 595-609.

Weissman, M. M., Bruce, M. L , Leaf, P J., & Holzer, C , III (1991). Affective disorders. In L. N. Robins & D. A. Regier (Eds.), Psychiatric disorders in America: The Epidemiologic Catchment Area Study New York: Free Press. Wilson, G. T, & Eldredge, K. L. (1991). Frequency of binge eating in bulimic patients: Diagnostic validity. Intemational Joumal of Eating Disorders, 10, 557-561. Winfield, I., George, L. K., Swartz. M., & Blazer, D. G. (1990). Sexual assault and psychiatric disorders among a community sample of women. American Joumal of Psychiatry, 147, 335-341. Young, M. A., Fogg, L. F, Scheftner, W. A., Keller, M. B., & Fawcett, J. A. (1990). Sex differences in the lifetime prevalence of depression: Does varying the diagnostic criteria reduce the female/male ratio. Joumal of Affective Disorders, 18, 187-192. Young, M. A., Scheftner, W A., Fawcett, J., & Klerman, G. L. (1990). Gender differences in the clinical features of unipolar major depressive disorder. Joumal of Nen'ous and Mental Disease, 178, 200-203.