• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of amjphAmerican Journal of Public Health Web SiteAmerican Public Health Association Web SiteSubmissionsSubscriptionsAbout Us
Am J Public Health. 2006 October; 96(10): 1766–1771.
PMCID: PMC1586134

Public Conceptions of Serious Mental Illness and Substance Abuse, Their Causes and Treatments: Findings from the 1996 General Social Survey

Sara Kuppin, DrPH and Richard M. Carpiano, PhD, MA, MPH


Objectives. We examined the degree to which lay beliefs about the causes of disorders may predict beliefs about what constitutes appropriate treatment.

Methods. We analyzed randomized vignette data from the MacArthur Mental Health Module of the 1996 General Social Survey (n=1010).

Results. Beliefs in biological causes (i.e., chemical imbalance, genes) were significantly associated with the endorsement of professional, biologically focused treatments (e.g., prescription medication, psychiatrists, and mental hospital admissions). Belief that the way a person was raised was the cause of a condition was the only nonbiologically based causal belief associated with any treatment recommendations (talking to a clergy member).

Conclusions. Lay beliefs about the biological versus nonbiological causes of mental and substance abuse disorders are related to beliefs regarding appropriate treatment. We suggest areas for further research with regard to better understanding this relationship in an effort to construct effective messages promoting treatment for mental health and substance abuse disorders.

Persons with mental health or substance abuse disorders may not seek help, at least not in the form of professional treatment (i.e., psychotherapy, medication), because these treatment options are perceived by the person, family, or social network as inappropriate or undesirable.1,2 The perceived undesirability or inappropriateness of professional treatment sources for mental health and substance abuse disorders by persons with these disorders is a major barrier to adequate treatment of mental health and substance abuse problems.1,2 Previous research indicates that the American lay public overwhelmingly agrees with mental health and substance abuse experts that mental health disorders are serious and that they require treatment.3 But what type of treatment does the public believe will help? Answering this question is essential so that public health professionals can construct and convey effective public health messages concerning the treatment of mental health and substance abuse disorders to the public.

Although recent research has assessed trends in lay beliefs about the causes of mental health and substance abuse problems,4 few studies have examined lay beliefs about treatment for these disorders. The literature suggests that the lay public demonstrates a preference for informal sources of help and holds a relatively negative view of, or unwillingness to use, professional—and particularly medical—sources of help, such as psychiatry and psychiatric medication.2,57 However, almost no research has examined the relation between causal beliefs and beliefs about treatment. Nevertheless, a recent study that used a small sample of undergraduates suggests that laypersons are more likely to perceive treatment options for depression, including psychiatric medication, as more helpful if causal beliefs are aligned with treatment focus.8

Understanding the relation of causal beliefs to beliefs about treatment for these disorders is essential for mental health and substance abuse treatment professionals who must construct effective messages to the public about the appropriate diagnosis and treatment of these disorders. Although beliefs may not be predictive of actual behavior, beliefs do contribute to public sentiment about the cause and appropriate treatment for mental illness and substance abuse disorders, and this in turn provides the context in which individual decisions about treatment are made.9 Therefore, better understanding the relationship between causal and treatment beliefs may illuminate the relationship between causal beliefs and the actual treatment-related behavior (such as help-seeking, treatment adherence, and continuation) of individuals within a community.

Recent messages to the public regarding the treatment of mental health and substance abuse disorders from medical, pharmaceutical, insurance, and some advocacy sources focus on biological causes of these disorders and typically emphasize biologically focused treatment options. Although it may seem logical that biologically focused treatments such as medication and other types of prescribed therapies would be the treatment choice for persons who perceive a biological cause for these disorders, we sought to evaluate this assumption by testing the following related hypotheses:

  1. Lay public beliefs regarding the appropriateness of specific treatments vary among 4 mental health conditions (depression, schizophrenia, alcohol abuse, and drug abuse).
  2. Beliefs about the cause of a disorder are associated with beliefs regarding the appropriateness of specific treatments.
  3. Biologically based (vs nonbiologically based) causal beliefs are more strongly associated with the endorsement of biologically focused treatment options.



The data for this study were obtained from the responses of subjects surveyed with the MacArthur Mental Health Module, a 57-item interview schedule, of the 1996 General Social Survey.3,10 Conducted since 1972, the General Social Survey is considered to be the premiere survey of US public opinion.3 Personal interviews were conducted with a nationally representative sample of 2904 noninstitutionalized adults living in the 48 contiguous US states, with a response rate of 76%. Of the 1444 people administered the MacArthur Mental Health Module, 1010 respondents answered all treatment option questions and, therefore, constituted the study sample. More extensive detail on the General Social Survey and the MacArthur Module is available elsewhere.3,10

Survey Design

The MacArthur Mental Health Module contained an experimental design in which each respondent was presented with a randomly selected vignette about an individual experiencing symptoms of 1 of 5 different mental health– or substance abuse–related conditions.3 Four of these met Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV)11 criteria for 1 of the following conditions: major depression, schizophrenia, alcohol dependence, or drug (cocaine) dependence. Alternatively, some respondents were instead presented with a fifth “control” vignette featuring a “troubled person” who was experiencing common problems of daily life but did not meet criteria for a DSM-IV diagnosis. Within each vignette, the featured individual’s gender, race (White, Black, Hispanic), and education (degree) were varied randomly. The distribution of respondent demographic characteristics for each vignette condition can be found in Table 1 [triangle].

Respondent Descriptive Statistics for Each Vignette Condition (n = 1010): United States, General Social Survey, 1996

After reading the vignette, each respondent was presented with, and asked to rate, a set of possible causes of the vignette person’s condition as well as presented with, and asked to evaluate the appropriateness of, treatment options that the vignette person could choose to alleviate his or her condition. We refer to these topics, respectively, as causal beliefs and treatment options.

Measures and Analyses

We examined 8 treatment options: see a psychiatrist; take prescription medication; check into a mental hospital; see a general doctor; go to a therapist (“or counselor, like a psychologist, social worker, or other mental health professional”); join a self-help group (“where people with similar problems help each other”); talk to family/friends; and talk to a clergy member (“such as a minister, priest, rabbi, or other religious leader”). Each of these dichotomous treatment option variables was coded yes = 1 or no = 0.

Five causal beliefs were examined: chemical imbalance (“a chemical imbalance in the brain”), genetics (“a genetic or inherited problem”), stress (“stressful circumstances in his/her life”), way raised (“the way s/he was raised”), and bad character (“his/her own bad character”). Respondents were asked how likely it was that the situation of the person in the vignette might have been be caused by each of these. These causal belief items were originally collected on a 4-point scale (ranging from “very unlikely” to “very likely”) and were collapsed for the present study into dichotomous variables (unlikely=0 or likely=1).

Our analytic strategy involved 2 steps. First, each of the treatment option variables was cross-tabulated with the 5 vignette conditions to determine response distributions within and across conditions. Second, binary logistic regression models were used to test associations between the causal belief and treatment option variables. These models statistically adjusted for gender, race, age, education, and vignette condition (entered as 4 separate, dichotomously coded, “dummy variables,” with troubled person treated as the referent category). We report the odds ratios and 95% confidence intervals of these models.

The amount of missing data differed across the causal belief and treatment option variables, raising concern about possible bias being introduced into the findings. Further analyses were conducted to determine the degree to which missing data for these variables might influence the results. Complete cases (cases having complete information on all 13 variables) were compared with cases containing missing information on any of these variables with respect to gender, race, age, and education. No statistically significant differences were found between the complete and incomplete cases except in terms of age (the mean age of complete cases was approximately 4.5 years younger than the cases with any missing data).


Treating Mental Illness and Substance Abuse

Table 2 [triangle] presents the vignette-specific and overall distribution of treatment options.

Distribution of Treatment Options Considered Appropriate (Percentage and Odds Ratio), by Vignette Condition Received (n = 1010)

For each vignette-treatment option cell, the percentage represents the percentage of total respondents receiving that particular vignette who selected that treatment option as an appropriate course of action for the person in their vignette. The bivariate odds ratio can be interpreted as the likelihood that a respondent who received that vignette condition (vs receiving another vignette condition) perceived the treatment option as appropriate (vs inappropriate). The total percentage of respondents in the sample who advocated each treatment option is indicated in the final row. (Although we present unadjusted odds ratios, additional analyses were conducted with binary logistic regression models that adjusted for respondents’ sociodemographics—the odds ratios obtained from both sets of analyses were nearly identical.)

Although the results indicate significant variation in treatment options according to the vignette condition in question, with respect to the total sample, all nonbiologically focused treatment options were more popular among the respondents than the biologically focused treatments. For example, talking with family and friends was advocated by almost 96% of the sample, whereas taking prescription medication was endorsed by less than 60%.

Depression and schizophrenia were the 2 conditions in which professional, biologically focused treatment options were most commonly viewed as appropriate. For example, 77.4% of the respondents who received the depression vignette felt that a psychiatrist was an appropriate treatment option for this condition. Compared with those who received other vignettes, respondents receiving the depression vignette were 1.6 times more likely to report a psychiatrist as an appropriate treatment option. Additionally, depression was significantly associated with recommendations for prescription medication and a general doctor, as well as therapist, which was the only nonbiologically focused, professional treatment significantly associated with this vignette. Similarly, receiving the schizophrenia vignette was significantly associated with seeing a psychiatrist and taking prescription medication (although the magnitudes of these associations were much greater, as indicated by odds ratios of 6.2 and 4.7, respectively). These respondents were also 3.6 times more likely to report checking into a mental hospital as an appropriate treatment and, conversely, approximately 1.75 times less likely to recommend talking to clergy.

The results for alcohol and drug dependence differed from the results for depression and schizophrenia in that they were less associated with seeking medical treatments than with therapists and self-help groups. Respondents who received the alcohol dependence vignette were more than 2 times less likely to report prescription medication as appropriate treatment. However, they were 2 times more likely to choose a therapist and more than 5 times more likely to select a self-help group. Drug dependence showed a similar pattern with respect to medication and therapists, but, unlike respondents who received the alcohol dependence vignette, respondents who received this vignette were significantly more likely to advocate checking into a mental hospital.

Last, a troubled person was associated with several medical and nonmedical treatment options. As indicated by the consistent pattern of odds ratios that are less than 1, these respondents were significantly less likely than those who did not receive this vignette to view any of these treatment options as appropriate.

Preferred Treatment Options

Table 3 [triangle] presents the results for multivariate binary logistic regression models that test the association between causal beliefs and treatment options. These models attempt to account for multiple causal beliefs that a respondent may have regarding the condition of the person in their vignette. For example, a respondent may believe that a chemical imbalance (a physical cause) and stress (a social cause) could each be likely causes for the situation experienced by the person in the vignette. Consequently, beliefs about these causes may entail different perceptions of appropriate treatment. Therefore, to account for the variation that multiple beliefs may be present in determining the appropriateness of treatment options, all the causal belief variables were simultaneously entered into the same model for each treatment outcome, also with control for gender, race, age, education, and vignette condition. This modeling approach has been used by others in previous research on mental illness perceptions.12

Binary Logistic Regression Models (Odds Ratios With 95% Confidence Intervals) of Treatment Options Regressed on Likely Causal Beliefs About Vignette Conditions (n = 1010)

The results indicate that some causal beliefs were associated with perceptions of treatment appropriateness. Chemical imbalance, genetics, and the way a person was raised were associated with several treatment options. Furthermore, a pattern is evident in the results whereby biologically focused treatment options (i.e., psychiatrist, prescription medication, mental hospitalization, and general doctor) were significantly more likely to be endorsed by those who perceived the cause to be biological.

Chemical imbalance and genetics (the only biologically based causes) were the only 2 causal beliefs associated (P ≤ .05) with any of the biologically focused treatment options. Chemical imbalance being viewed as a likely cause was significantly associated with all biologically focused treatment options as well as therapist, a nonbiologically focused formal treatment option.

Genetics also showed a pattern of association with biologically focused formal treatment. However, it was only associated with the formal treatments of prescription medication and mental hospitalization. Consistent with this preference for biologically focused, formal treatment, this causal belief also demonstrated an inverse association with talking with family and friends—respondents who believed that their vignette condition had a genetic cause were more than 2.2 times less likely to view talking with family or friends as an appropriate treatment option.

For nonbiologically focused causes, the way a person was raised was the only causal belief associated with any treatment options. Respondents who attributed their vignette condition to this cause were 1.7 times more likely to perceive talking to a clergy member as an appropriate treatment.


To promote adequate treatment of mental health and substance abuse disorders, it is essential to know how the lay public conceptualizes appropriate treatment for these disorders. When we examined this question with randomized vignette data from the MacArthur Mental Health Module of the 1996 General Social Survey and looked at totals, we found that nonbiologically focused treatment options were more commonly viewed as appropriate than biologically focused treatments. This finding is consistent with those of Swindle et al.5 Although it could be argued that “troubled person” is affecting this distribution—that respondents receiving this vignette were less likely to endorse professional treatments and therefore were skewing the overall distribution of treatment preferences—this preference for nonbiologically focused treatments persisted even when all respondents receiving the troubled person vignette were removed from the analysis. Interestingly, prescription medication was the second least likely option to be viewed by respondents as helpful (59.7%). This finding stands in contrast to the increasing use of medication to treat mental health and substance abuse disorders in the last few decades.

When compared with respect to the vignette conditions, beliefs regarding treatment appropriateness significantly varied according to the mental health or substance abuse condition in question. Medication was perceived as appropriate treatment for mental health disorders, as was a psychiatrist, general physician, and therapist. These treatments each received high levels of endorsement (greater than 75%) for their respective conditions. However, with respect to alcohol and drug dependence, prescription medication was significantly less likely to be perceived as an appropriate treatment option. By contrast, nonbiologically focused treatment options, such as a therapist, talking to family/friends, and a self-help group were significantly more likely to be viewed as appropriate solutions for these conditions.

These findings indicate that, despite their general preference for informal treatment, the majority of the lay public does recognize the need to seek formal (particularly biologically focused) treatment for depression and schizophrenia. However, the public is far more reluctant to endorse biologically focused treatments for substance abuse. Even though hospitalization was significantly endorsed as a treatment for drug dependence, therapists and self-help groups were more than twice as popular (as indicated by the percentages of respondents receiving these vignettes who endorsed such treatment).

Considering the association between causal beliefs and treatment preferences, our results also indicate that in the United States, the lay public’s beliefs about what treatment options are appropriate for mental illness and substance abuse are related to beliefs about the cause of the problem. Specifically, those who supported biological causal explanations of mental illness and substance abuse (62.0% for chemical imbalance and 47.1% for genetics) were significantly more likely to advocate biologically oriented formal interventions (e.g., a psychiatrist, prescription medication, hospitalization, and a general physician), compared with respondents who did not share this causal attribution. As for more socially rooted causal explanations of mental illness and substance abuse, only the belief that the condition was caused by the way an individual was raised was significantly related (P ≤ .01) to talking to a clergy member, a nonmedical, informal treatment option. Belief that stress or bad character was the cause of the vignette subject’s condition (endorsed, respectively by 87.8% and 46.4% of the entire sample) was not significantly predictive of any specific treatment option.

Overall, these findings suggest that beliefs in biological cause are significantly associated with beliefs about treatment appropriateness, and specifically beliefs about the appropriateness of formal treatment options. Although some respondents supporting biological causation also advocated informal sources of treatment, and some respondents perceiving more socially oriented causes viewed biological treatments as appropriate, only chemical imbalance or genetic causal beliefs were significantly more predictive (vs more socially rooted causal beliefs) of viewing formal treatment options as appropriate. Interestingly, nonbiologically oriented causal beliefs were not predictive of nonbiologically focused treatment recommendations.

One possible explanation for the association between biologically oriented causal and treatment beliefs, but not between nonbiologically focused beliefs and treatments, may be related to messages to the public over the past decade taking effect: the medical, insurance, and pharmaceutical industries as well as some antistigma groups in the mental health and substance abuse advocacy community have heavily emphasized the biological aspects of cause and treatment of these disorders. It is also important to consider the possibility that some respondents did not view the treatment option choices as distinctly biological versus non-biological, therefore blurring this distinction. For example, some respondents might have believed that general doctors are capable of focusing on issues of concern in a manner similar to that of therapists. Analysis of other items in the MacArthur Module not included in this study concerning respondents’ familiarity with mental illness and mental health treatment providers (either personally or in relation to others) may provide further insight about how respondents viewed the biological versus nonbiological emphases of various treatment options.

These findings offer important insights for furthering our understanding of how we think about the discrepancy between mental illness and substance abuse prevalence and treatment seeking and adherence. However, they must be considered with respect to the potential limitations of this study. Respondents were asked to express beliefs about causes and helpful treatment in response to a fictional person in a vignette, which may not reflect the behavior of respondents should they be faced with a similar situation in real life involving themselves or a loved one. Additionally, our study, although offering insight into treatment behavior, did not consider factors such as stigma or other treatment barriers (e.g., financial constraints and insurance issues), which may mediate the relation between causal beliefs and treatment beliefs. Finally, although respondents were presented with vignettes about 4 important mental health and substance abuse disorders, vignettes featuring other DSM-IV disorders (e.g., obsessive-compulsive disorder, posttraumatic stress disorder) may have yielded different results than those reported.

Further research is needed regarding why the US public does not endorse, overall, more formal, biologically oriented treatment options and, specifically, medication; these treatment options are the focus of many treatment messages and research from mental health and substance abuse experts. As mentioned earlier, stigma as well as other barriers to treatment, such as financial and insurance issues, may lead some to view formal treatment options as unhelpful, in spite of causal beliefs, because they are perceived as undesirable or inaccessible. In addition, more research is needed to explore the mechanisms by which causal beliefs influence treatment beliefs or vice versa. This study establishes that there is a relation between public beliefs about cause and the appropriate treatment of mental illness and substance abuse. Specifically, these findings indicate an association between biologically based causal and biologically focused treatment beliefs.

A richer understanding of the relation of causal beliefs to beliefs about appropriate treatment and its mediating factors is important for understanding the social context in which persons with mental health and substance abuse disorders are making decisions about treatment. Our findings suggest that effective messages to the public promoting formal treatment for these disorders need to address cause. Furthermore, the findings suggest that treatment providers should discuss cause with consumers and their families in cases in which formal treatment may be warranted to educate consumers and their families about possible causes of mental illness and substance abuse disorders, as well as gain important insights regarding specific beliefs that may influence consumer treatment behavior.


R. M. Carpiano co-authored this article while completing a Robert Wood Johnson Foundation Health and Society Scholar Fellowship at the University of Wisconsin, Madison. The authors thank Jo Phelan for her guidance and comments on earlier drafts of this article, as well as the reviewers for their valuable feedback.

Human Participant Protection
This study was approved by the Columbia University institutional review board.


Peer Reviewed

S. Kuppin originated the study and design, and R. M. Carpiano led the analysis of the project. Both authors contributed to the writing of the article.


1. Angermeyer MC, Matschinger H. Public attitude towards psychiatric treatment. Acta Psychiatr Scand. 1996;94:326–336. [PubMed]
2. Benkert O, Graf-Morgenstern M, Hillert A, et al. Public opinion on psychotropic drugs: an analysis of the factors influencing acceptance or rejection. J Nerv Ment Dis. 1997;185:151–158. [PubMed]
3. Pescosolido BA, Martin JK, Link BG, et al. Americans’ Views of Mental Health and Illness at Century’s End: Continuity and Change. Public Report on the Macarthur Mental Health Module, 1996 General Social Survey. Bloomington, Ind: Indiana Consortium of Mental Health Services Research, Indiana University; New York, NY: the Joseph P. Mailman School of Public Health, Columbia University; 2000.
4. Link BG, Phelan JC, Bresnahan M, Stueve A, Pescosolido BA. Public conceptions of mental illness: labels, causes, dangerousness, and social distance. Am J Public Health. 1999;89:1328–1333. [PMC free article] [PubMed]
5. Swindle R, Heller K, Pescosolido B, Kikuzawa S. Responses to “nervous breakdowns” in America over a 40 year period: mental health policy implications. Am Psychol. 2000;55:740–749. [PubMed]
6. Croghan TW, Tomlin M, Pescosolido BA, et al. American attitudes toward and willingness to use psychiatric medications. J Nerv Ment Dis. 2003;19:166–174. [PubMed]
7. Lauber C, Nortd C, Falcato L, Roessler W. Lay recommendations on how to treat mental disorders. Soc Psychiatry Psychiatr Epidemiol. 2001;36:553–556. [PubMed]
8. Iselin M, Addis M. Effects of etiology on perceived helpfulness of treatments for depression. Cognit Ther Res. 2003;27:205–222.
9. Strecher VJ, Rosenstock IM. The health belief model. In: Glanz K, Marcus Lewis F, Rimer BK, eds. Health Behavior and Health Education: Theory, Research, and Practice. 2nd ed. San Francisco, Calif: Jossey-Bass; 1997:41–59.
10. Davis JA, Smith TW. General Social Surveys, 1972–1996 [cumulative file]. Chicago, Ill: National Opinion Research Center. Ann Arbor, Mich: Inter-University Consortium for Political and Social Research; 1996.
11. Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition. Washington, DC: American Psychiatric Association; 1994.
12. Schnittker J, Freese J, Powell B. Nature, nurture, neither, nor: black-white differences about the causes and appropriate treatment of mental illness. Social Forces. 2000;78:1101–1132.

Articles from American Journal of Public Health are provided here courtesy of American Public Health Association
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • MedGen
    Related information in MedGen
  • PubMed
    PubMed citations for these articles

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...