American Sociological Association

SPQ Markowitz Snap

Stigma, Reflected Appraisals, and Recovery Outcomes in Mental Illness




Fred E. Markowitz

Northern Illinois University

Beth Angell

Rutgers University

Jan S. Greenberg

University of Wisconsin-Madison


Drawing on modified labeling theory and the reflected appraisals process, using longitudinal data from 129 mothers and their adult children with schizophrenia, we estimate models of the effects of mothers’ stigmatized identity appraisals of their mentally ill children on reflected and self-appraisals, and how appraisals affect outcomes (symptoms, self-efficacy, life satisfaction). Implications for modified labeling theory and social psychological processes in recovery from mental illness are discussed. 

For persons diagnosed with a mental illness, dealing with the many difficulties that the illness brings, including the management of symptoms that can interfere with functioning, regaining a positive sense of self, and leading a productive and satisfying life has come to be conceptualized as the process of recovery (Ralph and Corrigan 2005). Recovery is not considered an endpoint, but an ongoing process where these elements fluctuate over time, and may gradually improve. Personal accounts and attempts by researchers to explain the process consistently point to certain core outcomes, involving symptoms of the illness, self-concept (e.g., esteem, efficacy, identity), and socioeconomic well-being (e.g., employment, housing, relationships). While studies on recovery have begun to consider more elaborate causal processes that tie outcomes together, many of the social psychological aspects of recovery, especially those involving stigma and identity, with limited exception, remain unexamined (Markowitz 2001). 

            The stigma associated with mental illness is a major impediment to recovery, having the potential to transform a person’s identity “...from a whole and usual person to a tainted, discounted one (Goffman 1963, p. 3).” Mental illness is linked to an array of negative stereotypical traits (e.g., dangerousness, incompetence), it is somewhat misunderstood by the general public, and is often inaccurately and negatively portrayed in the media (Phelan et al. 2000; Martin, Pescosolido, and Tuch 2000). Research based on modified labeling theory has shown how internalized stigma is related to the loss of socioeconomic status, restricted social networks, lowered self-esteem, and diminished quality of life (Link and Phelan 2001; Markowitz 1998; Rosenfield 1997). Research has not fully examined how the stigmatizing attitudes and responses of others may impede recovery, however. Moreover, it has not reconciled the role of disturbing symptomatic behavior as causes of stigmatizing responses that, in turn affect recovery outcomes.   

            The role of the self-concept in stigma and recovery processes is a central concern, yet modified labeling theory-based research has been limited to studies of how stigma negatively affects global self-evaluation (esteem) and self-efficacy (sense of personal control) among persons with mental illness (Markowitz 1998; Rosenfield 1997; Wright and Gronfein 2000). More recently though, studies have begun to examine how stigma affects self-identities, or the ways in which persons with mental illness are viewed by self and others (Kroska and Harkness 2006, 2008). Critically, however, research has not specified an integrated model of stigma, one that links identity with symptoms and functioning. In this study, we seek to extend this line of research by incorporating insights from modified labeling theory with the reflected appraisals process of self-concept formation to examine how mental illness leads to a “spoiled identity” in terms of stigmatized self-image characteristics(e.g., unreliable, unintelligent, immature, and incompetent) that, in turn, affect recovery outcomes.


Labeling theory is an important explanatory framework that accounts for the effects of stigma associated with devalued statuses, such as “delinquent” or “mentally ill” (Becker 1963; Lemert 1967; Scheff 1984). The theory is rooted in the symbolic interactionist perspective within sociology (Blumer 1969; Mead 1934; Stryker 1980). One of the premises of symbolic interactionism is that responses to persons and actions are based on socially constructed meanings (“definitions of the situation”) that are drawn from shared cultural knowledge. Within this framework, self-conceptions result from perceptions of how significant others (e.g., family, friends, and teachers) view the self—the reflected appraisals process (Kinch 1963). Based on others’ responses to the self, we come to see ourselves as we think others see us (Cooley 1902). Self-conceptions that are linked to occupied social positions are role-identities (Stryker 1980). Persons occupy many normative roles (e.g., employee, spouse, and parent) with accompanying behavioral expectations. According to labeling theory, through mental health treatment, persons may be cast in the non-normative role of “mentally ill.” 

            Early versions of labeling theory specified the process by which deviant labels are applied and persons’ self-conceptions and social opportunities are altered.  Scheff (1984) argued that when behavior continually violates social norms, is highly visible (e.g., in “crisis” situations), and is regarded as serious, it may be viewed as evidence of “mental illness,” resulting in assignment of a psychiatric diagnosis. Scheff’s theory emphasized the role of formal labeling in setting into motion stigmatizing processes that lead to sustained symptomatic behavior(1984). Propositions of Scheff’s labeling theory have been strongly contested, with critics charging that negative outcomes for persons with mental illness are due to the impairment caused by the symptoms of mental disorder, not the stigma of labeling—a more strict ‘psychiatric’ framework (Gove 1982).Symptoms of many forms of mental illness are associated with social withdrawal, loss of interest in activities, irritability, and non-normative emotional responses. These symptoms can make social interaction and role performance very difficult. As a result, persons may be judged by others negatively—for example, as less competent, unpredictable, or potentially harmful. Rather than adopt a ‘symptoms versus stigma’ approach to understanding outcomes, models that incorporate both allow for a fuller understanding of the trajectory of recovery. 

In Link and colleagues’ modified labeling theory, the strong claim made by Scheff that labeling causes ‘careers in residual deviance’ is replaced by a more subtle approach to how stigma affects the course of illness and functional outcomes among those labeled through mental health treatment (Link 1987; Link, Mirotznik, and Cullen 1991). According to the theory, widely held stereotypical attitudes about persons with mental illness (e.g., as incompetent and dangerous) become personally relevant to an individual when diagnosed with a mental illness. Because of these attitudes, those diagnosed expect to be devalued and discriminated against. This anticipated rejection is directly experienced as demoralizing (i.e., self-esteem is lowered and depression increases). Also, to avoid rejection, persons who are labeled are hypothesized to adopt coping orientations, such as secrecy, disclosure, or social withdrawal, enhancing the effects of expected rejection by constricting social networks, leading to unemployment and lowered income. Thus, stigmatizing beliefs act as self-fulfilling prophecies. Drawing on the stress-process model(Pearlin et al. 1981), the theory further predicts that lowered self-esteem, constricted interpersonal networks, unemployment, and low income increase stress. Stress, in turn, places persons at risk for increased symptoms. In this way, while labeling and stigma are not the sole cause of sustained mental illness, theyindirectly affect the course of illness through changes in the self-concept and key social outcomes. Despite these advances, modified labeling theory studies have not fully incorporated the role of disturbing symptomatic behavior as causes of stigmatizing responses, and with limited exception (Kroska and Harkness 2006), studies have not considered the identity dimension of self-concept.

Stigma, Self-Concept, and the Reflected Appraisals Process

            The role of the self-concept in stigma processes is a central concern, yet modified labeling theory-based research has been limited mostly to examination of how expectations of rejection negatively affect global self-evaluation (esteem) and self-efficacy (i.e., sense of personal control) among persons with mental illness (Markowitz 1998; Rosenfield 1997; Wright and Gronfein 2000). It has only recently examined how stigma affects other dimensions of self, including self-meanings or personal attributes (Kroska and Harkness 2006, 2008). Stigma may adversely impact these dimensions of self, having important consequences for recovery. Persons who, for example, consider themselves as less competent, capable, or successful may feel demoralized and act in ways that live up to stigmatized self-images that reduce their quality of life by not making friends, furthering their education, or seeking jobs. Diminished quality of life may then lead to an increased risk of symptoms of psychiatric disorder (Markowitz 2001). 

            In a recent study, Kroska and Harkness (2006) show how, among persons with a serious mental illness, “stigma sentiments”—the evaluation (e.g., good vs. bad), potency (e.g., strong vs. weak), and activity (sharp vs. dull) (EPA) profile of “a person with mental illness” is related to corresponding dimensions of reflected appraisals (“how others see me”) and self-identities (“myself as I really am”). We build on their novel approach in several ways. First, we include the appraisals of the person with mental illness by a significant other, a key component of the reflected appraisals process.  Second, although there is some overlap, rather than the more general EPA dimensions of identity appraisals, we focus on a more specific set of characteristics associated with mental illness. Finally, we examine the causal impact of stigmatized identities—through the reflected appraisals process—on outcomes that represent dimensions of recovery for persons with mental illness. The model we propose is presented in Figure 1. It posits that initial levels of symptoms, functioning, and self-evaluation are likely to impact the ways in which significant others think about the person with mental illness. The more symptomatic and less socially engaged and competent persons with mental illness are, they are likely to be seen in more stigmatized terms by significant others. Persons with mental illness pick up on these appraisals expressed to varying degrees (reflected appraisals), which may then affect their identities (self-appraisals) in stigmatized ways. To the extent that persons see themselves in stigmatized terms, this is likely to adversely affect their symptoms/functioning and self-evaluation. Alternatively, to the extent that symptoms and functioning is not affected by others’ and self-appraisals, this suggests that it is simply the degree of stability in the underlying illness, rather than stigmatized identity that determines outcomes, in line with a more strictly medical, or “psychiatric” perspective.

Figure 1. Stigma, Reflected Appraisals, and Recovery Outcomes


           According to the reflected appraisals process, the self-concept is shaped, in large part, by the perceived responses of significant others, such as family, friends, or teachers (Gecas and Burke 1995; McCall and Simmons 1966; Stryker 1980). We focus on one of the most significant groups in the lives of persons with mental illness—family members.  The vital role of family members as caregivers, and the attendant burdens carried by these roles, is long recognized. So too, the role of the family’s emotional climate in contributing to relapse and other negative outcomes in the consumer is the subject of an extensive body of research on “expressed emotion” (see Avison 1999a,b). The study of stigma and families, however, remains limited to describing how stigma impacts the family members of persons with mental illness (Phelan, Bromet, and Link 1998; Struening et al. 2001). Even though families are often the targets of “courtesy” stigma, they may also inadvertently act as sources of stigma to their mentally ill family members. For example, studies of persons with mental illness report that, after employers and the general population, family members and mental health providers are a frequent source of stigmatizing responses, such as viewing respondents as less than competent, lacking understanding, making offensive comments, and expressing concern about potential dangerousness (Dickerson, Sommerville, and Origoni 2002). These expressions – much like the critical or over-involved comments that are implicated in the “expressed emotion” literature—are likely to impact the way that persons with mental illness think about themselves. Prior research, for example, has shown that significant others’ expectations are associated with role performance and the quality of patient-family relationships (Barrowclough et al. 2001; Greenley 1979). We hypothesize that the attitudes of family members of those with mental illness are important because they shape how those with mental illness come to think of themselves, in turn affecting recovery outcomes.

            Model Specification

            Following a similar approach used in prior research, we developed a set of 22 semantic-differential type measures of personal attributes consistent with mental illness stereotypes (e.g., safe/dangerous, success/failure, trustworthy/untrustworthy, unintelligent/intelligent, gentle/violent, competent/incompetent”) (Burke and Tulley 1977; Hoelter 1984; Schwartz and Stryker 1970). These items were administered to family members to assess significant others’ appraisals (e.g., “John is… trustworthy/untrustworthy, unintelligent/intelligent,” etc.). The same set of items was administered to those with mental illness to assess their reflected appraisals (e.g., “My mother thinks I am… trustworthy/untrustworthy, unintelligent/intelligent,” etc.) and their self-appraisals (e.g., “I am… trustworthy/untrustworthy, unintelligent/intelligent,” etc.). These measures are then included in a series of models where initial levels of symptoms, self-esteem, efficacy, functioning, and quality of life affect family appraisals of patients along stigmatized identity dimensions. Significant others’ appraisals, in turn, are predicted to influence reflected appraisals (mentally ill persons’ perceptions of their families’ view of themselves). Self-appraisals, in turn, affect outcomes. The model implies a set of mediated relationships where outcomes are due, in part, to appraisals.



Data were used from a longitudinal study of aging mothers of adult children with mental illness. We use data from waves 3 and 4 (collected at an 18–month interval) of the study since they contain the measures relevant to our model. Mothers were the primary family respondent and were recruited through community support programs, the local media, and the National Alliance for the Mentally Ill (n = 129).

Appraisal Measures

Friendly           Unfriendly

Trustworthy    Untrustworthy

Safe                 Dangerous

Unintelligent   Intelligent

Competent      Incompetent

Success            Failure

Able                Unable

Easygoing       Difficult

Organized       Disorganized

Dirty                Clean


A series of 22 items, referenced for self appraisals (“I am…”), reflected appraisals (“…my mother thinks I am…”), and significant other appraisals (“my son/daughter is…”) were included in the study.  Example items are listed below.

Outcome Measures

The Brief Symptom Inventory consists of 52 self-reported items several dimensions of symptoms in the past month, including somatization, obsession-compulsion, interpersonal sensitivity, depression, anxiety, hostility, paranoid ideation, and psychoticism (Derogatis and Melisaratos 1983).  Subjective life satisfaction was assessed using 22 items adapted from Lehman’s (1988) scale that asks respondents how they feel about living arrangements, family and social relationships, leisure activities, finances, employment, safety, and health. Self-efficacy (mastery)is measured by the average score on the widely used 8–item scale developed by Pearlin et al. (1981) that reflects the extent to which persons believe they have a sense of mastery, or personal control over circumstances and events in their lives.


There were several key findings. First, not surprisingly, symptoms and functioning are related to how family members think about their ill family members, how persons with mental illness think others perceive them, and how they perceive themselves, in terms of potentially stigmatized personal characteristics. Consistent with our hypotheses, initial levels of symptoms, self-efficacy, and quality of life are linked to the manner in which mothers appraise their sons and daughters with mental illness—those who are doing better in terms of symptoms and social well-being, as well as those with higher levels of personal control, are perceived by their mothers as more competent, capable, and healthy compared to those who are doing less well. The effects of these variables on reflected appraisals are explained, in part, by mothers’ appraisals. Also, part of the effects of symptoms, self-efficacy, and life-satisfaction on self appraisals is explained by others’ appraisals and reflected appraisals.

            Second, our findings show that mothers’ appraisals are reflected in patients’ perceptions of what their mothers think about them (reflected appraisals). Although it is impossible to test how appraisals are conveyed to the consumer using these data, one possible interpretation is that negative feedback (e.g., “you’re like a child, not an adult”) provided by mothers is at times conveyed directly, out of frustration and stress, and is part of the “conflicted relationships” that are often a consequence of mental illness (Karp 2001). Indeed, a body of research suggests that caregiver criticism (termed “expressed emotion”), both as rated by observers and as perceived by those with mental illness, is robustly associated with risk of symptomatic relapse (Butzlaff and Hooley 1998). Negative feedback can also be conveyed in a number of subtle ways that stigmatize persons with mental illness, such as sustained social exclusion or insidious comments (Dickerson et al. 2002; Jenkins and Carpenter-Song 2009). 

Third, beyond the link between mothers’ and reflected appraisals is our finding that reflected appraisals are related to self-appraisals, but not strongly so. The direction of the coefficients are, however, consistent with prior research showing that the ways in which persons think significant others perceive them affects their self-conceptions (Felson 1981, 1985, 1989).  It could be that persons with mental illness incorporate others’ appraisals into their self-appraisals in a less cognitive process, or that there is a greater degree of ambiguity in their perceptions of others’ appraisals. This leaves extension of the generalizability of the reflected appraisals process from dimensions of identity associated with academic and athletic ability and physical attractiveness among young persons to dimensions of stigmatized identity associated with serious mental illness among adults somewhat tentative. 

            Fourth, our findings show that, despite high levels of stability, symptoms, self-efficacy, and life-satisfaction are affected by stigmatized self-conceptions, consistent with the reflected appraisals process. It appears however, that self-appraisals have a comparatively greater impact on outcomes compared to mothers’ appraisals, the effects of which are not mediated by reflected appraisals. Together, these findings suggest that perhaps beyond clinical intervention (medication, counseling) implied by a strict medical model approach, recovery is, at least to some extent, a process that is influenced by the expectations and feedback provided by significant others in the lives of persons with mental illness. Significant others’ positive appraisals exert an effect that may be similar to that of social support. The presence of positive identity-related feedback may reduce symptoms while negative feedback may facilitate sustained symptoms. Moreover, stigmatized self-conceptions may reduce sense of control, empowerment in treatment programs, motivation to seek jobs and make friends, and thus contribute to diminished quality of life.

            The finding of a link between significant others’ appraisals and recovery outcomes is also consistent with previous research on expressed emotion (Greenley 1986). Perhaps critical comments from relatives induce shame that is directly internalized by the ill family member, thus leading persons with mental illness to think and act in ways that inhibit recovery. The present study suggests that perceived criticism on the part of the consumer is not simply illusory or an artifact of paranoid symptoms, but is, at least in part, a reflection of the opinions of care-giving family members. 


In this study, we applied the reflected appraisals process model to examine how stigmatized identity affects dimensions of recovery for persons with mental illness.  The findings of our study suggest a potential direction for extending modified labeling theory. An important step in further study is the need to include measures such as devaluation-discrimination beliefs in the model to examine how more widely held stigmatizing attitudes towards mental illness (expectations held by “most people in the community”) influence recovery through their effects on others’, reflected, and self-appraisals. In this way, following symbolic interactionist theory, we can link the attitudes of the “generalized other” with those of “significant others” to better understand self-concept formation and outcomes among persons with mental illness. 

            Mental illness represents a challenge to the study of self and identity, leading to several questions that may be guided by identity theories (Stets and Burke 2005). Unlike other medical conditions that have the potential to transform identity in positive ways (e.g., “cancer survivor”), because of its potential for disturbing behavior and the powerful stigma it carries, mental illness is likely to affect the self in more adverse ways (Albee and Joffe 2004). These effects may not be straightforward, however. For example, to what extent are the adverse effects of stigma contingent upon the salience of mental illness as a role-identity dimension relative to other dimensions? As those who have written about recovery indicate, work and social relationships are important sources of self-worth, offsetting the stigma of a diagnosis of mental illness, as well as helping to buffer the additional stresses that illness creates (Ralph and Corrigan 2005). Also, in terms of relationships with significant others, while we have emphasized consistency in stigmatized identity appraisals, how can discrepancies between role-identity expectations and performance be understood? For example, to what extent are appraisals affected by more specific discrepancies between the normative role expectations held by others (e.g., as sons or daughters) and behavior related to those roles? Similarly, how does the imbalanced exchange created by care-giving and the disruptions and dependency created by mental illness lead to further stigmatizing attitudes held by family members? 

            In sum, this study highlights the notion that recovery from mental illness is not simply a matter of controlling symptoms as indicated by a strictly ‘psychiatric’ perspective, but that it is, to a certain extent, a social-psychological process. The ways in which people think about persons with mental illness affects the beliefs and actions of those with mental illness, in turn shaping the trajectory of illness. Despite some limitations, given the generally favorable results of the present study, our preliminary study suggests that integrating modified labeling theory with reflected appraisals and identity formation processes may help further our understanding of how stigma impedes recovery.


Get Involved


ASA needs you to serve the discipline

Join a Section

Explore ASA's community of specialists in the discipline

Join ASA

Join or renew your membership