STIGMA, SELF-CONCEPT AND STIGMA RESISTANCE AMONG INDIVIDUALS WITH MENTAL ILLNESS

A dissertation submitted to Kent State University in partial fulfillment of the degree requirements for the degree of Doctor of Philosophy

by Natalie Bonfine May 2013

Dissertation written by Natalie Bonfine B.A., Kent State University, 2003 M.A., Kent State University, 2005 Ph.D., Kent State University, 2013

Approved by

Christian Ritter Dr. Christian Ritter

, Chair, Doctoral Dissertation Committee

Richard Adams Dr. Richard Adams

, Members Doctoral Dissertation Committee

Emily Asencio Dr. Emily Asencio Kristen Marcussen Dr. Kristen Marcussen Kristin Mickelson Dr. Kristin Mickelson Sara Newman Dr. Sara Newman

Accepted by Richard T. Serpe Dr. Richard T. Serpe

, Chair, Department of Sociology

James L. Blank Dr. James L. Blank

, Dean, College of Arts and Sciences

ii

TABLE OF CONTENTS

LIST OF FIGURES ………………………………………………..………………….iv LIST OF TABLES…………………………………………………..…………………v ACKNOWLEDGEMENTS…………………………………………..………………vii CHAPTERS

Page

1

INTRODUCTION AND STATEMENT OF THE PROBLEM………………1

2

THEORETICAL FOUNDATIONS AND RESEARCH QUESTIONS………9

3

DATA AND METHODOLOGY………………………………………….....55

4

RESULTS ……………………………………………………………………80

5

SUPPLEMENTAL ANALYSES AND RESULTS…………….…………..118

6

DISCUSSION AND CONCLUSIONS……………………….…………….134

REFERENCES……………………………………………………….…………......149 APPENDICES……………………………………………………….……………...167

iii

LIST OF FIGURES FIGURE

Page

2.1

Conceptual Model: Self-esteem………………………………………………53

2.2

Conceptual Model: Mastery………………………………………………….54

5.1

Standardized Estimates for Stigma, Defensive Strategies and Self-concept.……………………………………………………………..….129

iv

LIST OF TABLES TABLE

Page

3.1

Sample Characteristics: Means, Proportions, Standard Deviation and Alpha Reliability………………………………………………….……………...61

3.2

Correspondence of Theoretical Concepts, Research Concepts and Empirical Indicators………………………………………………….…………..62

3.3

Principal Component Analysis Factor Loadings for Empowerment Scale (28 items).………………………………………………………………….68

3.4

Rogers et al. (1997) Factor Structure of the Empowerment Scale (28 items).………………………………………………………………………..69

3.5

Principal Component Analysis Factor Loadings for Empowerment Scale, Not Including Self-Concept Items (13 items) …………………………....71

4.1

Descriptive Statistics for Control, Independent, Dependent and Intervening Measures: Means, Proportions, Standard Deviation and Alpha Reliability…..…84

4.2

Correlation Coefficients of Control Variables with Stigma, Self-Esteem and Mastery…………………………………………………………………………..86

4.3

Correlations of Dependent and Focal Measures…………………………….…...89

4.4

Exploratory OLS Regression Analysis of Self-Esteem on Stigma and Stigma Responses…………………………………………………………………...…....92

4.5

Exploratory OLS Regression Analysis of Mastery on Stigma and Stigma Responses…………………………………...……………………………..……..97

4.6

OLS Regression of Self-Esteem on Stigma and Stigma Responses……….…...105

4.7

OLS Regression of Self-Esteem on Interaction Effects of Stigma and Stigma Responses………………………………………………………………………108

4.8

OLS Regression of Mastery on Stigma and Stigma Responses............………..111

4.9

OLS Regression of Mastery on Interaction Effects of Stigma and Stigma Responses………………………….……………………………………….…..114 v

4.10

Summary of Research Questions and Evidence of Empirical Support................116

5.1

Post-hoc Regression Analyses of Self-Esteem on Multiple Stigma Response Measures …………………….…………………………………….…………...121

5.2

Post-hoc Regression Analyses of Mastery on Multiple Stigma Response Measures …………………….………………………….………...…123

5.3

Post-hoc Regression Analyses of Stigma Response Measures on Control Measures and Stigma…………………………….………………..……………125

5.4

Unstandardized and Standardized Estimate for Stigma, Defensive Strategies And Self-concept (N=221)……………………………………………….……..130

A.1

Post-hoc Analyses of Stigma and Self-Esteem without Time 1 Measure of Self-Esteem ……………………………………………….…….…175

A.2

Post-hoc Analyses of Stigma and Mastery without Time 1 Measure of Mastery ……………………………………………………....……177

A.3

OLS Regression of Self-Esteem on Criminal Justice Involvement, Stigma, and Stigma Responses………...……………………………………….179

A.4

OLS Regression of Mastery on Criminal Justice Involvement, Stigma, and Stigma Responses………...……………………………………….182

A.5

Bivariate Correlation Matrix of Dependent, Independent and Control Measures…...………………………………………...……………..…………..187

vi

ACKNOWLEDGEMENT I am grateful to many individuals for their assistance and support throughout my graduate career. I would like to thank Dr. Christian Ritter, my mentor and dissertation director. I am deeply grateful for the time, encouragement and guidance he has provided throughout the years. His support has helped me become the sociologist I am today. I would also like to acknowledge the time and thoughtful comments provided by the members of my dissertation committee: Dr. Richard Adams, Dr. Emily Asencio, Dr. Kristen Marcussen, Dr. Kristin Mickelson and Dr. Sara Newman. I am thankful to Dr. Richard Serpe for his guidance over the years, and to Dr. Rebecca Erickson for her advice, insight and support in graduate school. I would also like to thank Dr. Mark R. Munetz for his unwavering support and encouragement. I am also appreciative of his help in shaping my research ideas and interests. I will be forever grateful to my parents and family for their unfaltering love and support. I would like to thank my friends for understanding when I needed to focus on my work, but also for distracting me when I needed that too. Finally, I would like to thank Shawn Bonfine for believing in me, helping me through this, and for making me laugh. I cannot express how happy you make me. I am grateful to share this journey with you.

vii

CHAPTER 1

INTRODUCTION

The stigma of mental illness is “fundamentally a social phenomenon rooted in social relationships and shaped by the culture and structure of society” (Pescosolido and Martin 2007: 322). Stigma is a “mark” or attribute that distinguishes one person from another and links the marked person with socially devalued or undesirable characteristics (Goffman 1963; Link and Phelan 1995). Ultimately, such categorization will result in the stigmatized individual being “reduced from a whole and usual person to a tainted, discounted one” (Goffman 1963: 3). If one is exhibiting behaviors or expressing emotions that are contrary to the norms for social interaction, the individual will be viewed as abnormal or deviant. Stigma associated with mental illness can have far reaching consequences for the lives of individuals who are labeled. Stigma occurs when such “elements of labeling, stereotyping, separation, status loss and discrimination co-occur in a power situation that allows the components of stigma to unfold” (Link and Phelan 2001: 367). Stigma and discrimination related to mental illness may block certain life goals for individuals, such as living independently or obtaining stable and meaningful employment (Corrigan, Kerr and Knudsen 2005; Link 1982; Link and Phelan 2001; Rosenfield 1997). Further, public attitudes and stereotypes about mental illness may cause people to avoid getting treatment

1

2

that is needed because they do not want to be labeled or experience discrimination (Corrigan and Wassel 2008; Corrigan et al. 2005). Sociological theories assume that stigma results in the self-devaluation of individuals who are stigmatized. However, the evidence supporting this presupposition is mixed (Crocker and Major 1989). Stigma resistance is an emerging concept that may explain the conditions under which certain individuals are able to resist devaluation and discrimination by others (Thoits 2011). While stigma resistance has not received much theoretical or empirical attention to date, it has the potential to explain how the impact of stigma varies among those who are labeled. In this dissertation, I explore the link between stigma and self-concept. I also examine stigma resistance responses that people may enact or possess to ward off the potentially negative effects of the stigma of mental illness. Specifically, I examine the use of defensive strategies as well as empowerment as potential moderators of the relationship between perceived stigma and self-concept. These relationships are assessed longitudinally among a population of individuals with severe mental illness.

Evidence of the Impact of Stigma on Self-concept Theoretical and empirical evidence suggests that there is a harmful link between stigma and self-concept for individuals with mental illness (Corrigan, Watson and Barr 2006; Link 1987; Link, Mirotznik and Cullen 1991; Rosenfield 1997; and Wright, Gronfein and Owens 2000). Stigma may negatively impact self-concept by eroding selfesteem and/or self-efficacy (Corrigan et al. 2005; Link et al. 2001). Stigma is linked to

3

psychosocial outcomes, including reduced agency and quality of life, and increased levels of depression (Camp, Finlay and Lyons 2002; Livingston and Boyd 2010; Lundberg et al. 2009; Rosenfield 1992; Rosenfield 1997; Wright et al. 2000). Sociological theories provide insight into the process by which stigma may negatively impact self-concept. According to classic and modified labeling theories, individuals distinguish and label human differences. The labeled person is linked to undesirable stereotypes, attitudes or beliefs. From a symbolic interactionist perspective, individuals recognize the values, attitudes and beliefs of others in society. If the undesirable stereotypes and attitudes of others about mental illness are internalized, meaning the individual applies stereotypes to one‟s self, negative self-feelings will occur. There are two relevant assumptions that explain how stigma may negatively impact self-concept (Camp et al. 2002; Link et al. 1989). First, it is assumed that people with mental illness recognize and share the negative representations that others associate with mental illness. Second, it is assumed that being diagnosed with a mental illness makes that condition a central component of one‟s identity or self-concept. These assumptions indicate that individuals are aware of negative cultural values and representations of mental illness, and that those who identify with these devalued cultural depictions of mental illness will have more negative self-feelings because they are representative of one‟s identity. However, despite this theoretical and empirical evidence, other evidence suggests that the harmful effects of stigma on self-concept are not as strong or as lasting as might be predicted (Camp et al. 2002; Corrigan et al. 2005; Crocker and Major 1989; Gove

4

2004; Thoits 2011). Compared to stigma associated with other social characteristics (e.g. sexual orientation) or physical health conditions or disabilities, stigma associated with mental disorders has been shown to have the weakest association to mental health outcomes including self-esteem, quality of life and depression (Mak et al. 2007). That is, stigma associated with mental disorders may not have as strong of an impact on internal self-feelings as expected by labeling theories when compared to the stigma associated with other characteristics or conditions. This finding suggests that there is variability in the impact of stigma, as well as in the ways in which individuals respond to stigma, that results in differential impact of stigma on self-concept. Further, Corrigan and Watson (2002) discussed a paradox of stigma that suggests that there is much variation in the impact of stigma. The authors found that, while many individuals with mental illness suffer self-derogation as a consequence of internalizing stigma and discrimination, there is a subset of individuals who react energetically with anger or become empowered to advocate on behalf of themselves and others, while still others are indifferent to the effects of stigma with no apparent impact on self-concept (Corrigan and Watson 2002). The paradox, then, is that given these different responses, stigmatized individuals in general do not differ from non-stigmatized persons in selfesteem and they may even have higher levels of self-esteem (Corrigan and Watson 2002). This paradox suggests that there is a need to examine the degree that stigma experiences influence social psychological processes and the wide variation in responses to stigma.

5

Stigma Resistance Responses There are a number of responses that individuals with mental illness may use to reduce the impact of stigma. Some may react to stigma by internalizing the negative reactions of others, applying these negative attitudes towards one‟s self. Internalizing stigma in this way ultimately results in a devalued sense of self-concept and adopting an identity associated with mental illness (Corrigan et al. 2005; Corrigan, et al. 2006; Goffman 1963; Livingston and Boyd 2010). Others may react by engaging in defensive strategies, meaning those attempts to protect oneself from the negative attitudes and discrimination of others. Such strategies include hiding the fact that one has a mental illness, at least to the degree possible, or withdrawing from social interaction (Link et al. 2002; Link et al. 1989; Link et al. 1991). Such strategies are defensive because the individual is actively trying to avoid or limit exposure to negative attitudes, stereotypes, and discrimination that have been perceived or are anticipated. Still others may react to the threat of stigma by engaging in proactive resistance strategies. Stigma resistance strategies are “intentional, agentic responses to possible harm” that challenge or confront the negative attitudes of others (Thoits 2011: p. 11). Such strategies may be behavioral, such as rebuking someone for a tasteless joke or seeking to educate others about mental illness, or more cognitive, such as in the deflection of stigma by reducing threats to self-regard (e.g. saying to one‟s self, “that‟s not me”) (Thoits 2011). Stigma resistance strategies may be social as well, bringing people together in a collective effort to challenge negative attitudes about mental illness. Collective action increases social resources, promotes advocacy and activism and

6

legitimizes feelings of anger and injustice (Thoits 2011). Individual and social empowerment are critical components of proactive stigma resistance strategies because they represent an attempt of individuals or groups to gain power and control, where individuals may benefit from membership in a devalued social group. It is the process by which individuals may improve self-concept by actively confronting the sources of stigma. Resistance strategies may be the key to more fully understanding how stigma impacts the self-concept of individuals with mental illness.

Research Aims Given the contradictory evidence around a persistent negative link between stigma and self-concept, further investigation of these relationships is essential to understand how the stigma of mental illness may have a detrimental effect for some individuals with certain conditions, but not for everyone in the same way to the same degree. Sociological theory and research must continue to examine both positive and negative effects of labeling mental health conditions. I propose three general research aims to examine the relationship between stigma and self-concept among individuals with mental illness, and how individual responses to stigma impact this possible relationship. The first research aim is to examine the relationship between stigma and selfconcept, as measured by both self-esteem and mastery. Most research has typically focused on self-esteem, a measure of one‟s attitudes about oneself. This study will also focus on mastery as a measure of self-concept because it incorporates attitudes and

7

feelings of personal power and control over one‟s environment. Mastery may be of particular importance with regards to the second research aim, which is to better understand how individuals enact strategies that may ward off or resist the potentially negative effects of stigma on self-concept. Specifically, I will examine the use of defensive strategies (e.g. secrecy and social withdrawal) as a moderator in the theorized relationship between stigma and self-concept. Studies that have examined the use of defensive strategies report mixed results in their ability to reduce the negative effects of stigma on self-concept. As such, recent theoretical and empirical work has shifted towards a focus on other forms of coping with stigma that may enhance one‟s sense of power. Empowerment has been described as antithetical to stigma and represents a process by which individuals seek to build strength from identifying as a member of a stigmatized group. The third research aim is to critically examine empowerment and its underlying constructs to assess their applicability as measures of proactive responses to stigma and as protective factors for the self-concept of individuals with mental illness.

Summary of Dissertation by Chapter This dissertation is divided into six chapters. This present chapter (Chapter 1) introduces the issues under investigation and briefly discusses the research aims for this study. In Chapter 2, I review relevant theoretical perspectives that contribute to our understanding of the relationship between public stigma and self-concept. Chapter 2 also presents a discussion of stigma resistance in general, and provides theoretical support for

8

the use of defensive strategies and empowerment as stigma resistance responses. Chapter 2 concludes with a statement of the research questions. Chapter 3 describes the sample and methodology used to examine the research questions. Chapter 4 presents the results of initial, exploratory analyses, as well as the results of the main analyses related to the research questions. Chapter 5 contains a summary of results from post hoc analyses that further explore findings from the primary analysis. Chapter 6 contains a discussion of the key findings of this study, along with a summary of limitations of the study. I conclude with a discussion of future research directions.

CHAPTER 2

THEORETICAL FOUNDATIONS AND RESEARCH QUESTIONS

In this chapter, I discuss the theoretical perspectives that explain the impact of stigma on self-concept. This review of relevant literature is divided into six sections. First, I draw from a symbolic interactionist perspective to define and contextualize the self-concept. I begin by providing a brief overview of the symbolic interactionist conceptualization of the self as it relates to the evaluative process that occurs within social interaction. I also describe two components of self-concept, self-esteem and mastery, and discuss how they may be impacted through a process of labeling that occurs within social interaction. Second, I provide a definition of public stigma as the negative attitudes of others towards individuals with a socially devalued social status. I contrast public, perceived stigma from self-stigma, meaning that stigma which has been internalized by individuals. Third, I discuss classic and modified labeling theories. In the fourth section of this chapter, I discuss principles from symbolic interactionist and labeling theories that relate to the relationship between stigma and self-concept. The focus of this dissertation is to further elucidate the mechanisms that enable or disallow public, perceived stigma to impact the self-concept of individuals with mental illness. In the fifth section of this chapter, I provide a brief overview of resistance strategies that individuals enact when confronted with stigma. This discussion includes a 9

10

comparison of the plight of individuals with mental illness to other similar groups who are also stigmatized, and how membership or identification with such stigmatized groups is not necessarily prescriptive of damaged or devalued self-concept, and may be beneficial. Sixth, I discuss the use of defensive strategies and empowerment as resistant responses to stigma. I conclude this chapter with a statement of the research questions.

Symbolic Interactionist Conceptualizations of the Self The symbolic interactionist perspective is an explanatory framework that provides insight into the definition and development of the self. The symbolic interactionist perspective views both society and the self as a fluid and negotiated process in which individuals define and interpret the meaning of actions of others through social interaction (Blumer 1969; Goffman 1959; Mead 1934; Thoits 1999). Individuals communicate through a system of shared symbols (e.g. language, gestures, and signals). Through communication, individuals are able to reflect on how others view the self (Mead 1934). Similar to the looking-glass self, we derive self-meanings by reflecting on others‟ reactions to our self (Cooley 1902). In this way, the self is developed through a reflexive process that provides feedback to the actor of the consequences of how others view and respond to the self (Burke 1980; Rosenberg 1979). Goffman (1959) focused on the ritualistic nature of interaction in which individuals are both the process and products of social interaction. In The Presentation of Self in Everyday Life, Goffman (1959) described a dramaturgical approach that likened social interaction to a staged drama in which and all individuals are simultaneously actors

11

and audience to others. The dramaturgical perspective suggests that individuals engage in impression management by carefully selecting and controlling appearance, manner and personal behaviors. If successful, the individual presents a portrayal of him or herself within the interaction that he or she wishes to be conveyed to others. Acting is an attempt of the individual to control the information that is available to others and to define the situation in a way that is preferred by each actor (Cahill 2001). It is important to note that, from a symbolic interactionist perspective, the individual is a conscious and reflective actor, not a passive and predetermined object of larger social forces (Blumer 1969). The central mechanism for human interaction is the interpretation of one‟s own and others‟ behaviors (Blumer 1969; Mead 1934). This interpretation occurs through communication, a process by which we observe and interpret others‟ signals (e.g. language, gestures, facial expressions, etc.), and then reciprocate with our own (Turner 1986). This interpretation involves giving meaning to other social objects, judging the suitability of that meaning as it relates to one‟s self, and making decisions based on that judgment that affect oneself and others because of its influence on the interaction. Through this process, individuals are active in the development and maintenance of the self. The self may be conceptualized as consisting of three features that are developed and maintained within social interaction. These features include: 1) the self, that is, the social object that engages in interactive processes; 2) self-concept, meaning the individuals‟ understanding and interpretation of self; and 3) self-esteem, or one‟s

12

evaluation of self. I briefly describe each of these three features of the self in the following sections.

The Self Symbolic interactionism suggests that by taking the perspective of others, that is, viewing the self from the standpoint and values of others, we develop a sense of self based on reflected appraisals (Mead 1934). In this way, the self may be defined as a social object in an environment that is affected and impacted by the individual, as well as by other social agents (Mead 1934). This role-taking process allows us to see ourselves as we perceive others to see us. Through such reflection, “the individual experiences himself, as such, not directly, but only indirectly, from the particular standpoints of other individual members of the same social group as a whole to which he belongs” (Mead 1934: 138). Role-taking occurs during social interaction, as well as during imagined interaction and internal, cognitive processing within the mind (Mead 1934). The mind, defined as symbolic interaction toward the self, allows us to envision how others view ourselves, and it is through the mind that one reflects on and understands the meaning of other people‟s words and actions (Mead 1934; Rosenberg 1979). The self is a reflexive dialectic between Mead‟s (1934) concepts of the “I” and the “me.” The “I” represents the active agent that experiences, thinks and acts. The “I” incorporates the aspect of individuals that is unlearned, unplanned and innate. The “me” represents the perspective of oneself that one assumes when taking to role of specific

13

others or of the general community. It represents the conceptions of self that individuals adopt when describing the self (Gecas 1982; Mead 1934). This reflexive exchange between the “I” and the “me” is a process that produces the self, suggesting that the self is not an object that is shaped entirely at the hands of others, but that individuals are capable of choosing actions with consideration for social context. Both the “I” and the “me” must be present to reflect on one‟s self as an object, and also to engage in actions that cause activity. In this way, Mead viewed the self as a process requiring action, then reflection, then action. The self incorporates the perspectives of all those who come in contact with us. Mead (1934) defined this general perspective as follows: “the organized community or social group which gives to the individual his unity of self may be called „the generalized other.‟ The attitude of the generalized other is the attitude of the whole community” (p. 154). In taking the role of the generalized other, we incorporate the attitudes, dispositions, meanings, expectations and collective representations for situations within an orienting perspective for our own experiences (Turner 1986). That is, as we interact with others beyond the circle of significant others to whom we are closest, we incorporate the viewpoints of general others into our own conceptualization of our selves. Within Goffman‟s dramaturgical approach, the self is developed anew within each interaction (Goffman 1959). For Goffman, “self is not something that an individual owns but something others temporarily lend him or her” (Cahill 2001: 192). Thus, individuals attempt to control the impression that others have in order to present a self that is consistent with the individuals‟ self views. In doing so, individuals strive for

14

authentic interaction when possible. Further, individuals learn to have emotions and feelings attached to the selves that are presented to others. Goffman (1959) called this “face,” and he discussed how individuals feel shame if the face is not adequately represented or is inauthentic. The emotional attachment to one‟s projected self regulates one‟s conduct within interaction (Cahill 2001). There are social implications for those who are unable to control all aspects of one‟s projected self. If an individual has a visible, stigmatized condition (e.g. physical deformity), he or she will be challenged in presenting the self that he or she desires to maintain within interaction. Thus, he or she may alter the way that he or she interacts with others by limiting the people that he or she comes in contact with or by restricting the information that is shared. The individual may also make attempts to conceal the stigmatizing condition. Goffman (1959) describes the potential consequences to social interaction as follows: “When the individual employs these strategies and tactics to protect his own projections, we may refer to them as „defensive practices‟; when a participant employs them to save the definition of the situation projected by another, we speak of „protective practices‟ or tact. Together, defensive and protective practices comprise the techniques employed to safeguard the impression fostered by an individual during his presence before others” (p. 13-14). In sum, the self may be defined as an object that develops through social interaction through communication among individuals who use a system of symbolic language and shared meanings (Blumer 1969; Cooley 1902; Goffman 1959; Mead 1934; Rosenberg 1979). Through interaction, participants make reference to each other and attribute characteristics to each other. During actual or imagined interactions, one gauges the reactions of others towards oneself. Thus, the self is a process of reflexive action.

15

Self-concept If the self is understood to be the aspect of a person that has experiences, reflects on those experiences and acts upon self-understandings, self-concept, then, may be considered as the product of this cognitive and reflexive activity as it results in a conception of oneself as a physical, social and moral being (Gecas 1982; Thoits 1999). Self-concept is the product of reflexive cognitive action rooted in one‟s personal experiences while interacting with the social world, and represents the totality of an individual‟s thoughts and feelings towards oneself. Self-concept consists of the essential parts of the self, meaning our own understanding of our self as an object (e.g. as a man, optimist) (Gecas 1982; Rosenberg 1979; Thoits 1999). The self-concept is “not the „real self‟ but, rather, the picture of the self” (Rosenberg 1979: 7, emphasis in original), meaning that self-concept is understood to us as we perceive others to view our selves. Self-concept is developed through a process of reflected appraisals and social comparisons (Gecas 1982). Reflected appraisals are those that are adopted by taking the role of specific or general others. Social comparisons are those comparisons that individuals make to assess one‟s own abilities and qualities in comparison to others (Gecas 1982). Because self-concept is developed through personal interpretations of others‟ perceptions, it is an inherently social phenomenon, even though it is cognitive and internal. As Gecas (1982) suggested, “the content and organization of the self-concept reflects the content and organization of society” (p. 10). To impose order onto one‟s self-concept, individuals develop social identities based on social classification and self-definition (Rosenberg 1979; Thoits 1999). Social

16

identities are often rooted in the structural locations of the individual, such as one‟s race or gender, on are based on roles that individuals enact. Roles are positions that carry normative behavioral expectations and scripts for carrying out normative behaviors, and they connect individuals to cultural values and norms (Gordon 1976). Roles provide individuals with meaning which helps them to make sense of themselves. Individuals develop self-concept around these roles, also called role-identities. Stryker (1979) suggested that self-concept is a salience hierarchy of these roleidentities. Individuals adopt multiple role-identities and commit to each to varying degrees. The level of commitment to a specific role-identity depends on the degree that one is enmeshed within social relationships that depend on that identity (Stryker 1979). The level of salience depends on one‟s readiness to invoke various role-identities in given situations (Stryker and Serpe 1994). Thus, those role-identities that are more salient to an individual will have greater commitment and comprise the individual‟s self-concept to a greater degree. When others recognize and legitimize one‟s role-identity enactments, selfconcept is maintained (Thoits 1999).

Self-esteem: Evaluation of Self Self-esteem is the individual‟s overall evaluation or appraisal of the self (Cast and Burke 2002: Gecas 1982; Thoits 1999). Individuals engage in a process of selfverification in which they assess the degree to which feedback from social situations matches or confirms one‟s self-concept (Cast and Burke 2002). Individuals evaluate themselves based on their ability to achieve a match between an identity goal or ideal

17

(known as an identity standard) and perceptions drawn from the environment about the actual performance of the self (Burke 1991; Cast and Burke 2002; Swann, Stein-Seroussi and Giesler 1992). Thus, self-verification processes help individuals reconcile their perceived self-concept with the feedback that is received from others. Swann et al. (1992) contend that individuals seek a stable self-concept in order to maintain a predictable and negotiable social reality. Verifying an identity produces feelings of competency and self-worth, increasing one‟s sense of self-esteem. However, when the self-verification process is inconsistent with an individual‟s view of him or herself, individuals make attempts to reduce the disturbance in the self-verification process (Cast and Burke 2002). Individuals extricate themselves from social situations or shed the identity in order to avoid the negative feelings that arise from persistent discrepancies between situational meanings and identity standards (Cast and Burke 2002: 1048). There are multiple conceptualizations of self-esteem. First, self-esteem is conceptualized using both a global and a domain-specific approach. General or global self-esteem incorporates both positive and negative attitudes toward the self (e.g. Rosenberg‟s (1965) self-esteem scale). Global measures of self-esteem, such as Rosenberg‟s (1965) self-esteem scale, refer to overall positive and negative feelings about the self, and are explicitly designed not to be domain specific or conceptually unidemensional (Robins, Hendid and Trzesniewski 2001). Others have focused attention on the bideminsional nature of self-esteem which separates a positive dimension that examines self-affirming, self-confidence measures from a negative dimension which

18

focuses on self-deprecation or self-denigration (Owens 1993). The global self-esteem scale as constructed by Rosenberg (1965) has high construct validity and does not appear to be superior to versions of the scale that focus on positive or negative dimensions (Greenberger et al. 2003). Second, self-esteem is conceptualized as representing two different aspects of self-evaluations: self-esteem that is based on one‟s sense of worth or value (worth-based self-esteem) and self-esteem that is based on one‟s sense of power, efficacy and competency to exert agency, control or causality within one‟s environment (efficacybased self-esteem) (Gecas 1982; Cast and Burke 2002). This evaluation of efficacy may also be global (e.g. Pearlin and Schooler‟s (1978) mastery scale) or it may be specific to one‟s efficacy in a certain domain (e.g. the health-related locus of control scale as a measure of one‟s ability to control health behaviors, such as diet and exercise (Wallston et al. 1976)). I discuss both worth-based self-esteem and efficacy-based self-esteem in greater detail below. Self-esteem. Self-esteem is viewed as the degree to which an individual feels that he or she is a person of value. Rosenberg (1965) described self-esteem as the evaluation that individuals maintain towards him or herself, and which typically expresses approval or disapproval towards oneself. It is a concept grounded in social norms and values concerning interpersonal conduct (Gecas 1982). An individual receives feedback from others about the degree to which he or she is accepted and valued through reflected appraisals and social comparison. The more that one feels valued based on the feedback of this process, the more that worth-based self-esteem will be improved. A person with

19

high self-esteem will believe that he or she possesses many positive qualities and will have positive attitudes toward the self. However, worth-based self-esteem will be most at risk when an individual is faced with possible exclusion or devaluation from others (Cast and Burke 2002). Mastery. Efficacy-based self-esteem is the degree to which people see themselves as capable and efficacious (Cast and Burke 2002; Gecas 1982; 1989). It refers to the motivation of individuals to engage in certain actions, and is relevant to their expectation of the efficacy of such actions (Bandura 1977; Gecas 1989; Gecas and Schwalbe 1983). As individuals interact with others, they evaluate their effectiveness, abilities and causal agency (Gecas 1989). This process is realized through self-verification within interaction that provides feedback to individuals as to the efficacy of his or her actions. This component of self-esteem is more responsive to perceived effectiveness on the part of the individual in maintaining consistency between situational meanings (reflected appraisals from others) and identity standards (Cast and Burke 2002; Gecas and Schwalbe 1983). Efficacy is akin to mastery, which is the component of self-concept that reflects one‟s sense of power or personal control (Gecas 1989; Pearlin and Schooler 1978; Wright et al. 2000). This definition is compatible with a symbolic interactionist view of self as “active, creative and agentive” (Gecas 1982: 18). Gecas and Schwalbe (1983) discuss how individuals derive a sense of self “from the consequences and products of behavior that are attributed to the self as an agent in the environment” (emphasis added) (p. 79). Like self-esteem, mastery is an evaluation of one‟s self-concept. It differs from selfesteem because it focuses on one‟s abilities and competencies to enact various desired

20

effects on social outcomes. A person with a high sense of mastery would feel that he or she has power over things that happen to him or her. Thus, conceptualizing mastery as an evaluative measure of self-concept incorporates dimensions of agency and power, both of which may be integral to one‟s ability to overcome and resist negative effects of stigma and perceived discrimination.

Self-esteem and Mastery as Outcome Measures When examining self-concept as an outcome, researchers examine the processes that produce or inhibit self-esteem and mastery (Rosenberg 1979). While this is the focus of some empirical studies (see Bengtsson-Tops 2004), much work that focuses on a stress process framework views self-concept as an intervening concept in the process by which social stressors affect physical stress outcomes. Based on the stress process framework, self-esteem and mastery are viewed as personal coping resources that individuals have to reduce the harmful effects of stressors (Mirowsky and Ross 2003; Schieman 2002; Pearlin and Schooler 1978; Thoits 1995). Self-concept is thus conceptualized as a mediator in the process by which stressors impact health and psychological well-being outcomes, or as a moderator which exacerbates or inhibits the effects of stress on outcomes (Pearlin et al. 1981; Thoits 1999). The limitation with this approach as it relates to this study is that it does not examine how “the success or failure of coping strategies might enhance or undermined self-concept” (Thoits 1999: 358). To better understand self-concept‟s role in the stress process and the means by which it serves as a coping strategy, we need to examine those

21

factors that influence self-concept. While this dissertation is not a sociological assessment of the stress process itself, the findings may influence our understanding of the ways in which social stressors (e.g. stigma) impact self-esteem and mastery. In sum, the self develops through interaction, while the self-concept is the conception of oneself as a social being. The self-concept is ordered by social roles that provide meaning about who an individual is and what behaviors and actions are expected in social situations. Self-esteem is the evaluation or appraisal of the self-concept. Individuals engage in a process of self-verification to match self-feelings with the feedback received from others. Successful self-verification will result in positive selfevaluations, while negative self-verification will result in the individual making attempts to reduce discomfort that disconfirmation causes. There are two aspects of self-esteem: worth-based and efficacy-based self-esteem. Both aspects may be impacted by stigma, but the impact may occur in different ways. Worth-based self-esteem, meaning the feelings that one is a person of value, may be negatively affected by anticipated or experienced devaluation. Efficacy-based self-esteem is the evaluation of oneself as efficacious, agentive and in control. Efficacy-based selfesteem may relate to various strategies that individuals enact when confronted with stigma. In the following sections, I define stigma, discuss the social origins of stigma, and describe public, social stigma in comparison to internalized, self-stigma. I will then review classic and modified labeling theories which link the discussion on self-concept with the literature on stigma.

22

Defining Stigma Stigma is conceptualized as a mark or discrediting characteristic of an individual that is devalued by others with whom the individual interacts (Goffman 1963). In defining stigma, Goffman (1963) wrote, “The term, stigma, then, will be used to refer to an attribute that is deeply discrediting, but it should be seen that a language of relationships, not attributes, is really needed. An attribute that stigmatizes one type of possessor can confirm the usualness of another, and therefore is neither creditable nor discreditable as a thing in itself… stigma is a special kind of relationship between attribute and stereotype” (p.3-4). Thus, stigma exists within social interaction, and represents the separation of one individual (or group of individuals) from others based on cultural meanings of a devalued, discredited characteristic. Because stigma occurs through social interaction, the formation of stigma is inherently a social process. A stigmatizer perceives a mark or characteristics that is socially devalued or perceived as negative. Certain marks or cues may be more visible than others, such as those cues marked by physical deformity, while others may be more invisible or hidden, including the stigma of mental illness (Corrigan and O'Shaughnessy 2007; Goffman 1963). Once a stigmatizer perceives a mark, he or she endorses negative stereotypes that are learned and culturally sustained through a process of socialization (Balsam and Mohr 2007; Link et al. 1987; Stuber, Meyer and Link 2008; Thoits 2011). These stereotypes are knowledge structures that reflect the cultural beliefs and categories of a dominant social group (Corrigan and O'Shaughnessy 2007).

23

Often, stereotypes are formed around neutral observations about a certain group of individuals. However, they may also be formed around negative feelings or perceptions about individuals. If certain individuals behave in ways that are contrary to social norms for appropriate behavior, individuals may draw upon certain stereotypical beliefs to explain and understand the situation. If these stereotypes are enacted and endorsed by a stigmatizer, he or she may develop prejudicial attitudes towards the individual or may change behaviors to discriminate against the individual (Corrigan and O'Shaughnessy 2007; Stuber et al. 2008). Individuals rely on social cues to make these determinations within social interaction. However, for certain conditions, a label is needed for hidden stigmatizing conditions to become known. In this way, labeling of a condition by a person with authority or power to make such a determination plays an important role in the degree to which one will be stigmatized by others. Stigma operates at the structural, interpersonal and individual levels of society. At the structural level, stigma exists within institutional policies that allow for patterns of structural discrimination, exclusion and prejudice (Goldberg and Smith 2011; Link and Phelan 2001; Livingston and Boyd 2010). At an interpersonal level, social or public stigma is the negative attitudes of the general population held towards a group of individuals who are viewed as having a negative mark that links those individuals to undesirable characteristics (Link and Phelan 2001). Public stigma is the social endorsement of negatively held beliefs about stigmatizing characteristics or towards individuals who possess such characteristics. These negative beliefs are held by the generalized other and are endorsed through acts and portrayals of public discrimination

24

and prejudice (Livingston and Boyd 2010). Self-stigma occurs at the individual level and represents the loss of self-esteem or self-efficacy that occurs when people internalize public stigma (Corrigan et al. 2005). This dissertation will focus on a conceptualization of stigma as public or social stigma. Given the importance of social interaction in the creation of the self from a symbolic interactionist approach, I seek to better understand the degree to which negative feelings and attitudes of the generalized other impact the self-concept of people who are labeled with a stigmatizing condition. I do not focus on internalized stigma (or selfstigma). Internalized stigma relies too heavily on the assumption that public stigma, in fact, does have a negative impact on self-concept, and ignores the potential influence of moderators in the process by which stigma may affect self-concept. In sum, stigma may be defined as a discrediting mark or characteristic that is given to an individual who is expressing behavior or emotion that is counter to the culturally accepted expectations for normative behavior, thoughts or emotions. The individual will be viewed as deviant or abnormal, and will be labeled with a devalued social status. Mental illness is stigmatized because of public sentiments that individuals with mental illness are unpredictable, unkempt, irrational or dangerous (Link et al. 1999; Pescosolido et al. 1999, 2010; Phelan et al. 2000). If this public stigma negatively impacts the way that an individual views him or herself, then stigma becomes internalized and affects the self-concept. In the following section, I discuss classic and modified labeling theories that further explain this process.

25

Labeling Theories Labeling theories posit that individuals recognize and react to societal conceptions of mental illness (Link et al. 1989; Scheff 1963). Per Scheff (1963), labeling theory suggests that individuals who have been categorized by other people as deviant come to view themselves as deviant. Such categories are applied primarily by agents of social control, resulting in an official label that is attached to the individual. This label results in negative responses from others, which in turn causes the individual to adopt the expected social role of a person with mental illness. Thus an identity shift occurs for the labeled individual (Scheff 1963; Scheff 1974; Link et al. 1989). Labeling theory emphasizes the role of deviance in our conceptualization of mental illness. Scheff (1963) categorized mental illness as a form of residual deviance, meaning that the deviant behaviors do not violate a specific social rule, but represent a category of behaviors that appear wrong or culturally different. When the behavior gains sufficient public recognition to be viewed as problematic, or when a professional labels the behavior (e.g. by applying a diagnosis), the cultural definitions associated with mental illness take over and bring about the social consequences of being labeled, including social rejection and discrimination (Scheff 1963). Gove (1970) rejected this contention that the label is the reason for social rejection. Gove (1970) discussed that individuals are socially rejected because of deviant and norm-breaking behavior, not because they have been officially labeled. Social norms reflect the goals, priorities, and definition of

26

constructs and commonly held beliefs in society, and when these norms are violated by an individual exhibiting deviant behavior, social processes of rejection are enacted. Mental illness may also be viewed as deviant because it is counter to the dominant social norms of health. Illness in general is defined by deviance from a norm, which is based on social definitions and constraints, such as morality (Szasz 1960). The terminology previously used to describe mental illness has shifted from having a moral connotation, such as a manifestation of sin or demonic behavior, to a well/sick classification (Aneshensel and Phelan 1999; Ausubel 1961; Mechanic 1999). That is, the medical model now emphasizes a definition of mental illness as “sick” rather than “bad.” However, as Aneshensel and Phelan (1999) describe, the medical model emphasizes internal processes within individuals and diverts attention from social sources of stigma. While a label of sick may be preferable to a label of bad or evil, the medical model still shapes “sick” as a deviant status, which potentially results in stigma and negative social consequences.

Modified Labeling Theory Like labeling theory, modified labeling theory explains the process by which social labels, specifically the label of mental illness, impact self-concept. In discussing modified labeling theory, Link et al. (1989) proposed a 5-step process in which this occurs. First, societal conceptions of mental illness lead to general feelings of devaluation and discrimination towards mental illness, and through socialization, individuals learn these societal conceptions of mental illness (Link et al. 1987; Thoits 2011). Second, when

27

an individual is officially labeled with a mental illness the negative societal conceptions about mental illness become personally relevant to that individual (Link et al. 1989). Third, individuals respond to this label in a number of ways to mitigate the harmful effects of the negative label. Fourth, the negative consequences of others‟ reactions to the label of mental illness ultimately impact one‟s self-concept and social resources, and fifth, this process results in increased vulnerability to repeated or prolonged disorder and the continuation of the mental illness label (Link et al. 1989). Unlike classical labeling theory, modified labeling theory does not assume that the stigmatized individual accepts their categorization as self-descriptive (Thoits 2011). However, stigmatized individuals who do not accept the label are still exposed to the negative attitudes and perceptions of others, and the label may still negatively impact self-concept. Thus, the key element of labeling is that stigmatization and social rejection follow from a label that carries a negative connotation (Link et al. 1987). According to Link et al. (1987), “a label is a starting point that activates an array of beliefs about the designated person that may ultimately affect the level of acceptance or rejection such a person experiences” (p. 1474). Labeled individuals may take the perspective of others and define him or herself as others would, adjusting behavior to conform to the expectations of others. This self-fulfilling prophecy suggests that people with negative views alter behavior towards members of stigmatized groups so that a stigmatized individual behaves and ultimately views him or herself in a manner consistent with the negative stereotypes of the larger group (Crocker and Major 1989). Ultimately, this process may result in a shift in one‟s self-concept by affecting self-esteem or mastery.

28

Source of the Label An important consideration within classic and modified labeling theories is the source of the label. In describing labeling theory, Scheff (1963), suggested that a formal agent of social control is the primary source of labels. These include the doctors, psychiatrists or other professionals who assign psychiatric diagnoses (Phelan and Link 1999). However, others may also participate in the labeling process, such as lay persons who perceive behavior to be deviant, or even the individual who engages in self-labeling of one‟s own behavior (Gove 1970; Horwitz 1982; Thoits 1985). For instance, Thoits (1985) posited that “public, official labeling of one‟s rule breaking is not necessary for the emergence of a deviant identity; there can be private self-labeling” (p. 222). Thus, a formalized label of a mental disorder may not be necessary in order to experience the effects of being labeled. Drawing from Mead‟s (1934) concept of the generalized other, Thoits (1985) suggested that, even if others are not present, individuals may imagine the reactions of others and internalize these imagined responses. Self-labeling assumes then, that individuals recognize social values and norms for behavior, that there are cultural labels for the categories of norms for acceptable behavior that can be applied to oneself, and that there is motivation to conform to these norms (Thoits 1985). As individuals engage in self-reflection, they can imagine the reaction of others and may alter their self-conceptions accordingly. While self-labeling may occur, there is evidence that individuals are resistant to applying self-labels even when receiving mental health services. This unwillingness or uncertainty about self-labeling reflects a conception of mental illness as a dynamic

29

process, with symptoms that are transitory, variable or ambiguous (Moses 2009). Further, self-labeling is based on the assumption that individuals want to conform to social norms, which may not be the case for individuals who are seeking to clearly define, develop and understand one‟s self, such as adolescents (Moses 2009). Whether the resistance to engage in self-labeling is a reflection of the shifting nature of mental illness itself, or is a hesitancy to incorporate mental illness within one‟s self-concept, the process itself can be transformative for the individual.

Summary of Theoretical Principles Related to Stigma and Self-concept Thus far, I have presented the theoretical arguments that would suggest that stigma has a pervasive and negative influence on the self-concept of individuals with mental illness. While this may be true for certain individuals, there is evidence that stigma does not influence others in the same way or to the same degree. In this section, I summarize the key theoretical principles that may explain variation in the experience of stigma and its effects on self-concept. Specifically, I draw three themes from the above discussion of the impact of stigma on the development of self-concept. First, I discuss how processing the norms and values of the generalized other may result in the adoption of deviant identity; however, there is selectivity in the degree to which individuals adopt the perspective of the generalized other. Second, I discuss how self-concept focuses on certain role-identities that carry different weights or degrees of importance to the individual. I discuss how this may have an impact on the salience of a stigmatized identity. Third, I discuss the

30

importance and value placed on the agency of individuals in defining self-concept, and how this may relate to differential responses to stigma.

Rejecting the Views of the Generalized Other One of the key extensions of a symbolic interactionist approach is how the attitudes and perspectives of the generalized other plays a central role in the definition and labeling of a deviant status. By taking the perspective of others, or viewing one‟s self from the standpoint of the community and its social values, we see ourselves as others do (Blumer 1969; Mead 1934). According to Rosenberg (1979), “when the language of verbs becomes a language of nouns, either through formal certification procedures or general social recognition, the labeling process occurs, and produces additional elements of social identity” (p. 10-11). Thus, as individuals are classified and defined by others within social interaction, they take on labels given by others and incorporate them into self-concept or roleidentities. These labels may have positive or negative connotations, as in the case of a deviant or stigmatized label. Given this, it is assumed that possessing a stigmatized attribute that is not valued by the generalized other would have a negative effect on one‟s self-evaluation (Camp et al. 2002). However, evidence exists that individuals vary in the degree that they concur with others‟ views. Such disagreement may result in anger, defensive reactions or proactive responses to try to change the others‟ perspective (Camp et al. 2002; Fine and Asch 1988; Mak et al. 2007). As such, the negative impact of stigma on self-concept is not inevitable.

31

The Importance of a Stigmatized Identity to Self-concept The second theoretical principle that relates to the relationship between stigma and self-concept concerns the importance of the stigmatized condition to one‟s sense of self. The self-concept consists of the relationship of its parts, that is, the roles or identities, that individuals have. Those role-identities are organized hierarchically in a system of self-values, meaning the role-identities that people have are not of equal importance to individuals (Rosenberg 1979). Certain role-identities that comprise the self-concept have unequal centrality to the individual, and it is possible that identities related to a stigmatized condition are not incorporated within self-concept. There is a theoretical assumption that evaluations about one‟s role-identities are additive, combining to represent one‟s global sense of self-esteem (Thoits 1999). For instance, if one views oneself as witty, attractive, a hard worker and a good listener, all positive attributes, the assumption is that the positive self-evaluations of each of these specific aspects of self contribute to global self-esteem. While this may be true, this assumption does not take into account the importance of each component to one‟s selfconcept. We derive global self-esteem from qualities that matter to us as individuals, or those that are rated as central to our conceptualization of our self (Rosenberg 1979). In other words, “if we attempt to understand persons‟ global self-esteem, we must appreciate the differential contribution to global self-esteem of the esteem attached to self-components which themselves are regarded by persons as more or less central” (Stryker and Serpe 1994: 18). If this is so, then those aspects of self-concept (roleidentities) that are most highly valued will contribute most to global self-esteem. Those

32

aspects of self-concept that are deemed less important will not impact self-esteem as much. If an individual does not view the status of mental illness as an important or central component of their self-concept, that individual may not incorporate the stigmatized identity within his or her self-concept. Goffman (1961) purported that a patient of a mental hospital will convert his or her self-concept to match the institutional definition of self as a mental patient. Thus, patients who have been institutionalized for a long period of time will adopt a self-concept as one who is sick, dependent and in need of care. However, while there may be acceptance of the patient role and the presence of symptomatic behavior, this does not necessarily translate into acceptance of being a person with mental illness within one‟s self-evaluation (Townsend 1976). Individuals may protect or preserve self-concept by minimizing the emphasis on oneself as a person with mental illness. This would reduce the ability of the mental illness label (and all of its negative connotations) to harm self-concept.

Personal Agency in Defining the Self-concept A third theoretical principle that may illuminate how stigma has a differential impact on self-concept is the role of personal agency in defining self-concept. Acknowledgement of the attempts of individuals to reject or resist such labels reflects a theoretical trend that focuses on an assertive and agentive self to create and maintain positive self-concepts (Gecas 1982). According to Rosenberg (1979), individuals selectively choose which other social actors are significant others, which social

33

comparisons are made, and which aspects of one‟s self-concept are central or most salient to the individual. This selectivity implies an active choice on the part of the individual to create or maintain a self-concept that is desirable to that individual. The degree to which one is successful at this will impact how one evaluates self-concept, and may relate to feelings of self-worth or efficacy. Given the importance of agency in shaping one‟s self and self-evaluations, it is necessary to examine both self-esteem (worth-based self-esteem) and mastery (efficacybased self-esteem) as central components of self-concept (Rosenfield 1992; Wright et al. 2000). Mastery represents the extent that one feels in control or powerful over one‟s life circumstances, and it is an essential concept to the investigation of how individuals draw upon personal resources and paradigms to respond to stress, including stigma (Thoits 2011). Compared to studies that examine self-esteem, a focus on mastery incorporates one‟s feelings of control and ability to change one‟s situation. Mastery may be particularly important when examining the effects of stigma on self-concept, or in the assessment of approaches that individuals enact to reduce or limit negative effects of stigma. In sum, while the symbolic interactionist perspective and labeling theories suggest that negative cultural values related to mental illness will be viewed as self-descriptive, there is also theoretical support that the degree to which this process occurs varies. Individuals may be aware of negative stereotypes held by the generalized other, but they may also not define themselves by these cultural depictions. Further, individuals may cognitively reframe the degree of importance of the mental illness identity, thus reducing

34

the potential threat of stigma associated with a deviant category. Finally, individuals enact a degree of agency in selecting which aspects of self-concept are most central. In the following sections, I discuss how stigma resistance is a critical theoretical and empirical construct in fully understanding the effect of stigma.

Resisting the Negative Effects of Stigma As discussed, there is theoretical and empirical evidence to suggest that the stigma of mental illness may have a detrimental impact on self-concept (Corrigan et al. 2006; Livingston and Boyd 2010; Munoz et al. 2011). However, there is also evidence that suggests that some individuals are not as negatively affected by stigma as might be predicted or expected based on symbolic interactionist and labeling theory frameworks, and that the negative effects of stigma are not felt by all individuals with mental illness (Thoits 2011). In other words, there is wide variation in the degree to which stigma exerts a negative impact on an individual‟s sense of self. Personal agency and choice of responses to stigma have not been thoroughly examined, and much is unknown about how individuals enact strategies to counter the effect of stigma. Thoits (2011) observed that there are multiple, unexplored responses to stigma that may explain the modest effect of stigma on self-concept for certain individuals. Such responses may have either beneficial or detrimental consequences, but an individual‟s use of resistance strategies should not be ignored. As such, additional research is needed to more fully understand how individuals respond to perceived stigma.

35

In this section, I discuss two stigma resistance orientations that individuals may adopt in response to stigma: defensive strategies and empowerment.

Stigma Resistance: Defensive Strategies As discussed, people with mental illness are likely to be aware of negative cultural representations of mental illness. Once labeled, these negative representations become personally relevant (Link et al. 1989). In response, some individuals adopt defensive strategies to overcome the negativity associated with a stigmatized condition (Camp et al. 2002; Link et al. 1991; Markowitz 1998; Thoits 2011). Link et al. (1991) and Link and colleagues (1989) discussed withdrawal and secrecy as coping responses to stigma. Withdrawal involves limiting social interactions to those who know about and tend to accept one‟s stigmatized condition, and secrecy involves concealing one‟s condition from others (Link et al. 1991: 304). From a modified labeling perspective, secrecy and withdrawal are self-protective strategies that are designed to distance oneself from the stigmatizing attitudes of others. There is mixed evidence as to the effectiveness of secrecy and social withdrawal on reducing the harmful influence of stigma. For instance, Link and colleagues (1991) reported that social withdrawal had a detrimental effect on demoralization (an internal feeling state) and unemployment (an objective social condition), and that neither secrecy nor withdrawal diminished negative labeling effects on psychological distress or demoralization. Others found no evidence of the effectiveness of social withdrawal as a strategy that protects self-esteem (Ilic et al. 2011; Wright et al. 2000), while others have

36

found that secrecy has a strong, negative impact on self-esteem (Ilic et al. 2011). Link et al. (1991) suggest that the use of defensive coping strategies hurts individuals‟ selfconcept because it inadvertently limits social opportunities and induces psychological malaise among those who invoke such responses.

Stigma Resistance: An Empowerment Approach Further examining the interplay of personal agency and social empowerment is one theoretical direction that may explain the relationship between stigma and self. While some coping responses to the stress posed by stigma may be defensive reactions, other responses indicate agentive, proactive measures that individuals take to assuage the effects of stigma. Empowerment has been described as the antidote to or opposite of stigma (Corrigan et al. 2005; Lundberg et al. 2008), and may be an effective stigma resistance response. However, it is unknown how empowerment affects the potential relationship between stigma and self-concept. Empowerment is defined as the means to gain control over one‟s life conditions and critically understand and influence the structural constraints in which one lives (Perkins and Zimmerman 1995; Rogers et al. 1997). It is a multi-level construct that operates at the individual, organizational and social or community levels (Zimmerman 1995). According to Rappaport (1987), “empowerment conveys both a psychological sense of personal control or influence and a concern with actual social influence, political power, and legal rights” (p. 121). With regards to mental illness, the goals of

37

empowerment are to diminish stigma and discrimination, overcome illness and promote recovery from illness (Corrigan and Garman 1997). The empowerment ideology emerged from the self-help and community psychology movements which followed deinstitutionalization (Dickerson 1998; Segal, Silverman and Temkin 1995). This empowerment ideology developed, in part, as a response against the domination of the psychiatric professional establishment, and it encourages persons with mental illness to take control over their lives, reduce reliance on professionals and develop support networks with other similar individuals who also have mental illness (Dickerson 1998). An empowerment approach has been incorporated into treatment and service programs by increasing client decision-making in the treatment process, improving social skills and networking opportunities, encouraging supportive peer interactions, and framing service providers as partners or collaborators instead of as authoritative experts (Corrigan and Garman 1997; Perkins and Zimmerman 1995; Segal et al. 1995). Empowerment is a broad construct distinct from self-concept. In developing a theory of empowerment, Rappaport (1987) described it as “not only an individual psychological construct, it is also organizational, political, sociological, economic and spiritual” (p. 130). To Rappaport (1987), empowerment is inherent in racial and economic justice, legal rights, human needs, health care, education and community. It is as relevant to organizations, neighborhoods and communities as it is to individuals (Rappaport 1987: 130). Empowerment, however, has primarily received attention in the psychological literature because of its emphasis on internal, psychological processes (e.g.

38

sense of personal control, self-efficacy and self-esteem) and because of its popularity in a variety of mental health services and programs. This focus has downplayed some of the social components of empowerment, such as group advocacy, and ignores social context and forces that exert influence over individuals (Peterson and Zimmerman 2004; Riger 1993). This is problematic because it focuses attention on or within individuals and detracts focus from the collective actions of groups of individuals. Empowerment theory, research and interventions link individuals to the larger social environment (Perkins and Zimmerman 1995). While empowerment is at the forefront of many treatment programs, it has been understudied as a social phenomenon that influences and is influenced by social forces. Empowerment has been cited as an essential element to ward off or combat stigma, yet research on empowerment has not focused on how individuals who have successfully overcome stigma have achieved this outcome (Corrigan and Garman 1997; Shih 2004). Empowerment is similar to the process of coping as both are concerned with gaining control over one‟s environment to mitigate the influence of stress (Gutierrez 1994). However, the concept of empowerment goes beyond coping. The focus on coping is on how individuals themselves adjust to the effects of stress, either through the use of social support resources or by enlisting psychosocial resources to buffer the effects of stress. The empowerment perspective, however, is concerned with how individuals actively attempt to adjust or eliminate stressful situations (Gutierrez 1994). Feelings of control are significantly associated with better health outcomes, goal setting and problem solving, as well as reducing uncontrollable stressors (Gutierrez 1994; Thoits 2006).

39

Coping processes seek to avoid or prevent negative consequences, whereas the goal of an empowerment perspective is to enact processes that change the social world and create positive outcomes (Shih 2004). If coping is a reactive response to adversity that drains available coping resources, empowerment is an active attempt to overcome adversity by drawing strength from the experience of adversity (Shih 2004). The empowerment perspective looks at the relationship of individuals to their social environments, yet from the perspective of how the environment is impacted by the individual. From this perspective, “empowering interactions can be an important mediator of stressful life experiences by encouraging health and action oriented responses to the social environment” (Gutierrez 1994: 208). Illness and identity: Developing an empowered self. When facing a chronic or stigmatized illness, individuals may define new roles and identities for themselves based on the illness condition (Crossley 1997; Onken and Slaten 2000). The link between an illness condition and its impact on self-concept is based on an assumption that the debilitating condition is central to one‟s self-concept, leading to social comparisons between sick and well social groups (Fine and Asch 1988). Such social comparisons make illness synonymous with helplessness, dependency, and passivity. The connection between illness and identity extends Parson‟s (1951) concept of the sick role. In brief, the sick role is a set of “institutionalized expectations” around the social condition of illness (Parsons 1951: 436). There are four exemptions and obligations of the sick individual. First, the individual is exempt from normal social role responsibilities. Second, the sick individual is exempt from responsibility for the illness,

40

meaning that the individual is expected to recognize that his or her condition “must be taken care of” (Parsons 1951: 437). Third, the individual must express a desire to get well, as the illness itself must be viewed as undesirable. Fourth, the sick individual is obligated to seek professional help to correct the condition. Implicit in the sick role model is that the illness is an undesirable state that limits the abilities of the afflicted individual, and that professional or outside help must be sought to correct the situation. Individuals who are ill are then removed from social interaction and are viewed as separate and distinct from healthy, normal others. This distinction may have some positive effects, particularly if the adoption of the sick role increases empathy and social support from close network ties (Perry 2011). However, there are likely simultaneous negative effects of being categorized as ill. In terms of the effects of stigma of mental illness on self-concept, it is possible that adoption of the sick role for individuals with chronic health conditions may negatively influence self-concept as individuals who are stigmatized may internalize negative reactions of a healthy generalized other. Conversely, empowerment in the development of an illness identity challenges the traditional attitudes about illness and its impact on self-concept. An empowerment perspective suggests that individuals seek to build strength from identifying as a person with a certain illness or condition. This encourages the individual to accept labels but refute the negative stigma associated with such labeling. Crossley (1997) discussed the development of an “empowered self” as a possible response of individuals who reject the

41

traditional model of the sick role, particularly if the illness condition is stigmatized or misunderstood by others (p. 1870). In developing an empowered self, individuals enact strategies that build strength and derive power from the otherwise discredited identity as an ill person. Individuals become empowered to define their conditions in ways that overcome stigma of a devalued status. For example, constructing oneself as a “survivor” or “fighter” has positive, moralistic connotations suggesting agency, hard work and determination of this achieved status (Crossley 1997: 1866). The empowered self may improve the selfconcept of the individual and limit the negative influence of social devaluation. Given the above discussion, social empowerment may be advantageous for individuals with mental illness as it suggests the exercise of personal agency in developing an identity around an illness condition. This identity promotes in-group social connectivity, which leads to better self-esteem, self-efficacy, resiliency in the face of adversity, adaptability and optimism (Balsam and Mohr 2007; Shih 2004; Thoits 2011). The process of identifying with a stigmatized group may eliminate stress-inducing tension around hiding one‟s condition from others, reduce the shame associated with the disabling condition, and it may reduce negative attitudes of others (Corrigan 2003; Lee and Craft 2002; Onken and Slaten 2000). As Link and Phelan (1995) state, “to the extent that people with mental illness are empowered to „come out‟ and challenge stereotypes, the broader public may come to view mental illness differently than they do now” (p. 375). This strategy has been effective for other marginalized groups that have sought to overcome diversity and establish equality, such as those seeking racial and gender

42

equality, as well as within the gay, lesbian, bisexual and transgendered movement (Balsam and Mohr 2007; Corrigan and Garman 1997; Gershon, Tschann and Jemerin 1999; Quinn and Earnshaw 2011). Strategies that shift power from the dominant group to the stigmatized group may reduce the influence of stigma. Thus, stigma may be negated through proactive approaches to change or reduce stigma, such as community activism, protest, education and interpersonal contact (Corrigan et al. 2005; Link and Phelan 1995). Individuals who self-identify as a person with mental illness may also turn to similar others and peers to gain understanding, acceptance and support related to the condition. As individuals make connections with one another to reduce the effects of stigma, they may also begin to develop an identity as a collectivity (Thoits 2011). By creating a social collective, individuals with mental illness may become empowered to challenge negative stereotypes and discrimination and advocate for social justice. This process could have broad social consequences and may also improve the self-concept of individuals involved. In sum, empowerment is a concept that has been under-studied by sociologists. Empowerment is distinct from one‟s sense of self-concept, and goes beyond internal coping processes. Individuals develop a sense of resiliency by overcoming adversity and creating positive outcomes (Shih 2004). Examining empowerment as a social activist movement will inform social research on the influence and power of the social collectivity of disadvantaged groups. Further research will also improve our understanding on how empowerment operates across different social characteristics, in different social contexts, and across the life span (Zimmerman 1995). Finally, from a

43

social psychological perspective, we may better understand empowerment as it contributes to the development of self-concept and identity. It is important to understand empowerment as a social process by which individuals draw strength and power. Thus, incorporating measures of empowerment within sociological analyses of stigma resistance strategies may help researchers explain why some individuals with mental illness do not have negative or adverse reactions to social stigma. In this study, I posit that an empowerment orientation may mitigate the impact of stigma on self-concept. However, I must first discuss some measurement issues related to empowerment that must be addressed in order to understand its impact on the relationship between stigma and self. Measuring empowerment. Empowerment is typically conceptualized as a psychological construct that incorporates self-oriented measures to examine internal, social psychological orientations (Peterson and Zimmerman 2004; Riger 1993; Segal et al. 1995; Thoits 2011). Empowerment is closely related to self-concept as those who have a strong sense of empowerment can be expected to have high levels of self-esteem. Also, empowerment is by definition a construct that integrates perceptions of personal control (mastery) with one‟s behaviors to exert control because it is related to “control over one‟s treatment and one‟s life” (Corrigan et al. 2005). While empowerment is often viewed as an individual construct, it is distinct from one‟s sense of self or personality (Segal et al. 1995). In developing a valid measure of empowerment, it is necessary to examine the correlation of measures of self-concept (e.g. self-esteem, self-efficacy and mastery) and empowerment (Segal et al. 1995). However,

44

common operationalizations of empowerment incorporate measures of self-concept as underlying components within the construction of empowerment measures. This blurs our understanding of how empowerment may be manifested within the everyday lives of individuals with mental illness. Incorporating self-concept within the construction of empowerment measures is problematic in two ways. First, such an emphasis may not fit with the realities of many people with mental illness. Low personal regard, negative self-feelings and lack of insight into an illness condition may be definitional characteristics of certain mental illness conditions (Dickerson 1998; Shih 2004). Thus, measures of empowerment may be inaccurate or inadequately defined if they, by definition, rely on the presence of a certain level of self-esteem, mastery, or awareness of the illness condition. Second, and most relevant to the present analysis, incorporating self-concept measures within the concept of empowerment masks important differences between self-oriented and socially-oriented aspects of empowerment and confounds any assessment of the relationship among stigma, empowerment and self-concept. This measurement issue is apparent when examining the construction of a commonly used measure of empowerment- the Boston University Empowerment Scale (Rogers et al. 1997). This measure was developed with input from a group of individuals with mental illness who were engaged in mental health services and who were active in the self-help movement. Rogers et al. (1997) held meetings to get input on the psychological components of empowerment (e.g. assertiveness, having decision-making power, having access to information and resources). These groups developed consensus

45

of certain dimensions of empowerment, including control over one‟s life, achievement of goals, self-esteem and self-efficacy (Rogers et al. 1997). Rogers et al. (1997) developed survey items to represent these dimensions by drawing upon other commonly used valid and reliable measures, including Rosenberg‟s (1965) self-esteem scale and Rotter‟s (1966) locus of control scale. The authors created a 28-item scale consisting of five factors: self-esteem/self-efficacy; power/powerlessness; community activism and autonomy; optimism and control over the future; and righteous anger (See “Appendix A: Measurement of Self-concept, Stigma and Responses to Stigma” for the individual items). These factors represent self-oriented as well as general or community-oriented aspects of empowerment. The process used to develop Rogers et al.‟s (1997) measure of empowerment reflects the tendency of current research to focus on self-concept oriented measures, including self-esteem and mastery, as definitional components of empowerment. However, this process confounds these concepts with others that are embedded within the same measure, which may mask the effects of these underlying concepts or may misattribute the driving mechanism within empowerment. Community activism and righteous anger. An empowerment orientation may reduce the harmful effects of stigma on self-concept. However, because it is a broad concept with many different definitions, conceptualizations and measurements, research is needed to determine which components of empowerment relate to stigma and selfconcept. Two measures from Rogers et al.‟s (1997) Empowerment Scale, community

46

activism and righteous anger, are theoretically predicted to relate to stigma resistance, but have received little empirical research to date. I briefly discuss each concept below. First, when constructing the Empowerment Scale, Rogers et al. (1997) included items pertaining to an individual‟s orientation towards the ability of social others (or social collectives) to overcome obstacles. This measure of community activism may explain differential responses among individuals with mental illness to the threat of stigma on self-concept. Corrigan and Watson (2002) and Thoits (2011) discussed three possible responses to stigma. These responses included: 1) acceptance and internalization of the stigmatized status, 2) ignoring the stigma, and 3) challenging stigma. This third response incorporates one‟s identification with similar others in a collectivity to engage in efforts to change the perception of others. Such an approach towards community activism may empower individuals to challenge others and to advocate on behalf of themselves (Corrigan and Watson 2002; Thoits 2011). This may in turn affect self-esteem and mastery (Corrigan et al. 1999). However, as Sibitz et al. (2011) suggest, “further research should focus on the development of a more robust measure of stigma resistance, possibly extending the range of measurements, e.g. toward the impact of community activism” (p. 321). A second concept included within the Empowerment Scale that relates to stigma resistance is righteous anger (Rogers et al. 1997). Corrigan and Watson (2002) noted that certain individuals will respond “forcefully” to stigma with righteous anger (p. 36). However, other stigmatized individuals will not internalize stigma because they either may not be aware of it or because they believe that the stigmatization is illegitimate (Mak

47

et al. 2007). If stigma is viewed as illegitimate, individuals may react with righteous anger towards the prejudice or discrimination that is experienced or perceived (Deegan 1990; Mak et al. 2007). Such a response may empower people to become active participants in their own treatment or in challenging negative stereotypes about mental illness (Corrigan and Watson 2002; Thoits 2011). However, further specification of the concept of righteous anger is needed, particularly as it may represent an active orientation towards stigma resistance. To better understand what is at the core of the concept of empowerment, it is necessary to critically examine how empowerment is constructed and to assess meaningful conceptual differences between concepts that are (or may be) conflated within one measure. This is of particular importance when trying to disentangle the effect of stigma and empowerment on psychological processes, including self-concept. In this dissertation, I present exploratory analyses of the impact of two measures derived from Rogers et al.‟s (1997) Empowerment Scale. This study will assess an orientation towards community activism as a measure of stigma resistance, as well as an orientation towards righteous anger. Both empowerment measures are conceptualized as possible resistance responses to the threat of stigma. This study is an empirical assessment of these theoretical suppositions.

Understanding Stigma Resistance among Community Mental Health Samples Most studies have focused on empowerment among mental health service utilizers who are participants in client-run, self-help programs, also known as a “clubhouse”

48

model of mental health services (Dickerson 1998). The clubhouse model developed out of client dissatisfaction with professionally-led treatment programs (Corrigan and Garman 1997). These programs are typically developed and operated by individuals with mental illness who are clients of the mental health system (Link, Castille and Stuber 2008; Markowitz 1998; Rosenfield 1992, 1997; Thoits 2011), although there is variation in the degree that professionals without mental illness are included within the treatment model (Corrigan and Garman 1997). Underlying this model is the assumption that, because of lived experiences, the client has unique insight about mental illness and how it should be treated (Corrigan and Garman 1997). In addition to providing for a variety of services to meet the basic needs of clients, empowerment-oriented clubhouse programs offer vocational, psychosocial and social opportunities for participants (Mowbray et al. 2006; Rosenfield 1997). The clubhouse model encourages participants to take an active role in planning treatment-oriented programs, such as group therapy meetings, as well as in planning various social activities, daily maintenance of the facility and developing skills related to self-care. However, research on the clubhouse model of mental health services is still needed. Although there is some evidence that the clubhouse model reduces hospitalizations, increases independent living skills, helps clients find jobs and improves feelings of self-efficacy and self-esteem (Corrigan and Garman 1997; Mowbray et al. 2006), little is known about which components of the clubhouse model drive these findings. The clubhouse model is a highly specialized form of mental health services, and clubhouse programs may be markedly different from traditional models of community

49

mental health care. The clubhouse model emphasizes support, self-sufficiency and personal empowerment, and these values may not be as apparent within a traditional community mental health context, and many users of traditional mental health services may feel disenfranchised from participating in treatment decisions (Dickerson 1998). Traditional mental health centers exist within a fragmented system of care and often lack adequate funding and resources, which limits the mental health system from offering an array of services that encourage a recovery-based approach to mental health care (Mechanic and Rochefort 1990; Torrey 2008). Thus, the clubhouse model may not be representative of mental health service utilizers in an outpatient, community setting. Participation in a clubhouse model may also be dependent upon a number of factors. For instance, Mowbray et al. (2006) reported that the majority of clubhouse participants were male, white, and middle aged. Other factors, such as adherence to a treatment program, the degree of psychiatric disability, presence of co-occurring substance use issues, and/or financial resources, may also signal differences between the population of mental health service clients who engage in clubhouse programs and those in the general population of mental health service users. This study broadens this research by assessing aspects of empowerment among individuals with mental illness who vary in level of disability, who are engaged in general community mental health services and are not necessarily affiliated with a clubhouse or other empowerment-oriented program. Such research addresses a gap in knowledge related to the construction of empowerment measures, but also to their

50

applicability among a sample of individuals engaged with a traditional model of mental health services. Research Questions In this chapter, I have summarized theoretical approaches that may explain the link between stigma and self-concept. I have also discussed stigma resistance responses to stigma, including defensive strategies and empowerment. These mechanisms may be used to explain how some stigmatized individuals eschew or overcome the effects of negative labels that are applied. Further, I have provided a brief critique of the construction of empowerment and have made suggestions for further research in this area. In this section, I propose several research questions to examine these relationships. This dissertation focuses on the relationship between stigma and self-concept among individuals with mental illness, while also considering the role of stigma resistance responses that individuals enact to ward off the potentially negative effect of stigma. My primary conceptual relationship of interest is to better understand the effect of stigma on self-concept, which I operationalize as two outcome measures- self-esteem and mastery. Thus, I have research questions that focus on self-esteem as the primary outcome, and questions related to mastery. I first discuss the research questions that focus on self-esteem as the primary dependent variable of interest, and follow this with a statement of the questions related to mastery. Theoretical assertions and empirical findings suggest that stigma has a negative impact on self-concept by reducing self-esteem. However, additional research is still

51

needed to document this theoretical claim. Given this, a primary research question for this study is: Research question 1: Does perceived, public stigma have a negative effect on self-esteem? Further, there is evidence that individuals engage in strategies to defend against negative reactions of others to protect or preserve one‟s self-concept. If that is the case, such strategies would have a moderating effect on the primary focal relationship between stigma and self-concept. In other words, the effect of stigma on self-concept (e.g. selfesteem, mastery) would vary in relation to the presence or absence of stigma resistance strategies. However, the evidence of the effectiveness of such strategies to lessen negative effects of stigma on self-concept is limited (Link et al. 1991). Given this, I pose several research questions to determine the moderating effect of resistance strategies on the potential relationship between stigma and self-concept. If individuals use defensive strategies, the effect of stigma on self-concept would be different for those individuals than if defensive strategies are not used. Since defensive strategies, which include strategies of hiding one‟s mental illness from others and withdrawing from social interaction, may have a negative impact on self-concept, the use of such strategies should exacerbate the harmful impact of stigma. Thus, a related research question is as follows: Research question 2: Do social withdrawal and secrecy moderate the potential relationship between perceived stigma and self-esteem?

52

Empowerment may be another means of resisting stigma, although this theoretical assumption has not been tested. Further, one of the research aims of this study is to provide a critical assessment and more in-depth analysis of the concept of empowerment. I have argued that commonly used conceptualizations of empowerment confound related constructs within one measure. This is particularly problematic for this assessment of the effects of stigma on measures of self-concept as the dependent variables of this analysis are embedded within the measure of empowerment as it has been traditionally conceptualized. This may mask the effect of other important aspects of empowerment that may be indicators of stigma resistance strategies. Given these concerns, I examine two factors related to the construct of empowerment, community activism and righteous anger, as proactive and energetic empowerment orientations that may impact the link between stigma and self-concept. If empowerment is a positive or beneficial means of coping with stigma, then it would be expected that dimensions of empowerment, community activism and righteous anger, would have ameliorative effects on the relationship between stigma and selfconcept. In other words, the adoption of an empowerment orientation that focuses on community activism and righteous anger by some individuals should mitigate negative effects of stigma on self-concept. Thus, additional research questions are: Research question 3: Does community activism moderate the relationship between perceived stigma and self-esteem? Research question 4: Does righteous anger moderate the relationship between perceived stigma and self-esteem?

53

The research questions pertaining to self-esteem are depicted in Figure 2.1.

Figure 2.1. Conceptual Model: Self-esteem In addition to using self-esteem as an outcome measure in this relationship, I assess the effect of stigma on mastery because this concept incorporates an individual‟s sense of personal power to overcome adversity. Below are the related research questions that focus on mastery as the primary outcome measure. See also Figure 2.2. Research question 5: Does perceived, public stigma have a negative effect on mastery? Research question 6: Do social withdrawal and secrecy moderate the potential relationship between perceived stigma and mastery? Research question 7: Does community activism moderate the relationship between perceived stigma and mastery? Research question 8: Does righteous anger moderate the relationship between perceived stigma and mastery?

54

Figure 2.2. Conceptual Model: Mastery The proposed analyses will examine the moderating effects of defensive responses and empowerment measures on the relationship between stigma and selfconcept. Given the recent theoretical assumptions related to stigma resistance, it is assumed that the use of such strategies would fundamentally alter the effect of stigma on self-concept for those individuals who engage in defensive strategies or adopt an empowerment approach. Specifically, empowerment approaches may lessen or ameliorate negative effects of stigma on self-concept, while the use of defensive strategies may exacerbate or worsen the effects of stigma. These relationships will be assessed in the following chapters.

CHAPTER 3

DATA AND METHODOLOGY

In this chapter, I describe the data and sample used for this study. First, I begin with an overview of the data collection process, and discuss how the sample was generated and the characteristics of the sample. I then provide information about the measures used within this analysis. I conclude by discussing the plan for analysis for examining the research questions set forth in the previous chapter.

Data and Sampling Data Source The data used for this project are from a longitudinal study of individuals living with mental illness who are engaged in community-based mental health services in an urban area in Northeast Ohio (“The Quality of Life of People with Mental Illness” Principal Investigators: Christian Ritter and Mark R. Munetz). This study sample is comprised of over 200 individuals with a broad range of psychiatric diagnoses. Thus, this study presents a more representative sample of individuals with mental illness, and will allow for the examination of the extent to which the effect of stigma on self-concept may differ for those who have been diagnosed with different mental disorders.

55

56

One of the original aims of the Quality of Life of People with Mental Illness study (I will refer to this study as Quality of Life, in short) was to examine the consequences of jail diversion programs on a population of individuals living with mental illness who had had some contact with the criminal justice system. In general, there is a high level of involvement of individuals with mental illness and the criminal justice system. A study conducted by the National Alliance on Mental Illness found that 44 percent of individuals with severe mental illness in their sample had been arrested or detained by police (Hall et al. 2003). Further, approximately 10 percent of jail inmates have a mental disorder (Treatment Advocacy Center 2009). This issue is the focus of a body of research beyond the scope of this dissertation that examines how individuals may be diverted from the criminal justice system and linked to the mental health system. However, this aim did direct the recruitment processes for the Quality of Life study which has implications for the present study. Specific jail diversion programs of interest in the original Quality of Life study include Crisis Intervention Team (CIT) training for law enforcement officers and Mental Health Court (MHC) specialty court docket programs, and some individuals within the sample were recruited from these sources. The sample consists of individuals who had had encounters with the criminal justice system, either through interaction with a CIT officer or via referral to or participation in the MHC program, and a control group of individuals with mental illness who had no involvement with the criminal justice system at the outset of the study. Approximately 52 percent of the final sample for this present study became eligible to

57

participate in this study as a consequence of some degree of suspected criminal activity that resulted in criminal justice involvement. The majority of these became known to the research team via their eligibility for or participation in the MHC program (46 percent), while a small percentage of the final sample (6 percent) had been referred to a community mental health emergency center by CIT-trained law enforcement officers. Still others who did not have contact with the criminal justice system (48 percent of the final sample for this dissertation) were recruited by study personnel reaching out to clients of a community-based mental health center (30 percent) and by reaching out to those who were referred to the community mental health emergency center by a referral source other than the police (18 percent). Thus, this sample is not a random sample of individuals with mental illness, and the recruitment sources were a function of the sampling parameters for the original Quality of Life study. The Quality of Life study provides an opportunity to better understand psychosocial processes of a sample of individuals with severe mental illness who are engaged in community mental health treatment and who may or may not have had some contact with the criminal justice system. While it is an important topic, criminal justice involvement of individuals with mental illness is not a focal issue for this present study as I am seeking to investigate how stigma associated with mental illness relates to selfconcept regardless of criminal justice involvement. I will, however, examine the potential effect of prior contact with the criminal justice system on the research questions posed previously (see “Appendix B: Supplemental Technical Data Analyses” for additional detail).

58

Sample Recruitment and Data Collection Activities All eligible and potential participants who were recruited for possible participation in the Quality of Life study were contacted by members of the research team who provided information about the study and assessed the level of interest in participating. If potential participants were interested, trained members of the research team followed-up to schedule an appointment for interview at a community mental health treatment facility. The research team arranged for cab service if transportation to the interview was needed. Interviewers met each participant and escorted them to a private interview room to review the consent process and conduct the interviews. Each interviewer was certified in human subjects research and trained in the medical emergency and privacy protocols required by the mental health treatment agency. Each interview itself lasted approximately 60 minutes. Interviews were conducted with 370 individuals from September 2002 through November 2005. 262 individuals participated in a follow-up interview approximately 6 months after the initial interview (February 2003 through March 2006). This resulted in a 71 percent retention rate between the first and second interviews. Participants were compensated $20 for participating at Time 1, and $25 for participating at Time 2. The Quality of Life interviews focused on multiple concepts to assess participant well-being and quality of life. The interview included measures to assess perceived stigma, self-concept (e.g. mastery, self-esteem), perceptions about mental illness, social support and social networks, empowerment and use of services. This interview also contained items to assess one‟s general life satisfaction and satisfaction across different

59

domains, including satisfaction with living situation, leisure activities, family, social relations, finances, work (or school), safety, and physical health (Lehman 1997). The measures relevant to this study are further described in the sections that follow. In addition to conducting interviews, a team of clinical supervisors employed by the research team reviewed medical charts of all participants in the study in order to obtain all psychiatric diagnoses, including co-occurring substance use diagnoses. If there was more than one diagnosis within the chart, the clinical supervisor then made a determination about which diagnosis was the principal, or primary, diagnosis. Each medical chart was reviewed by two clinical supervisors to determine inter-rater reliability. Medical charts with discrepancies among reviewers were assessed by a psychiatrist who made the final determination of the principal diagnosis. Even with this careful check, there will still a number of study participants where no diagnosis was available. Thus, while the sample of those who participated in the Time 2 survey administration was 262, the final sample size for this study was reduced to 221 due to cases that were missing on psychiatric diagnosis. All study participants provided informed consent for all aspects of the study, including the interview and medical chart review. The Institutional Review Boards (IRB) of Kent State University and the Northeastern Ohio Universities College of Medicine (NEOUCOM, now Northeast Ohio Medical University) approved this study (initial study protocol numbers: 02-250 (KSU); 02-032 (NEOUCOM)). The original study protocol has been approved as an exempt study by the IRB at Northeast Ohio Medical University (protocol # 10-066). The IRB at Kent State University has approved the present study as

60

a project involving secondary data analysis (protocol #12-299). Please see Appendix C for copies of the notices of IRB approval.

Sample Description Demographic characteristics of this sample can be found in Table 3.1. The average age of the sample is just over 40 years old. Thirty nine percent of the sample is female. Half of the sample (50 percent) is white, 42 percent is African American, and 8 percent is of another race, primarily Hispanic. On average, the sample has completed fewer than 12 years of education ( ̅ = 11.87, SD = 1.983), and the average monthly income of the study participants is approximately $675, or $8,100 annually. Given that this sample consists of individuals with mental illness, I included several binary measures for psychiatric diagnoses. Over half of the sample have a principal diagnosis of schizophrenia (51 percent), and 19 percent have a principal diagnosis of depressive disorder. Seventeen percent of the sample have a diagnosis of bipolar disorder, and 13 percent have a principal diagnosis of “other,” representing an inclusive category of all other psychiatric diagnoses that may have been a principal diagnosis, including anxiety disorders, substance use diagnoses, and post-traumatic stress disorder, among others. I include control measures that relate to various roles held by members of the sample, particularly as role statuses may be of particular importance to the self-concept of individuals (Camp et al. 2002; Simon 1997). As shown in Table 3.1, 22 percent of the sample is employed either on a part-time or full-time basis. The majority of the study

61

participants are not married, with 55 percent never married, and an additional 37 percent formerly married (either divorced or widowed). Only 8 percent of the sample is currently married or partnered. Finally, 58 percent of the sample has children. Table 3.1. Sample Characteristics: Means, Proportions, Standard Deviation and Alpha Reliability (N = 221) Variable Mean S.D. Range Demographic factors Age 40.26 10.203 18 – 64 Gender (1=female) .39 .489 0–1 White .50 .501 0–1 African American .42 .495 0–1 Other race .08 .274 0–1 Education (years completed) 11.87 1.983 7 – 18 Income (monthly) 675.48 487.701 0 – 3200 Psychiatric diagnoses Schizophrenia .51 .501 0–1 Depressive disorders .19 .393 0–1 Bipolar disorder .17 .374 0–1 Other diagnoses .13 .333 0–1 Role-based statuses Employed .22 .413 0–1 Never married .55 .498 0–1 Formerly married .37 .483 0–1 Currently married/cohabiting .08 .274 0–1 Respondent has children .58 .495 0–1

Measures This section briefly describes the primary concepts of interest for this study, including self-concept (dependent measures), perceived stigma (independent measure) and potential moderating measures related to stigma resistance response items, as well as demographic and control measures, including psychological functioning and social resources. The conceptualization for these measures is further detailed in Table 3.2.

62

Table 3.2. Correspondence of Theoretical Concepts, Research Concepts and Empirical Indicators Theoretical concepts Research concepts Empirical indicator Self-concept Self-esteem Self-Esteem Scale Rosenberg (1965) Mastery

Personal Control / Mastery Scale (Pearlin and Schooler 1978)

Stigma

Perceptions of public endorsement of negative views towards people with mental illness

Perceived Devaluation and Discrimination Scale (Link 1987)

Stigma resistance responses

Defensive coping orientation/strategies

Social withdrawal Secrecy (scale adapted from Link et al. 1991 and Wright et al. 2000)

Empowerment

Righteous anger Community activism (scales adapted from Rogers et al. 1997)

Gender

Dichotomous measure (Female = 1)

Race

African American, white, other

Age

Age (coded in years)

Education

Number of years of education

Financial resources

Monthly income

Psychiatric diagnosis

Primary Axis I psychiatric diagnosis at time of entry into study (per medical chart) [Primary categories: schizophrenia; bipolar disorder; major depression; other diagnoses (including substance use disorders)]

Role-based statuses

Employment status (1=working) Marital status Parental status

Sample characteristics

63

Table 3.2 (continued). Correspondence of Theoretical Concepts, Research Concepts and Empirical Indicators Psychological Stability of psychological Depression (CES-D) at time 1 functioning distress Anxiety (SDS) at time 1 (functionality) Participant‟s level of Interviewer appraisal: Level of comprehension difficulty in comprehending questions Interviewer observation: significant problem during the interview Interviewer observation: significant problem was comprehension Interviewer observation: significant problem was some other concern Social resources Social support Frequency of contact with family members (Lehman) Frequency of contact with social relations (Lehman) Social embeddedness

Extensiveness of network (Fischer) Number of social ties

Self-concept Self-esteem. Self-esteem is assessed using Rosenberg‟s (1965) measure of selfesteem. This 10-item measure incorporates two dimensions of self-esteem based on the respondent‟s positive (self-worth) and negative (self-deprecation) attitudes towards self (per Wright et al. 2000). The positive self-esteem indicators include “I feel that I am a person of worth, at least the equal of others,” “I feel that I have a number of good qualities,” “I am able to do things as well as most other people,” “I take a positive attitude toward myself,” and “On the whole, I am satisfied with myself.” The negative self-esteem items include “At times I think I am no good at all,” “I feel I do not have much to be proud of,” “I certainly feel useless at times,” “I wish I could have more respect for myself,” and “All in all, I feel that I am a failure.” The self-esteem scale is

64

coded using a six-item Likert scale ranging from “strongly disagree” to “strongly agree,” and items were coded such that higher scores indicate higher levels of positive selfesteem. The alpha reliability for this scale is .875. Mastery. Mastery is measured by a seven item scale (Pearlin and Schooler 1978). Items in this scale relate to the control one feels over his or her life, ability to solve problems in one‟s life, feeling helpless, feeling pushed around, belief in control over one‟s future and belief in one‟s ability to accomplish things. The response categories for this scale range from “strongly disagree” to “strongly agree”. The range for this scale is from 7, indicating low levels of mastery, to 42, indicating high levels of mastery. The alpha reliability for mastery at Time 2 is .866.

Stigma Devaluation/ discrimination is included as a measure of perceived, public stigma. This is a 12-item scale (α = .875) assessing the extent to which an individual believes most people will devalue or discriminate against a person with mental illness (Link 1987). Response categories are a Likert scale ranging from “strongly disagree” to “strongly agree” for questions related to how “most people” view people with mental illness. The range for this scale is from 12, indicating low levels of perceived devaluation or discrimination, to 72, indicating high levels perceived devaluation or discrimination. Sample items include “Most people feel that having a mental illness is a sign of personal failure” and “Most employers will pass over the applications of a former mental patient in

65

favor of another applicant.” This measure is the primary independent measure as it is conceptualized to have deleterious effects on the self-concept of individuals.

Stigma Resistance Responses Defensive strategies. This study uses Wright et al.‟s (2000) measure of defensive strategies. This is an 11-item scale to measure an individual‟s tendency to cope with stigma by keeping their illness secret or withdrawing from potentially stigmatizing situations. Respondents answered “yes” (1) or “no” (0) to a series of statements assessing whether they would use particular defensive strategies, meaning those strategies that would minimize possible stigmatization or discrimination through withdrawal or inaction, in a variety of situations. Sample items are “When you meet people for the first time do you ever tell them that you were once mentally ill?” and “Do you think it is a good idea to keep your history of mental illness a secret?” This measure is adapted from Link and colleagues (1991) and includes items related to withdrawal and secrecy coping responses (See Appendix A for the individual items that are drawn from Link et al. (1991)). Items are coded so that higher scores indicate greater use of defensive strategies of secrecy and withdrawal (α = .614). Empowerment. The Boston University Empowerment Scale incorporates 28-items to assess overall personal empowerment. This measure was developed by Rogers et al. (1997) based on input from a group of individuals with mental illness who were involved in the self-help movement. The measure incorporates internally-oriented factors related to self-esteem, self-efficacy and autonomy. However, other items within the scale indicate a

66

more external or socially-based orientation and focus on power and community activism. Finally, there are attitudinal measures related to optimism and righteous anger. For this dissertation, I specifically focus on two dimensions of empowerment discussed by Rogers et al. (1997): community activism and righteous anger. These dimensions represent active or energetic aspects of empowerment that are conceptually distinct from self-esteem and mastery. Further, each of these concepts has been suggested as either the antithesis of stigma or as an expected reaction to stigma that is perceived to be unjust (Corrigan and Watson 2002). The scale assessing community activism (α = .704) includes four measures such as “people working together can have an effect on their community” and “working with others in my community can help to change things for the better.” The measure of righteous anger include three items: “Getting angry about something never helps,” “People have no right to get angry just because they do not like something” and “making waves never gets you anywhere” (reverse coded) (α = .687). Response categories for items in both scales range from strongly disagree to strongly agree (6-point Likert scale). The variables are coded such that higher scores indicate agreement, and negatively worded items were reverse coded for consistency. To arrive at these measures, I conducted analyses on the factor structure of the full empowerment scale to disentangle the self-concept oriented measures from other constructs. I examined the inter-item reliability and correlations among items after removing the self-concept (self-esteem and mastery) measures from the full scale. To do this, I conducted Principal Component Analysis (PCA), an exploratory factor analysis

67

data reduction technique, with Varimax rotation for the 28-items of the full empowerment scale. Given that my sample is of comparable size and make-up to Rogers et al.‟s (1997) initial study sample (e.g. individuals with mental illness who are engaged in mental health services), it could be expected that the factor structure would be similar. For these preliminary analyses, all 28 items loaded onto one or more factors with a factor loading of .40 or greater. Table 3.3 lists the 28-item factor analysis with the factor loading score. Eigenvalues and the total variance explained are also shown for each factor. The data sorted into seven factors. While there are similarities among the factor structure of my data and that reported by Rogers et al. (1997), my results did not replicate an identical factor structure. For comparison, Table 3.4 shows the factor structure reported by Rogers et al. (1997). As I have argued, a measure that incorporates self-concept items within the construct of empowerment may be misleading and confounding, particularly within a social psychological analysis of the impact of stigma on one‟s self-concept. As such, I conducted additional exploratory factor analyses of the empowerment scale by removing those items that are indicative of self-concept (self-esteem and mastery). To do this, I removed all items that ask the respondent to think of their sense of self within the context of the question. Many of these items are derived from Rotter‟s (1966) Locus of Control scale (e.g. “When I make plans, I am almost certain to make them work”), Rosenberg‟s (1965) self-esteem scale (“I have a positive attitude about myself”) and the Self-Efficacy Scale (“I see myself as a capable person”) (Sherer and Adams1983).

68

69

70

By removing these self-oriented items in this exploratory way, I am able to disentangle those items within the concept of empowerment that may confound our understanding of any relationship between stigma and self-concept. Removing the selforiented items from the empowerment scale maps onto Rogers et al.‟s (1997) in that the items from the Self-esteem/Self-efficacy (Factor 1) and the self-oriented items from the Power/Powerless (Factor 2) factors are removed (see Table 3.4). This analysis of the remaining 13 items yielded a 4-factor structure. Table 3.5 lists the 13-item factor analysis and factor loading scores of these items. Based on this factor analysis, I examined the items associated with each factor. Items that loaded onto the first factor are those measures that relate to Rogers et al.‟s (1997) factor of “community activism.” However, the items measuring autonomy did not load onto the community activism factor as they did for Rogers et al. (1997), with the exception of one item (“People should try to live their lives the way they want to”), which had a factor loading of .41. Other items related to autonomy (e.g. “People have a right to make their own decisions, even if they are bad ones”) did not load onto this factor. This suggests that the measure of community activism represents one concept, community activism, and not a blended concept of community activism and autonomy as suggested by Rogers et al. (1997). Reliability analyses of these five items further suggests that community activism and autonomy are separate constructs because the scale reliability would be improved if the item assessing autonomy (“People should try to live

71

Table 3.5. Principal Component Analysis Factor Loadings for Empowerment Scale, Not Including Self-Concept Items (13 items) Factor and Scale Item Loading Factor 1 Working with others in my community can help to change things for the .82 better People working together can have an effect on their community .70 a People have more power if they join together as a group .67 Very often a problem can be solved by taking action .65 People should try to live their lives the way they want to .41

1 3

Factor 2 Making waves never gets you anywhere Getting angry about something never helps People have no right to get angry just because they don‟t like something You can‟t fight city hall (authority)

.75 .73 .71 .61

Factor 3 Getting angry about something is often the first step toward changing it People are only limited by what they think is possible People have more power if they join together as a group a

.73 .70 .40

Factor 4 People have a right to make their own decisions, even if they are bad ones Experts are in the best position to decide what people should do or learn

.79 .70

Eigenvalue = 2.47, variance explained = 19.0 percent; 2 Eigenvalue = 2.24, variance explained = 17.3 percent; Eigenvalue = 1.29, variance explained = 10.0 percent; 4 Eigenvalue = 1.29, variance explained = 9.9 percent

their lives the way they want to”) is dropped from the scale (Cronbach‟s α is improved from .668 to .704). Thus, reliability analyses of the four community activism items support this assertion of construct validity, as these items have adequate internal consistency. The items that loaded onto the second factor are indicative of the “righteous anger” construct described by Rogers et al. (1997). Four items had sufficient factor loading scores to indicate adequate consistency among the items. When scaled,

72

Cronbach‟s α was improved by dropping one item (“you cannot fight city hall (authority)”). Removing this item increased the Cronbach‟s α score to .686. The items that loaded onto the third factor are suggestive of perceptions about one‟s ability to bring about change. However, these items also did not hold up to reliability testing and could not be incorporated into one scale item. The items that loaded on the fourth factor are indicative of measures of autonomy (per Rogers et al. (1997)). Again, however, reliability analyses indicated that these items did not have adequate internal consistency or validity. These exploratory findings using PCA factor analysis suggest that there are two aspects of empowerment, righteous anger and community activism, that are conceptually different from other items included within the full empowerment measure. These two measures are not significantly correlated with each other (r = -.085, p= .207), providing further evidence that they represent two distinct dimensions of the concept of empowerment that merit further analytical and theoretical consideration.

Psychological Functioning As each member of the sample lived independently in the community at the time of the interview, and was not in a psychiatric hospital, it is assumed that there is some level of social and psychological functioning despite the presence of a severe mental illness. I assess psychological functioning to control for any possible effect related to the study participant‟s level of ability or capacity to engage with the research process. While a measure of global psychological functioning, such as the Global Assessment of

73

Functioning or the Multnomah Community Ability Scale (MCAS) would have served to assess the participant‟s degree of social and psychological functioning (Barker et al. 1994; Hall 1995), such measures were not available in the data. As such, I use several proxy measures to assess the level of psychiatric functioning and the potential impact that may have on the validity of the data collected. First, I included the participant‟s level of anxiety and depression at the Time 1 survey administration. These measures provide an assessment of the stability of mental well-being over time and are relevant to measures of self-concept. The measure for anxiety is the Symptoms of Distress Scale, which is adapted from the Symptom Checklist-90 (SCL-90) and the Brief Symptom Inventory (BSI) (Derogatis 1975; 1977). The Symptoms of Distress Scale was developed by the Behavioral Health Services Division and is part of the Mental Health Statistics Improvement Program (MHSIP) (Ganju 1999). This scale consists of items which ask the respondent to report the extent to which he or she was bothered by symptoms experienced within the past week. The scale consists of 15 items, such as “feeling nervousness or shakiness inside,” “being suddenly scared for no reason” and “feeling so restless you couldn‟t sit still.” Response categories range from 1 (“not at all”) to 5 (“extremely”). Cronbach‟s α score = .930. Depression was measured using the 20-item Center for the Epidemiologic Studies-Depression (CES-D) scale (Radloff 1977). This measure asks respondents to report how frequently within the past week he or she experienced a variety of feelings and emotions, such as “felt depressed,” “felt lonely,” or “felt happy.” Several behavioral items are included as well, such as “had restless sleep” or “had crying spells.” The

74

response categories for these items are 0 (“rarely or never”), 1 (“some or a little of the time”), 2 (“a lot of the time”) and 3 (“most or all of the time”). Cronbach‟s α for this scale is .933. Second, as a means of assessing functionality, I also included some measures provided by the interviewer about his or her perceptions of the respondent during the course of the interview. These measures include the interviewer‟s assessment of the level of difficulty of the participant in comprehending the questions, whether or not there was a significant problem during the interview, and if so, if that problem was related to comprehension or some other concern (e.g. a health issue that may have impacted the progression of the interview). At the conclusion of the interview, the interviewer noted the level of difficulty the respondent had in comprehending the questions. Response categories are none, slight, a fair amount, and a lot (range 1 – 4). I also included a binary measure indicating whether or not the interviewer perceived a significant problem during the interview (1=yes). For those cases where a significant problem was reported, I include two binary measures to determine if the problem was related to the respondent comprehending the questions (1=yes), or if the problem was related to some other concern that arose during the course of the interview (e.g. participant became physically ill during the interview) (1=yes).

Social Resources In order to better understand the possible relationships among the primary measures of interest (stigma, self-concept and stigma resistance responses) it is necessary

75

to incorporate a more complete understanding of the array of social resources that one has that may impact the relationship among my focal measures. Social networks and social support are vital components of sociological research and refer to one‟s connectedness to others in society (Barrera 1986; Turner and Marino 1994). Social networks are comprised of those ties an individual may go to as a source of social support, and considers the availability of ties to provide aid. Social support refers to aspects of support, either received or perceived, from individuals in the network. Network theorists posit that it is the number of ties, or the perceived level of availability of those ties, that is beneficial to health and well-being, while a support perspective would argue that it is the strength of those ties that is beneficial (Barrera 1986; Lin and Peek 1999; Smith and Christakis 2008). In this study, I examine the social embeddedness and social networks aspects of social support. Social embeddedness is a relevant construct for this analysis because it represents the “flip side of social isolation” and relates to a psychological sense of community (Barrera 1986: 415). There is a potential connection between social embeddedness and adopting stigma resistance strategies. Lack of embeddedness, that is, social isolation, may correlate with the use of defensive strategies, including social withdrawal and secrecy, while social engagement may be linked to community activism. I included several control measures to assess one‟s level of social engagement and embeddedness. First, I include a measure of the frequency of contact with one’s family members, including how often the respondent talks to family on the telephone and how often the respondent gets together with family. Response categories ranged from 1

76

(“not at all”) to 5 (“at least once a day”). Respondents who reported there were no family members to contact were given a code of 1 to indicate no contact. These two items were summed and were correlated (r= .312; p< .001). I also included a measure to assess the frequency of contact with social relations. This measure includes seven items assessing how frequently the respondent has contact with other people who are not in his or her family. This item combines the frequency in which the respondent does things with a close friend, visits with someone outside of the respondent‟s home, calls someone (other than a relative), writes a letter to someone, emails someone, does something that was planned ahead of time, and how often the respondent spends time with someone who is considered a special or close friend. Response categories ranged from 1 (“Not at all”) to 5 (“at least once a day”). These seven items were summed, with an alpha reliability of .748. I also included two measures to assess one‟s degree of social embeddedness. First, I included a count of the number of social network ties the respondent named ( ̅ = 4.23, SD = 2.66). Second, I included a measure of the extensiveness of one‟s social network (Fischer 1982). This item is created by summing the total number of persons available in one‟s social network who are available to the respondent, either to provide different types of assistance or social support (e.g. “In the past six months, who has helped you with household tasks?” Or “In the past six months, with whom have you discussed personal worries?”) or by serving an important role to the respondent (e.g. “In the past six months, who has been a partner?”). The measure is calculated by subtracting the number of areas where there is no one to provide support from 9 (the total number of areas to be covered),

77

and then combining this with the number of areas where the respondent has named multiple persons who provide support. For each area where there is more than one supporter named, a value of .5 is added. Thus, a measure is created that summarizes the extent of availability of one‟s social network.

Sample Characteristics All analyses included demographic variables as controls, including age, gender, race, education, and income. Age is measured in years. Gender is coded as a dummy variable with 1=female, 0=male. Race is coded into three dummy measures, with 1=white (0=all other races), 1=black (0= all other races), and 1=other race (0=white and black). The level of education is coded as the number of years of school completed. Income is a self-report measure of one‟s monthly income in U.S. dollars. As this sample consists of individuals with severe and persistent mental illness, I included the primary psychiatric diagnosis. All clients of the community mental health agency have a primary Axis I diagnosis, with the possibility that there are other co-occurring Axis I or II diagnoses. Psychiatric diagnoses were collected by a team of clinical supervisors employed by the research team as previously described. For this study, primary psychiatric diagnosis is assessed by a series of dummy variables representing schizophrenia, depressive disorders, bipolar disorder, and all other diagnoses. I also included role-based statuses, including marital status, parental and employment statuses. Marital status is coded as three dummy measures representing 1=never married (0=all other categories), formerly married (including divorced and

78

widowed) =1 (0=all other categories) and currently married or cohabitating=1, 0=all other categories). Parental status is coded as a dummy variable to assess if the respondent has any children (1=yes, 0=no). Employment status is a dummy variable to indicate whether the respondent is employed, either full- or part-time, or not (1=yes, 0=no).

Analytic Strategy Using this longitudinal survey data, I explore the primary, focal relationships among stigma and self-concept by examining the effects of perceived devaluation/ discrimination on self-esteem, as well as its effect on mastery. Because of the longitudinal nature of the data, I am able to control for Time 1 levels of self-esteem and mastery. While I do not anticipate a change in each measure over a six month time period, particularly since there was no intervention, including the Time 1 measure of each dependent variable provides a stability measure and reduces potentially spurious variability. I use these two dependent measures because each represents a separate dimension of self-concept, with self-esteem assessing general, positive and negative feelings towards oneself, and mastery assessing feelings towards one‟s sense of efficacy, power and control. I then examine the moderating effects of defensive strategies, community activism and righteous anger on the relationships between stigma and self-concept as such responses to stigma may impact the relationship between stigma and self-concept. One goal of this study is to better understand the use of defensive strategies and empowerment as stigma resistance responses. Defensive strategies relates to withdrawal

79

from social connections and hiding aspects of oneself from others. Empowerment in general and the measures used for this study, community activism and righteous anger, are conceptually related to feeling in control and eliciting change in one‟s environment. Given this, it is conceivable that each stigma response may impact self-esteem and mastery in different ways. Further, I will use a variety of analytical approaches to test the research questions outlined above. I will report descriptive statistics (e.g. mean, median, range) for all measures for the sample, as well as correlations among all measures. I use factor analysis and scale construction techniques to assess the reliability of the proposed measures. For the multivariate analyses, I utilize ordinary least squares (OLS) regression and structural equation modeling. All analyses are conducted using SPSS Version 19 and AMOS Version 21. It is possible that relationships beyond those suggested by the research questions may persist. Such relationships will be assessed and modeled in post-hoc analyses if their presence is detected, but these relationships are not predicted by the initial research questions posed for this study. I organize the analyses presented in the next chapters by focusing first on the research questions with self-esteem as the outcome measure, and then with mastery as the dependent measure.

CHAPTER 4

RESULTS

In this chapter, I present the results of data analyses to address the central research questions for this study. I review the descriptive statistics and correlations for the various measures that I have included. I discuss results related to the Ordinary Least Squares (OLS) regression analyses examining the relationships between stigma and self-esteem and stigma and mastery. I also test for interactive effects of stigma resistance responses on these potential relationships.

A Note on the Presentation of Results I have organized the discussion of findings from the correlation and OLS regression analyses into three sections: (1) Descriptive Statistics (beginning on page 81): In this section, I present sample characteristics and describe correlations for the various control, independent and dependent measures for this analysis. (2) Initial Exploratory Analyses (beginning on page 90): In this section, I discuss exploratory analyses among all control measures that theoretically may impact the relationship between stigma and self-concept. The goal of these initial exploratory analyses is to determine the demographic and control measures that are most closely related to the key measures of interest in this study, which I then use in all subsequent 80

81

analyses. However, given the number of measures under consideration, and the correlation of these items, there are considerable threats of multicollinearity present within these exploratory analyses. I address these concerns and correct for them in the following section. (3) Analysis of Research Questions (beginning on page 101): This section contains the research analyses of the research questions outlined in Chapter 2. Specifically, I examine the effect of public stigma on self-esteem and mastery, as well as the possible moderating effect of defensive strategies, community activism and righteous anger. I examine these relationships while holding constant the demographic and other control measures that were significantly correlated with the key measures of interest. It should also be noted that, given the exploratory nature of these analyses, I have relaxed the probability threshold to include p-values of .10 to identify statistically weak yet theoretically meaningful relationships.

Part 1. Descriptive Statistics In addition to the sample‟s demographic characteristics described in Chapter 3 (See Table 3.1), I examined several other factors that may influence the theorized relationships among my focal measures. Specifically, I included control measures to assess psychological functionality and the participant‟s level of comprehension at the time of the interview, as well as the social resources an individual has that may influence the effect of stigma. Descriptive statistics for these items are presented in Table 4.1.

82

For anxiety (Time 1), the average score on the Symptoms of Distress Scale was 32.99 (SD = 13.035), with a range of 15 – 75, a moderate score. The average score for depression at Time 1, as measured by the CES-D, was 23.03 (SD = 13.031) with a range of 0 – 60. This average is higher than the established threshold for depression among the general population, which is a score of 16 (Radloff 1977). This indicates a high degree of depressive symptoms at the first administration of the survey among this sample. To assess the respondent‟s level of difficulty in comprehending the questions asked, I included items based on the interviewer‟s observations immediately following the interview. There was no reported difficulty in the respondent comprehending the questions for 193 cases (87 percent). Further, there were no interviews that were rated with a level of difficulty in comprehension as “a lot.” Twenty one (10 percent) interviews were rated as “slight” difficulty, and seven (3 percent) were rated as a “fair amount” of difficulty in comprehension. The presence of any significant problem occurring during the interview was reported among 9 percent of interviews (n=20). Of these incidents, comprehension was a significant problem for 12 cases, and some other concern (e.g. participant was not feeling well) accounted for 15 cases (categories are not mutually exclusive). Table 4.1 further presents the descriptive data regarding the availability of social resources. Respondents reported relatively frequent contact with family members, with a mean of 7.18 (SD=2.023; range: 2 – 10). The mean score for frequency of contact with social relations was 20.33 (SD=5.530; range: 7 – 35). With regards to the number of social network ties, the majority of respondents (85 percent) named between 1 and 6

83

persons who were available to assist them in a number of different areas ( ̅ = 4.23). Three percent named “no one” and 12 percent named seven or more. The range for this measure is 0 to 18. The mean score for Fischer‟s (1982) measure assessing extensiveness of network is 8.03 (SD=1.250; range: 4.5 = 13), indicating a moderate level of depth within one‟s social network. I included the Time 1 measures for the two dependent variables for this studyself-esteem and mastery. It is important to adjust for the Time 1 measure of the dependent variables because the Time 1 measure of each is the variable that would be most highly correlated with the Time 2 measure. Including the Time 1 stability measures as controls reduces extraneous variability from those relationships that may exist among Time 2 independent variables affecting the Time 2 dependent variable. In other words, the Time 1 stability measure is included to highlight actual variability that is occurring among the independent and dependent measures net of the effect of the Time 1 measure, and reduces or eliminates spurious variability that does not truly exist. The Time 1 measure of selfesteem had a mean score of 40.74 (SD=10.524). For Time 2, the mean score is 41.48 (SD=10.010). Thus, the mean score of self-esteem is relatively high and stable across survey administrations. The mean of mastery at Time 1 is 28.74 (SD=6.648) and at Time 2 is 28.73 (SD=6.429), also suggesting a relatively high and stable score across time.

84

Table 4.1. Descriptive Statistics for Control, Independent, Dependent and Intervening Measures: Means, Proportions, Standard Deviation and Alpha Reliability (N = 221) Variable Mean S.D. Range Alpha Psychological functioning Anxiety (T1) 32.99 13.035 15 – 75 .930 Depression (T1) 23.03 13.031 0 – 60 .933 Difficulty in comprehending Qs 1.16 .444 1–4 Interviewer observation: sig. problem .09 .288 0–1 Comprehension is sig. problem .05 .227 0–1 Other concern is sig. problem .07 .252 0–1 Social resources Frequency of contact (family) (T2) 7.18 2.023 2 – 10 .475 Frequency of contact (social) (T2) 20.33 5.530 7 – 35 .748 Number of network ties 4.23 2.662 0 – 18 Extensiveness of network 8.03 1.250 4.5 – 13 Focal measures Self-esteem (T1) 40.74 10.524 10 – 60 .908 Self-esteem (T2) 41.48 10.010 10 – 60 .875 Mastery (T1) 28.74 6.648 7 – 42 .812 Mastery (T2) 28.73 6.429 7 – 42 .866 Devaluation/discrimination (T2) 46.98 10.258 12 – 72 .875 Defensive strategies (T2) 6.29 2.581 0 – 11 .614 19.24 2.678 4 – 24 .704 Empowerment: community activism (T2) 9.45 3.275 3 – 18 .687 Empowerment: righteous anger (T2)

The primary independent measure for public stigma, perceived devaluation/ discrimination, had a mean score of 46.98 (SD=10.258). I included three measures to assess stigma resistance responses among the sample. The measure of defensive strategies has a mean score of 6.29 (SD=2.581), indicating that participants reported engaging an average of just over six different defense strategies during interactions with others. Community activism had a mean score of 19.24 (SD=2.678), and righteous anger had a mean score of 9.45 (SD=3.275). Table 4.2 presents Pearson‟s correlation coefficients (r) for the demographic and psychological functioning measures with the primary independent (devaluation/

85

discrimination) and dependent measures (self-esteem and mastery). No demographic factors are significantly correlated with either self-esteem or mastery. However, being female and years of education are both positively correlated with devaluation/ discrimination (r=.15, p