American Journal of Orthopsychiatry © 2017 Global Alliance for Behavioral Health and Social Justice

2017, Vol. 87, No. 5, 575–584 http://dx.doi.org/10.1037/ort0000250

Mental Illness Disclosure Decision Making

Rohini Pahwa Anthony Fulginiti
New York University University of Denver

John S. Brekke and Eric Rice

University of Southern California

Disclosure related to mental illness has been linked to various positive outcomes, including better
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

mental health. However, many individuals with serious mental illness (SMI) continue to practice
This document is copyrighted by the American Psychological Association or one of its allied publishers.

non-disclosure. Even though disclosure inherently occurs within the context of one’s social relation-
ships, research has generally conceptualized mental illness disclosure as an individual level phenom-
enon and neglected to consider preferences concerning to whom an individual discloses and the
factors that influence this decision. The current study uses the disclosure decision-making model
(DD-MM) by Greene (2009) to better understand the processes of mental illness disclosure preference
and selective disclosure for individuals with SMI (n ⫽ 60) using multivariate random intercept
logistic regression with an emphasis on the constituent factors of disclosure preference at both
individual and relational levels. The majority of participants were found to practice selective
disclosure, with 68% of the participants identifying at least 1 network member to whom they could
disclose. Family members and friends were central to the selective disclosure process, comprising the
greatest proportion of network members who, both were and were not identified as preferred
confidants. Women were found to show higher odds of preference for mental illness disclosure than
men. Having lower perceived social support was associated with lower odds of disclosure preference.
Among relational factors, greater relationship availability and lower dyadic tangible social support
were associated with lower odds of disclosure preference. Practice and research implications of using
social network analysis to get a deeper understanding of disclosure and disclosure preference are
discussed, including implications for future interventions targeting stigma reduction.

Public Policy Relevance Statement

The current study supports the use of a social network approach to understand the mental
illness disclosure process. The results suggest that disclosure patterns are affected by an
individual’s characteristics (i.e., gender, perceived social support) as well as relationship
characteristics with their social network members (i.e., kind of relationship, duration of
relationship, and availability of support).

S
erious mental illness (SMI) is a leading cause of disability mental illness has been known to decrease the likelihood of dis-
worldwide (Wiley-Exley, 2007). Of the 235 million peo- closure (Carpenter & Greene, 2013; Link, 1982), even though it
ple comprising the U.S. adult population (U.S. Census has been linked to a host of positive intrapsychic and interpersonal
Bureau, 2011), 13.6 million are living with a SMI and facing outcomes, such as freeing up cognitive and affective resources
pervasive stigma (Schomerus et al., 2012). Stigma associated with along with helping to elicit social support (Garcia & Crocker,
2008; Reavley & Jorm, 2014; Rüsch, Brohan, Gabbidon, Thorni-
croft, & Clement, 2014) and develop a supportive social network
This article was published Online First April 10, 2017. (Bos, Kanner, Muris, Janssen, & Mayer, 2009; Corrigan & Lundin,
Rohini Pahwa, Silver School of Social Work, New York University; 2001). In a workplace setting, mental illness disclosure may facil-
Anthony Fulginiti, Graduate School of Social Work, University of Denver; itate an individual’s access to special adjustments or accommoda-
John S. Brekke, Department of Adults and Healthy Aging, Suzanne tions (Brohan et al., 2014).
Dworak-Peck School of Social Work, University of Southern California;
Existing research among those with SMI has generally sought to
Eric Rice, Department of Children, Youth and Families, Suzanne Dworak-
Peck School of Social Work, University of Southern California.
understand disclosure in the context of help-seeking and occupational
Correspondence concerning this article should be addressed to Rohini settings (Chen, Lai, & Yang, 2013) as well as an individual level
Pahwa, Silver School of Social Work, New York University, 1 Washington phenomenon where an individual with a “concealable” (Rüsch et al.,
Square North, New York, NY 10003. E-mail: [email protected] 2014) stigmatized identity either chooses to disclose or not (Greene,

575
 576 PAHWA, FULGINITI, BREKKE, AND RICE

2009). Disclosure behavior has been linked with a host of individual A more comprehensive understanding of disclosure prefer-
level factors, like gender (Brohan et al., 2012; Derlega & Chaikin, ence must include an investigation into both individual and
1976), psychiatric symptomatology (Ilic et al., 2014), and social relational level factors associated with whether individuals
support (Garcia & Crocker, 2008). The literature has not given due choose to disclose. The current article takes the dialogue further
credence to the fact that disclosure is inherently a social activity, by using social network analysis to deepen our understanding of
whereby an individual prefers to disclose to another individual within selective disclosure and by using multilevel modeling to simul-
the context of social relationships (Rice, Comulada, Green, Arnold, & taneously examine various individual and relational level de-
Rotheram-Borus, 2009) and involves active decisions about disclo- terminants of mental health disclosure. Enhancing our under-
sure. Greene, Derlega, and Mathews (2006) have talked about disclo- standing of disclosure is critical to facilitate help-seeking and
sure experiences as transactions between the “discloser and the dis- inform mental health service delivery efforts among individuals
closure targets” (Greene et al., 2006, p. 416) and has called for a with SMI.
deeper investigation into various relational variables in theoretical The specific aims of the study are: (a) To descriptively explore
understanding of the process of decision making about disclosure the phenomenon of selective disclosure of individuals with SMI by
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

(Greene et al., 2006). Support for a deeper investigation of the process understanding the patterns of disclosure and non-disclosure at the
This document is copyrighted by the American Psychological Association or one of its allied publishers.

of disclosure decision-making can also be found in the existing relational level. (b) To identify and examine the relative influence
literature on the phenomenon of selective disclosure, which involves of individual-level and relational-level factors associated with
a discriminant approach to what and with whom one discloses about mental illness disclosure preference.
their mental illness (Corrigan, Watson, & Barr, 2006). Recent evi-
dence indicates that selective disclosure serves as an adaptive identity
management strategy (Ilic et al., 2014), which is optimal for enhanc-
Theoretical Framework for Mental Illness
ing social support while minimizing stigma (Bos et al., 2009). How-
Disclosure Decision-Making
ever, little is known about the characteristics of and relationships with Despite the benefits associated with mental illness disclosure, most
candidates that increase the likelihood of disclosing to them. This of the recent literature has had a limited focus on disclosure decision-
underscores both a gap in the present literature about disclosure and making process. A comprehensive conceptual framework is important
the need to better understand the decision-making process behind for a deeper understanding of individual and relational factors that
selecting a potential candidate for disclosure. contribute to the process of disclosure. A conceptual framework
Although a key determinant of disclosure preference that has would also provide a potential roadmap for addressing factors to be
emerged in the literature centers on the suitability of prospective targeted by a potential disclosure intervention. We have tailored the
confidants (i.e., to “whom” to disclose), there is a need to move disclosure decision-making model (DD-MM) by Greene (2009) and
beyond the broad categories of the types of relationship and also applied it to individuals with severe mental illness (refer to Figure 1).
study other key network level variables that might play an impor- This disclosure DD-MM has been previously applied to various
tant role in the disclosure process, such as network sources of concealable stigmatizing identities, like HIV (Greene, Derlega, Yep,
social support, greater accessibility, or quality of the relationships & Petronio, 2003), and has also been applied to suggest a stigma
(Fulginiti, Pahwa, Frey, Rice, & Brekke, 2015; Granovetter, 1973; intervention for mental health research by Carpenter and Greene
Greene et al., 2006). For instance, social support has been dis- (2013). An advantage of the DD-MM is that it was developed for
cussed within the context of social networks (Cohen & Wills, application across motivational contexts (Greene et al., 2012), which
1985) and has been linked to the disclosure decisions about con- is important given that people with serious mental illness can be
cealable stigmatized statuses (e.g., Fulginiti et al., 2015) but the motivated to disclose for many voluntary (e.g., seeking community
extent to which the perceived availability or receipt of support support, giving hope to peers, and fighting stigma) and involuntary
affect mental illness disclosure preference within dyadic relation- (e.g., “forced disclosure” because of symptom severity) reasons (Bril-
ships is unknown. Barniv, Moran, Naaman, Roe, & Karnieli-Miller, 2016); a broader

Assessment of Information
(individual level factors)
Age, gender, marital status,
symptom levels, psychosocial
functioning, stigma, perceived social
support
Disclosure Preference

Assessment of Receiver (relational

level factors)
Relationship type, relationship
duration, relationship availability,
relationship quality and relational

Figure 1. Disclosure Decision Making model on disclosure preference. See the online article for the color
version of this figure.
 MENTAL ILLNESS DISCLOSURE 577

discussion of voluntary versus involuntary disclosure occurs below. lar disorder, or major depressive disorder; and should have used
The model, which was conceptualized for understanding the likeli- outpatient mental health services at the site for at least 2 months. The
hood of health disclosure, suggests that individuals make multiple exclusion criteria included having a diagnosis of mental retardation,
assessments as part of their disclosure making process. an identifiable neurological disorder, or a primary diagnosis of drug or
At one level, individuals perform an assessment of information alcohol abuse or dependence in the previous 6 months. Potential
that include factors about the discloser and other information that participants were identified by their service providers on the basis of
might impact disclosure decisions. For an individual with SMI, the inclusion and exclusion criteria and referred to the research team
this level would include individual factors like stigma, symptoms, when they expressed interested in participating in the study. After
psychosocial functioning, perception of social support and other providing consent, participants completed a battery of self-report
demographic factors like age, gender, and marital status. At an- questionnaires (i.e., individual-level information) and a social network
other level, individuals perform an assessment of potential receiv- interview (i.e., relational-level information) in a face-to-face inter-
ers that include different aspects of people around them as candi- view. Institutional Review Boards at University of Southern Califor-
dates for disclosure (Greene et al., 2012). For an individual with nia and Los Angeles County Department of Mental Health approved
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

SMI, this level would include relational levels factors like rela- the study.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

tionship type, quality, length, availability, and support. Assess-

ment of factors on the individual and relational level (or assess-
ment of information and potential receiver in the DD-MM) would Measures
lead to decisions about preference for disclosure and candidates for
disclosure. Social network interview. The social network data was
For perspective, it is important to consider that disclosure can be gathered using a semistructured interview using a protocol that guided
voluntary or involuntary. Consistent with more contemporary the collection of personal network data from each participant. The
communication scholarship (e.g., Greene et al., 2006), we use the participants were first informed that the purpose of the interview was
term disclosure to mean the intentional sharing of information (i.e., to collect information about their social networks over the previous 2
voluntary disclosure) but the reality is that information can be months. Interviewers then used a “name generator” method that
shared in unintentional ways (i.e., involuntary disclosure). For included a series of prompts, which were read aloud, to elicit a list of
example, the notion that disease progression can signal the pres- the participant’s social network members using the following set of
ence of illness has often been discussed (e.g., Stutterheim et al., statements: “Think about the last 2 months. I want to know about the
2011). Psychiatric symptoms can become so severe that outside people you have had contact with, whom you talk to, hang out with,
observers suspect mental illness from behavioral signs without any talk on the phone with, or email, text, or talk with on the Internet.” To
verbal admission of it (e.g., Bril-Barniv et al., 2016). Furthermore, promote recall of the complete network rather than a biased subset of
situational or contextual circumstances can compel disclosure, network members (e.g., family members), the series of prompts
such as when someone is essentially mandated to involve collateral following the name generator consisted of reading a list of different
support during an emergency room visit to avoid psychiatric relationship types to the participant, including the following: “friends;
hospitalization or during a hospitalization to expedite discharge. family; case workers; people you know from the neighborhood;
Lastly, in what can be described as an aspect of “third-party girlfriend/boyfriend; husband/wife; people you know from agencies;
disclosure” (Greene et al., 2003), a person may voluntarily disclose friends you knew before you started using mental health agencies or
information about their mental illness to a family member who may have moved away.” Participants were subsequently asked a
then shares that information with others without permission from series of questions about the attributes of each social network member
the diagnosed person (e.g., Chen et al., 2013). Although our and their relationships with each social network member. Further
current study seeks to clarify the multilevel influences of voluntary information about the social network interview method is available in
disclosure, there is no assumption of equivalence to involuntary Rice (2010); Rice et al. (2009). This method yielded data on relational
disclosure in regards to causes or consequences. level factors described below.

Method
Relational-Level Factors
Participants and Procedure
The social network interview yielded information on a multitude
The study sample consisted of 60 ethnically diverse individuals of relational-level factors, including: (a) disclosure preference, (b)
with SMI treated in a publically funded mental health clinic in Los relationship type, (c) relationship duration, (d) relationship quality,
Angeles County in 2012. Three trained interviewers holding Masters- (e) relationship availability, and (f) social support.
level clinical degrees conducted the interviews. A senior social net-
work researcher provided training and supervision to the interviewers Disclosure preference. Mental illness disclosure prefer-
to promote consistency in the administration of the network interview ence within participant social networks was assessed using a
protocol across research team members following the technique out- question that asked about disclosure preference (Whom would you
lined in Rice (2010). Participants were outpatients recruited from a not wish to tell that you have a mental illness?). To account for
community-based outpatient mental health clinic in the United States. individuals in their network who already knew about their mental
To be eligible for participation, the individuals had to be between 18 illness, the following question was also asked: From the people
and 65 years of age; have a diagnosis of schizophrenia spectrum who already know about your mental illness, who do you wish
disorder (schizophrenia, schizoaffective disorder, or psychosis), bipo- didn’t know that you have a mental illness? The list of all persons
 578 PAHWA, FULGINITI, BREKKE, AND RICE

in a participant’s network (potential disclosure candidates) served levels of functioning (e.g., How often do you meet with friends and
as the set of possible response options, requiring participants to acquaintances?). The current study used the total outcome score
indicate whether or not they would disclose to each network calculated by summing the results of the four items (Strauss &
member (i.e., relational-level disclosure measure). This factor was Carpenter, 1974). The internal reliability for the measure was .73
used as the dependent variable in the analysis. Each network in the current study.
member who was identified by the participant as someone to
whom they would not wish to disclose about their mental illness or Stigma. The Internalized Stigma of Mental Illness scale
wished that the network member did not already know about their (ISMI; Ritsher & Phelan, 2004; Ritsher-Boyd, Otilingam, & Gra-
mental illness was coded as 0. All other network members were jales, 2003) was used to measure internalized stigma. Each item
coded as 1, indicating that the participant would not mind disclos- was measured on a 4-point Likert scale ranging from “strongly
ing to them. disagree” to “strongly agree”, with higher scores indicating greater
stigma (e.g., I am embarrassed or ashamed that I have a mental
Relationship type. For type of relationship, participants illness). A single factor, consisting of the four subscales of the
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

were asked, “Who is this person/what is their relationship to original ISMI (Alienation, Stereotype Endorsement, Discrimina-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

you?”; this variable was created to consist of four mutually exclu- tion Experience and Social Withdrawal) was used, as per the
sive categories, including family, service providers (including case recommendation of the authors on further factor analytical find-
workers, psychiatrists, and peer-providers), friends, and acquain- ings (Ritsher & Phelan, 2004; Ritsher Boyd et al., 2003; Stevelink,
tances (including neighbors, colleagues, and other acquaintance). Wu, Voorend, & van Brakel, 2012). This single factor yielded an
␣ of 0.87 in the current study. The scale is frequently used in
Relationship duration. Relationship duration was mea- research on individuals with SMI, with evidence supporting high
sured in years with an item that asked, “How long have you known internal and test–retest reliability as well as construct validity
this person?” Relationship quality was measured with an item (Ritsher & Phelan, 2004; Ritsher Boyd et al., 2003).
using a 5-point scale that asked “How close are you to this
person?” Higher scores indicated greater closeness. Perceived social support. The 4-item short version of
the Medical Outcomes Study Social Support survey was used to
Relationship availability. Relationship availability was measure perceived availability of social support (MOSSS; Sher-
measured with an item using a 4-point scale that asked, “How bourne & Stewart, 1991). Items are rated on a 5-point Likert scale,
frequently do you have a contact with this person?” Higher scores ranging from none of the time to all of the time, with higher ratings
indicated more frequent contact. indicating greater support (e.g., How often is each of the following
kinds of support (e.g., someone to do something enjoyable with)
Relational level social support. Social support was available to you?). The internal reliability for the measure was .75
measured on four dimensions (tangible, informational/emotional, in the current study. The scale has been commonly used in research
positive social interactional, and affectionate social support) with similar populations (Pahwa, Smith, McCullagh, Hoe, &
adapted from the 4-item short version of the Medical Outcomes Brekke, 2016; Robitaille, Orpana, & McIntosh, 2011).
Study Social Support survey (Sherbourne & Stewart, 1991) that
were each measured using dichotomous items that asked the par- Demographic variables. Information pertaining to de-
ticipant whether or not they would seek that type of support from mographic factors, including age, gender, and ethnicity was ob-
each member of their social network. tained using a self-report survey.

Data Analysis
Individual-Level Factors
Mental illness disclosure preference, a social network level variable
Psychiatric symptom severity. Psychiatric symptoms that represented whether or not a network member in an individual’s
were assessed using the 14-item Colorado Symptom Index network was a preferred candidate for disclosure, was regressed on
(CSI; Shern, Lee, & Coen, 1996). CSI is a self-report measure individual-level factors (i.e., measures that vary across participants)
assessing psychological or emotional difficulty in the past 30 and relational-level factors (i.e., measures that vary across network
days. Items are rated on a 5-point Likert scale, ranging from not members). Two-level multilevel logit models were used, which al-
at all to at least every day, with higher ratings representing lowed for the simultaneous modeling of disclosure preference across
more difficulty (e.g., How often have you felt depressed?). The participants (i.e., interindividual variation) as well as disclosure pref-
internal reliability (Cronbach’s ␣) for the measure was .90 in erence across network members within participant social networks
the current study. (i.e., intraindividual variation). Because the disclosure practices of a
participant to their social network members are likely to be highly
Psychosocial functioning. The Strauss and Carpenter correlated, the introduction of a random intercept was indicated.
Functional Outcomes scale was used to measure functional out- Relational-level factors constituted Level 1 covariates and individual-
comes (Strauss & Carpenter, 1972, 1974). The scale contains four level factors constituted Level 2 covariates. The multilevel analysis
discrete dimensions (duration of noninstitutionalization, social proceeded in two steps. First, a series of univariable logistic regression
contacts, useful employment, and symptoms), each of which is analyses were used to examine associations between disclosure and
rated using a single item. Items are rated on a 5-point scale, with each individual-level and relational-level factor. Second, factors that
variable response anchors whose higher scores indicate higher were either found to be significantly associated with disclosure in the
 MENTAL ILLNESS DISCLOSURE 579

univariable logistic regression analysis or were variables of interest or variable logistic regression models examined the associations be-
known to be theoretically linked to disclosure were entered into a tween each individual-level and relational-level factor and disclosure
multivariable multilevel logistic model. Data were analyzed using the preference. Table 3 describes the results from the univariable logistic
xtmelogit command in Stata version 13 (StataCorp, 2013) and the regression analyses. Among the individual-level factors, gender, in-
models were fit using maximum likelihood estimation (MLE). Odds ternalized stigma, and perceived social support were found to be
ratios (OR) and associated 95% confidence intervals (CIs) were used associated with disclosure preference. Specifically, women (OR ⫽
to determine significance. 9.577, p ⬍ .01) and people with higher levels of perceived social
Although there is minimal guidance for making decisions about support (OR ⫽ 1.308, p ⬍ .05) exhibited higher odds of preference
requisite sample sizes for each level of a multilevel model to for disclosure. Individuals with higher stigma levels exhibited lower
maximize power (Scherbaum & Ferreter, 2008), Maas and Hox odds of preference for disclosure (OR ⫽ .880, p ⬍ .10). Among the
(2005) found that a sample size exceeding 50 at the highest level relational-level factors, relationship type, tangible social support, and
of the model produces unbiased and accurate estimates of regres- emotional-informational social support were found to be associated
sion coefficients, variance components, and SEs (Maas & Hox, with disclosure preference. Specifically, individuals showed higher
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

2005). Given that the current study uses a two-level model, with odds of disclosure preference toward service providers (OR ⫽ 18.584,
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Level 2 (i.e., the highest level) representing the participants, our p ⬍ .01), people who provided tangible social support (OR ⫽ 7.947,
sample size (n ⫽ 60) is adequate. Of note, the sample size at Level p ⬍ .05), and people who provided emotional-informational social
1, representing the number of social network ties, is 866. Further support (OR ⫽ 2.674, p ⬍ .05).
examples of multilevel modeling with network data can be found Table 3 also presents the findings from the multivariable logistic
in works by Kennedy et al. (2010); Snijders (1996). regression model that includes variables found to be significantly
associated with disclosure preference in the univariable logistic
regression analysis or were variables of interest or known to be
Results theoretically linked to disclosure. Among the individual-level fac-
In terms of the total sample, 53% of participants were female and tors, gender, internalized stigma, and perceived social support
over 66% were racial/ethnic minorities. The average age of the were found to be associated with mental illness disclosure prefer-
participants was 47 years (SD ⫽ 9.31). Thirty-five percent were ence in the multilevel analysis. The odds of disclosure preference
diagnosed with schizophrenia, schizoaffective disorder or psychosis, (or choosing a network member as a potential disclosure candi-
33.3% with a mood disorder, and 31.7% were diagnosed with major date) for women were found to be six times the odds of disclosure
depression. In terms of relational tie characteristics, family members preference for men (OR ⫽ 6.502, p ⬍ .05). Every unit increase in
consisted of the majority of their social networks (45%), followed by internalized stigma led to an 11% decrease in the odds of disclo-
friends (24%) and, mental health providers, including peer providers sure preference (OR ⫽ .891, p ⬍ .10). Every unit increase in
(17%). Participants knew their network members for an average of 18 perceived social support led to a 26% increase in the odds of
years and knew an average of 36% of their network members since disclosure preference (OR ⫽ 1.258, p ⬍ .10). Among the
before getting a diagnosis of a mental illness. Approximately 31% of relational-level factors, relationship type, and tangible social sup-
their networks provided them with affectionate social support, 26% port were found to be significantly associated with disclosure
with positive social interaction, 20% with emotional support and 11% preference. Specifically, as compared with family, service provid-
provided tangible support. Table 1 provides descriptive statistics on ers were 29 times more likely to be chosen as preferred candidates
the individual characteristics of the participants along with the rela- for disclosure (OR ⫽ 29.004, p ⬍ .01) and individuals showed
tional characteristics between each participant and their social net- eight times higher odds of disclosure preference with the network
work members. members who were identified as sources of tangible social support
The first aim of the study was to explore the phenomenon of (OR ⫽ 8.932, p ⬍ .05).
selective disclosure by understanding the disclosure patterns at
individual and relational levels. On an individual level, 68% of the
participants had disclosed their mental illness to at least one
Discussion
confidant. More than 13% of the sample had disclosed to more The risks associated with mental illness disclosure for an indi-
than three members of their network. Overall, with the total vidual living with SMI complicate the decision-making about
number of ties being 866, less than 10% of the social network disclosure and preferences about disclosure candidates. The idea of
members across all participant social networks had not been told disclosure is increasingly being linked to deliberate choices about
about the mental illness. This means that over 90% of the social disclosure and preferences about potential disclosure candidates,
network members named by participants were confidants for dis- especially since decisions about disclosure can have significant
closure. Of those who were not the chosen confidants for disclo- positive and negative repercussions (Greene et al., 2012). The
sure, 45% were family members, 27% friends, 24% acquaintances current study simultaneously used individual and relational level
and just 2% service providers. On the other hand, of those who data to provide an understanding of the factors associated with
were disclosed to, 45% were family members, 23% were friends, mental illness disclosure preferences for individuals with severe
18% were service providers and only 13% were acquaintances. mental illness using the DD-MM by Greene (2009). Before dis-
Table 2 provides the patterns of selective disclosure for individuals cussing our first aim of exploring the phenomenon of selective
with SMI. disclosure, it is important to highlight that the current sample had
The second aim of the study was to identify and examine the very high rates of disclosure. Sixty-eight percent of the participants
relative influence of individual-level and relational-level factors asso- had disclosed their mental illness to at least one confidant. More
ciated with mental illness disclosure preference. The series of uni- than 13% of the sample had disclosed to more than three members
 580 PAHWA, FULGINITI, BREKKE, AND RICE

Table 1. Demographic Characteristics of the Participants in the Sample (n ⫽ 60, Level 2) and
the Relational Ties Between Participants and Their Social Network Members (n ⫽ 866, Level 1)

Factor Mean (SD) n (%)

Individual-level factors
Age (in years) 47.40 (9.31)
Gender
Male 28 (46.70)
Female 32 (53.30)
Ethnicity
Non-Hispanic Whites 20 (33.30)
Ethnic minorities (including Latinos, Blacks, and Asians) 40 (66.70)
Relationship status
Married/living with partner 7 (11.70)
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Not married/living with a partner 53 (88.30)

This document is copyrighted by the American Psychological Association or one of its allied publishers.

Employment status
Employed 4 (6.70)
Not employed 56 (93.30)
Diagnosis
Schizophrenia spectrum disorder (including
schizophrenia, schizoaffective disorder, and psychosis) 21 (35.00)
Bipolar disorder 20 (33.30)
Major depression 19 (31.70)
Psychosocial factors
Symptomatologya 33.22 (11.72)
Psychosocial functioningb 10.22 (2.01)
Internalized stigmac 17.67 (7.89)
Social supportd 13.53 (4.06)
Relational-level factors
Relationship type
Family 394 (45.49)
Friend 206 (23.79)
Service provider 145 (16.74)
Others 121 (13.97)
Relationship duration 16.21 (17.82)
Relationship quality 3.59 (1.30)
Relationship availability (3 or more contacts in past month) 668 (77.10
Prediagnosis affiliation 313 (36.10)
Social support
Tangible support 96 (11.10)
Emotional-informational 174 (20.10)
Positive social interactional 222 (25.60)
Affectionate support 265 (30.60)
From mental health community 235 (27.10)
a
Measured by Colorado Symptom Inventory (Shern et al., 1996). b Measured by Straus and Carpenter
Functional Outcome Scale (Strauss & Carpenter, 1972). c Measured by Internalized Stigma of Mental Illness
scale (Ritsher Boyd et al., 2003). d Measured by the 4-item short version of the Medical Outcomes Study Social
Support Survey (MOSSS; Sherbourne & Stewart, 1991).

of their network. This finding could be attributed to the clinical The current study yielded some interesting findings about se-
nature of the sample used in the study. Because the participants lective disclosure. While family members were found to be the
were currently seeking services for their mental illness, it would primary confidants about an individual’s mental illness (with 46%
directly impact their rates of disclosure. This phenomenon could of total individual’s network members disclosed to being family
also be attributed to the free recall name generator used in the members), they were also found to be the primary candidates of
current study, whereby, participants respond to a prompt that non-disclosure (with 45% of the total network members not dis-
defines some criteria (Rice et al., 2014), which in turn enables the closed to also being family members). This finding underlines the
nomination of both strong and weak ties (Brewer, 2000; McCarty ambivalence experienced by individuals with SMI when making
& Govindaramanujam, 2005). This finding speaks to the relevance decisions about mental health disclosure. While on one hand,
and applicability of social network methodology to gather in depth individuals do not want to talk about their mental illness with
information about individuals with SMI to enable more nuanced important network members (like family) because of a fear of
and effective service delivery. burdening them (Brohan et al., 2012; Chen et al., 2013) or antic-
 MENTAL ILLNESS DISCLOSURE 581

Table 2. Selective Disclosure Patterns for Individuals With Serious Mental Illness (SMI; N of
Network Members ⫽ 884)

Disclosure/non-disclosure N (%) N (%)

Prevalence of disclosure and non-disclosure at individual level (n ⫽ 60)

a

Disclosersa 41 (68.30)
Non-disclosers 19 (31.70)

Prevalence of non-disclosure at relational level (n ⫽ 866)

Network members not disclosed to 75 (8.66)
Network members disclosed to 791 (91.34)

Patterns of selective disclosure at relational level

Relationship type Disclosure patterns (n ⫽ 791) Non-disclosure patterns (n ⫽ 75)
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Family 360 (45.50) 34 (45.30)

Friend 186 (23.50) 20 (26.70)
Mental health provider 143 (18.00) 2 (2.70)
Acquaintances 102 (12.90) 19 (24.00)

Note. Individual-level disclosure prevalence is a metric for disclosure at Level 2, providing information about
the proportion of participants who have identified social network members to whom they have and have not
disclosed their mental illness. Relational-level disclosure prevalence is a metric for disclosure at Level 1,
providing information about the proportion of all social network members who have been confidants or
candidates for disclosure. The differences in sample sizes in the table are a function of differences in sample sizes
at Level 1 (n ⫽ 866) and Level 2 (n ⫽ 60).

ipation of stigma (Quinn et al., 2014), family members are also the that worsening psychiatric symptomatology may increase the like-
most preferred disclosure candidates because of an increased like- lihood of mental illness disclosure (Greene et al., 2012; Holmes &
lihood of obtaining support and resources (Bos et al., 2009). This River, 1998), especially to close family members and friends, as
also speaks to the importance of quality versus types of social well as the link between psychological distress and general self-
network relationships, wherein close family and friendship ties disclosure (e.g., Fever Model; Stiles, 1987). However, the possi-
might increase the likelihood of disclosure while weak family and bility exists that different types of symptoms may affect disclosure
friendship ties might decrease the likelihood of disclosure. decisions differently (Fulginiti et al., 2015). For instance, even
For our second aim, the current study used social network though anxiety (Stiles, Shuster, & Harrigan, 1992), as well as
methodology and multilevel modeling to provide a comprehensive schizophrenia (Brohan et al., 2012) have been linked to higher
examination of mental illness disclosure preference in a multieth- disclosure, there is no guarantee that other common symptoms of
nic sample of individuals with SMI using an adaptation of the serious mental illness (e.g., severe depression, mood disorders)
DD-MM by Greene (2009). Distinct from prior research, which will similarly influence decisions about disclosure. If different
has typically shed light on either individual or relational factors symptom types indeed affect disclosure in different ways then
and focused on whether an individual discloses of not, the current using a global measure of symptom severity could result in the
study simultaneously examined the relative influence of both types failure to detect an effect. Relatedly, although our findings did not
of factors while studying mental illness disclosure preference as a show that functional level impacted disclosure preference, the use
precursor to disclosure. Consistent with work on depression dis- of a global measure limited the ability to examine the impact of
closure (Garcia & Crocker, 2008) and meta-analytic findings on domain-specific functioning on disclosure preference. For exam-
HIV disclosure (Smith, Rossetto, & Peterson, 2008), social support ple, one might expect that social functioning may be more relevant
was found to be significantly associated with disclosure preference to disclosure preference than functioning in activities of daily
at the individual level, implying that whether or not a network living. Additionally, the current study only provides a preliminary
member is chosen as a candidate for disclosure is related to an perspective on the potential relationship between functional level
individual’s general perception of social support. Also consistent and mental illness disclosure preference and this relationship
with prominent work in the area of SMI (Corrigan & Rao, 2012; would benefit from further exploration, particularly given the
Garcia & Crocker, 2008) and other stigmatized conditions (Smith centrality of functioning to rehabilitation efforts.
et al., 2008), greater internalized stigma was also found to be In terms of the relationships between demographic factors and
associated with disclosure preference. The findings contribute to mental illness disclosure preference, results from the multilevel
the existing work in the literature on the importance of various modeling showed gender to be the only factor associated with
psychosocial factors, like social support and stigma, in facilitating disclosure preferences. Consistent with the majority of prior work
decisions about disclosure. In terms of clinical factors, no relation- on general self-disclosure, women were more willing to choose a
ship was observed between psychiatric symptom severity or func- network member as a potential disclosure candidate than men
tional level and mental illness disclosure preference. The null (Derlega & Chaikin, 1976). This finding is consistent with litera-
finding pertaining to symptoms is surprising given the proposition ture that links mental illness self-disclosure to perceived gender
 582 PAHWA, FULGINITI, BREKKE, AND RICE

Table 3. Univariable and Multivariable Random-Intercept Logistic Regression of Disclosure Preference About
Mental Illness (N ⫽ 60)

Univariablea Multivariable
Factor OR 95% CI OR 95% CI

Individual-level factors
Females 9.577ⴱⴱⴱ [1.816–50.491] 6.502ⴱⴱ [.960–44.05]
Age 1.008 [.919–1.106] — —
Minorities .487 [.706–3.361] 1.178 [.153–9.621]
Married/partner .197 [.018–2.100] — —
Symptomsb 1.020 [.914–1.052] 1.035 [.949–1.129]
Psychosocial functioningc .963 [.950–1.094] .800 [.500–1.281]
Internalized stigmad .880ⴱⴱ [.782–.989] .891ⴱ [.779–1.019]
Perceived social supporte 1.308ⴱⴱ [1.041–1.643] 1.258ⴱ [.966–1.638]
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Relational-level factors
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Relationship type (ref group ⫽ family)

Friend 1.058 [.485–2.303] 1.731 [.681–4.404]
Service provider 18.548ⴱⴱⴱ [3.480–98.844] 29.004ⴱⴱⴱ [4.801–174.912]
Others .334ⴱⴱ [.461–.766] .754 [.289–1.964]
Relationship duration .996 [.978–1.015] — —
Relationship quality 1.233 [.940–1.617] — —
Relationship availability .918 [.671–1.258] — —
Social support
Tangible support (help with daily chores) 7.947ⴱⴱ [1.522–41.490] 8.932ⴱⴱ [1.465–54.441]
Emotional-informational (share personal problems) 2.674ⴱⴱ [1.010–7.102] 1.964 [.616–6.260]
Positive social interaction (do something enjoyable with) .916 [.399–2.102] — —
Affectionate support (make you feel loved and wanted) 1.461 [.648–3.292] — —

Note. OR ⫽ odds ratio; CI ⫽ confidence interval.

a
Dependent variable: measured using a question asked during the network interview on disclosure about mental illness. b Measured by Colorado
Symptom Inventory (Shern et al., 1996). c Measured by Straus and Carpenter Functional Outcome scale (Strauss & Carpenter, 1972). d Measured by
Internalized Stigma of Mental Illness scale (Ritsher Boyd et al., 2003). e Measured by the 4-item short version of the Medical Outcomes Study Social
Support Survey (MOSSS; Sherbourne & Stewart, 1991).
ⴱ
p ⬍ .10. ⴱⴱ p ⬍ .05. ⴱⴱⴱ p ⬍ .01.

norms whereby it is easier for women to disclose their mental tween disclosure preference and social support at multiple levels;
illness (Derlega & Chaikin, 1976) and contradicts other literature at individual level as perception of social support and at relation-
whereby, women might have a harder time disclosing depending ship level by asking individuals to identify all the relationships that
on the context of disclosure (Brohan et al., 2012). This suggests the were a source of tangible or emotional/informational social sup-
need for further investigation into the mental illness disclosure port. The significant relationship between disclosure preference
practices of men and a need for tailoring disclosure and stigma and social support at both individual and relationship levels, which
related interventions specially geared toward men. is consistent with the literature (Garcia & Crocker, 2008; Smith et
The present study also expands our understanding of things al., 2008), further speaks to the importance of social support in the
about relationships that affect mental illness disclosure preference disclosure decision-making process and a need to use more so-
and the process of selective disclosure. In the current study, an phisticated methodologies, like social network analysis and mul-
individual’s disclosure preference was significantly associated tilevel modeling, to understand the importance of complex phe-
with several relational variables. Compared with family, individ- nomenon like social support in the lives of individuals with SMI.
uals showed higher preference for disclosure with caseworkers. Our study supports the previous finding in the literature that
This finding is consistent with literature suggesting that disclosing individuals make decisions regarding disclosure on the basis of
to caseworker and other service providers is paramount to a assessments made at multiple levels (Greene et al., 2012), empha-
successful client-provider relationship (Seeman, 2013). This find- sizing that we need to better understand these deeply personal
ing might also reflect the fact that communications between a decision-making processes before embarking on any potential in-
consumer and a service provider are expected to be confidential, tervention to increase the likelihood of disclosure. This also has
which could potentially increase the likelihood of disclosure. Ad- implications for future interventions and clinical work. In light of
ditionally, tangible social support on the relational level was found previous evidence linking disclosure preference to internalized
to be significantly associated with mental illness disclosure pref- stigma (Hielscher & Waghorn, 2015; Pachankis, 2007) and the
erence, which indicates that network members who provide social literature supporting that interventions targeting mental health
support are more likely to be the candidates of mental illness disclosure are more effective than those targeting internalized
disclosure. stigma (Carpenter & Greene, 2013; Corrigan & Fong, 2014),
The relationship between disclosure preference and social sup- disclosure appears to be a particularly attractive intervention point
port deserves a special mention. We studied the relationship be- that carries implications for stigma reduction. This study also
 MENTAL ILLNESS DISCLOSURE 583

speaks to the importance of social support in the selective disclo- Brohan, E., Evans-Lacko, S., Henderson, C., Murray, J., Slade, M., &
sure process and for disclosure interventions, and provides further Thornicroft, G. (2014). Disclosure of a mental health problem in the
insight into the process by which individuals with SMI decide employment context: Qualitative study of beliefs and experiences. Ep-
whom to choose for their mental illness disclosure. idemiology and Psychiatric Sciences, 23, 289 –300. http://dx.doi.org/10
There are a few important limitations to the current study that .1017/S2045796013000310
need to be considered while interpreting the results. First, this Brohan, E., Henderson, C., Wheat, K., Malcolm, E., Clement, S., Barley,
E. A., . . . Thornicroft, G. (2012). Systematic review of beliefs, behav-
study used cross-sectional data, which make the results associa-
iours and influencing factors associated with disclosure of a mental
tional and not causal. However, inferences about causality are
health problem in the workplace. BMC Psychiatry, 12, 11–24. http://dx
often difficult to make in field research because of many uncon- .doi.org/10.1186/1471-244X-12-11
trolled background sources of variance. Second, our modest sam- Carpenter, A., & Greene, K. (2013). Designing a disclosure-focused stigma
ple size and use of purposive sampling mean that our findings intervention for mental health research. In M. H. Eaves (Ed.), Applica-
might not be generalizable to the entire population of individuals tions in health communication (pp. 15–30). Dubuque, IA: Kendall Hunt.
with SMI. To mitigate the effects of small sample size and pur- Chen, F. P., Lai, G. Y., & Yang, L. (2013). Mental illness disclosure in
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

posive sampling, we gathered the data at multiple levels and Chinese immigrant communities. Journal of Counseling Psychology, 60,
This document is copyrighted by the American Psychological Association or one of its allied publishers.

included data from individuals in both high and low intensity 379 –391. http://dx.doi.org/10.1037/a0032620
treatment modalities to increase the generalizability of the sample. Cohen, S., & Wills, T. A. (1985). Stress, social support, and the buffering
Most of the participants were either Non-Hispanic White or La- hypothesis. Psychological Bulletin, 98, 310 –357. http://dx.doi.org/10
tino. Hence, the findings might not be representative of other .1037/0033-2909.98.2.310
ethnic or cultural groups. Lastly, because the sample consisted of Corrigan, P. W., & Fong, M. W. M. (2014). Competing perspectives on
individuals actively engaged in services, our study does not ac- erasing the stigma of illness: What says the dodo bird? Social Science &
Medicine, 103, 110 –117. http://dx.doi.org/10.1016/j.socscimed.2013.05
count for individuals with severe mental illness whose lack of
.027
insight or disagreement about having a mental illness could lead to
Corrigan, P. W., & Lundin, R. (2001). Don’t call me nuts: Coping with the
less disclosure. Despite the limitations of the current study, this
stigma of mental illness. Chicago, IL: Recovery Press.
research provides a foundation for further exploration of disclosure Corrigan, P. W., & Rao, D. (2012). On the self-stigma of mental illness:
decisions and preferences in a variety of the ways. Stages, disclosure, and strategies for change. Canadian Journal of
Future studies should include diverse subgroups to understand Psychiatry, 57, 464 – 469.
cultural differences in disclosure about mental illness. Especially Corrigan, P. W., Watson, A. C., & Barr, L. (2006). The self-stigma of
since conceptualizations about mental illness is rooted in culture, mental illness: Implications for self-esteem and self-efficacy. Journal of
there needs to be further exploration of patterns of disclosure Social and Clinical Psychology, 25, 875– 884. http://dx.doi.org/10.1521/
preferences according to the discloser and their disclosure candi- jscp.2006.25.8.875
dates’ ethnic and cultural affiliations. Further studies would also be Derlega, V. J., & Chaikin, A. L. (1976). Norms affecting self-disclosure in
wise to include instrumentation that assesses different types of men and women. Journal of Consulting and Clinical Psychology, 44,
psychiatric symptoms or functioning, which could affect disclo- 376 –380. http://dx.doi.org/10.1037/0022-006X.44.3.376
sure preferences. Furthermore, although the current study provided Fulginiti, A., Pahwa, R., Frey, L. M., Rice, E., & Brekke, J. S. (2015).
some interesting and important insights into various individual and What factors influence the decision to share suicidal thoughts? A mul-
tilevel social network analysis of disclosure among individuals with
relational factors associated with mental health disclosure prefer-
serious mental illness. Suicide and Life-Threatening Behavior, 46, 398 –
ences, the concepts require a larger sample size using a longitu-
412. http://dx.doi.org/10.1111/sltb.12224
dinal methodology to establish causal relationships and stronger
Garcia, J. A., & Crocker, J. (2008). Reasons for disclosing depression
conclusions. Lastly, more work is needed to explore the conditions matter: The consequences of having egosystem and ecosystem goals.
and consequences of involuntary disclosure as well as how the Social Science & Medicine, 67, 453– 462. http://dx.doi.org/10.1016/j
nature of disclosure (voluntary vs. involuntary) and different dis- .socscimed.2008.03.016
closure motivations impact outcomes in the context of mental Granovetter, M. S. (1973). The strength of weak ties. American Journal of
illness. Sociology, 78, 1360 –1380. http://dx.doi.org/10.1086/225469
Greene, K. (2009). An integrated model of health disclosure decision-
Keywords: disclosure preference; mental illness; stigma; social
making. In T. D. Afifi & W. A. Afifi (Eds.), Uncertainty, information
network analysis; social support
management, and disclosure decisions: Theories and applications (pp.
226 –253). New York, NY: Routledge.
References Greene, K., Derlega, V. J., & Mathews, A. (2006). Self-disclosure in
Bos, A. E. R., Kanner, D., Muris, P., Janssen, B., & Mayer, B. (2009). personal relationships. In A. L. Vangelisti & D. Perlman (Eds.), The
Mental illness stigma and disclosure: Consequences of coming out of the Cambridge handbook of personal relationships (pp. 409 – 428). Cam-
closet. Issues in Mental Health Nursing, 30, 509 –513. http://dx.doi.org/ bridge, England: Cambridge University Press. http://dx.doi.org/10.1017/
10.1080/01612840802601382 CBO9780511606632.023
Brewer, D. D. (2000). Forgetting in the recall-based elicitation of personal Greene, K., Derlega, V. J., Yep, G. A., & Petronio, S. (2003). Privacy and
and social networks. Social Networks, 22, 29 – 43. http://dx.doi.org/10 disclosure of HIV in interpersonal relationships: A sourcebook for
.1016/S0378-8733(99)00017-9 researchers and practitioners. Mahwah, NJ: Erlbaum.
Bril-Barniv, S., Moran, G. S., Naaman, A., Roe, D., & Karnieli-Miller, O. Greene, K., Magsamen-Conrad, K., Venetis, M. K., Checton, M. G.,
(2016). A qualitative study examining experiences and dilemmas in Bagdasarov, Z., & Banerjee, S. C. (2012). Assessing health diagnosis
concealment and disclosure of people living with serious mental illness. disclosure decisions in relationships: Testing the disclosure decision-
Qualitative Health Research, 27, 573–583. http://dx.doi.org/10.1177/ making model. Health Communication, 27, 356 –368. http://dx.doi.org/
1049732316673581 10.1080/10410236.2011.586988
 584 PAHWA, FULGINITI, BREKKE, AND RICE

Hielscher, E., & Waghorn, G. (2015). Managing disclosure of personal lish and French versions of the Medical Outcomes Study social support
information: An opportunity to enhance supported employment. Psychi- scale. Health Reports, 22, 33– 40.
atric Rehabilitation Journal, 38, 306 –313. http://dx.doi.org/10.1037/ Rüsch, N., Brohan, E., Gabbidon, J., Thornicroft, G., & Clement, S. (2014).
prj0000127 Stigma and disclosing one’s mental illness to family and friends. Social
Holmes, E. P., & River, L. P. (1998). Individual strategies for coping with Psychiatry and Psychiatric Epidemiology, 49, 1157–1160. http://dx.doi
the stigma of severe mental illness. Cognitive and Behavioral Practice, .org/10.1007/s00127-014-0871-7
5, 231–239. http://dx.doi.org/10.1016/S1077-7229(98)80008-4 Scherbaum, C. A., & Ferreter, J. M. (2008). Estimating statistical power
Ilic, M., Reinecke, J., Bohner, G., Röttgers, H., Beblo, T., Driessen, M., . . . and required sample sizes for organizational research using multilevel
Corrigan, P. W. (2014). Managing a stigmatized identity-evidence from modeling. Organizational Research Methods, 12, 347–367. http://dx.doi
a longitudinal analysis about people with mental illness. Journal of .org/10.1177/1094428107308906
Applied Social Psychology, 44, 464 – 480. http://dx.doi.org/10.1111/jasp Schomerus, G., Schwahn, C., Holzinger, A., Corrigan, P. W., Grabe, H. J.,
.12239 Carta, M. G., & Angermeyer, M. C. (2012). Evolution of public attitudes
Kennedy, D. P., Wenzel, S. L., Tucker, J. S., Green, H. D., Jr., Golinelli, about mental illness: A systematic review and meta-analysis. Acta
D., Ryan, G. W., . . . Zhou, A. (2010). Unprotected sex of homeless Psychiatrica Scandinavica, 125, 440 – 452. http://dx.doi.org/10.1111/j
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

women living in Los Angeles county: An investigation of the multiple .1600-0447.2012.01826.x

levels of risk. AIDS and Behavior, 14, 960 –973. http://dx.doi.org/10 Seeman, M. V. (2013). When and how should I tell? Personal disclosure of
This document is copyrighted by the American Psychological Association or one of its allied publishers.

.1007/s10461-009-9621-3 a schizophrenia diagnosis in the context of intimate relationships. Psy-

Link, B. (1982). Mental patient status, work, and income: An examination chiatric Quarterly, 84, 93–102. http://dx.doi.org/10.1007/s11126-012-
of the effects of a psychiatric label. American Sociological Review, 47, 9230-6
202–215. http://dx.doi.org/10.2307/2094963 Sherbourne, C. D., & Stewart, A. L. (1991). The MOS social support
Maas, C. J. M., & Hox, J. J. (2005). Sufficient sample sizes for multilevel survey. Social Science & Medicine, 32, 705–714. http://dx.doi.org/10
modeling. Methodology, 1, 86 –92. http://dx.doi.org/10.1027/1614-2241 .1016/0277-9536(91)90150-B
.1.3.86 Shern, D., Lee, B., & Coen, A. (1996). The Colorado Symptom Inventory:
McCarty, C., & Govindaramanujam, S. (2005). A modified elicitation of A self-report measure for psychiatric symptoms. Tampa, FL: Louis de la
personal networks using dynamic visualization. Connections, 26, 61– 69. Parte Mental Health Institute.
Pachankis, J. E. (2007). The psychological implications of concealing a Smith, R., Rossetto, K., & Peterson, B. L. (2008). A meta-analysis of
stigma: A cognitive-affective-behavioral model. Psychological Bulletin, disclosure of one’s HIV-positive status, stigma and social support. AIDS
133, 328 –345. http://dx.doi.org/10.1037/0033-2909.133.2.328 Care, 20, 1266 –1275. http://dx.doi.org/10.1080/09540120801926977
Pahwa, R., Smith, M. E., McCullagh, C. A., Hoe, M., & Brekke, J. S. Snijders, T. (1996). Stochastic actor oriented models for network change.
(2016). Social support-centered versus symptom-centered models in The Journal of Mathematical Sociology, 21, 149 –172. http://dx.doi.org/
predicting functional outcomes for individuals with schizophrenia. Jour- 10.1080/0022250X.1996.9990178
nal of the Society for Social Work and Research, 7, 247–268. http://dx StataCorp. (2013). Stata Statistical Software: Release 13. College Station,
.doi.org/10.1086/686770 TX: StataCorp LP.
Quinn, D. M., Williams, M. K., Quintana, F., Gaskins, J. L., Overstreet, Stevelink, S., Wu, I., Voorend, C., & van Brakel, W. (2012). The psycho-
N. M., Pishori, A., . . . Chaudoir, S. R. (2014). Examining effects of metric assessment of internalized stigma instruments: A systematic
anticipated stigma, centrality, salience, internalization, and outness on review. Stigma Research and Action, 2, 100 –118. http://dx.doi.org/10
psychological distress for people with concealable stigmatized identi- .5463/SRA.v1i1.11
ties. PLoS ONE, 9, e96977. http://dx.doi.org/10.1371/journal.pone Stiles, W. B. (1987). I have to talk to somebody. In V. J. Derlaga & J. H.
.0096977 Berg (Eds.), Self-disclosure. Theory, research, and therapy (pp. 257–
Reavley, N. J., & Jorm, A. F. (2014). Willingness to disclose a mental 282). New York, NY: Plenum Press.
disorder and knowledge of disorders in others: Changes in Australia Stiles, W. B., Shuster, P. L., & Harrigan, J. A. (1992). Disclosure and
over 16 years. Australian and New Zealand Journal of Psychiatry, 48, anxiety: A test of the fever model. Journal of Personality and Social
162–168. http://dx.doi.org/10.1177/0004867413495317 Psychology, 63, 980 –988. http://dx.doi.org/10.1037/0022-3514.63.6
Rice, E. (2010). The positive role of social networks and social networking .980
technology in the condom-using behaviors of homeless young people. Strauss, J. S., & Carpenter, W. T., Jr. (1972). The prediction of outcome in
Public Health Reports, 125, 588 –595. schizophrenia. I. Characteristics of outcome. Archives of General Psy-
Rice, E., Comulada, S., Green, S., Arnold, E. M., & Rotheram-Borus, M. J. chiatry, 27, 739 –746. http://dx.doi.org/10.1001/archpsyc.1972
(2009). Differential disclosure across social network ties among women .01750300011002
living with HIV. AIDS and Behavior, 13, 1253–1261. http://dx.doi.org/ Strauss, J. S., & Carpenter, W. T., Jr. (1974). The prediction of outcome in
10.1007/s10461-009-9554-x schizophrenia. II. Relationships between predictor and outcome vari-
Rice, E., Holloway, I. W., Barman-Adhikari, A., Fuentes, D., Brown, ables: A report from the WHO international pilot study of schizophrenia.
C. H., & Palinkas, L. A. (2014). A mixed methods approach to netowk Archives of General Psychiatry, 31, 37– 42. http://dx.doi.org/10.1001/
data collection. Field Methods, 26, 252–268. http://dx.doi.org/10.1177/ archpsyc.1974.01760130021003
1525822X13518168 Stutterheim, S. E., Bos, A. E. R., Pryor, J. B., Brands, R., Liebregts, M., &
Ritsher Boyd, J., Otilingam, P. G., & Grajales, M. (2003). Internalized Schaalma, H. P. (2011). Psychological and social correlates of HIV status
stigma of mental illness: Psychometric properties of a new measure. disclosure: The significance of stigma visibility. AIDS Education and
Psychiatry Research, 121, 31– 49. http://dx.doi.org/10.1016/j.psychres Prevention, 23, 382–392. http://dx.doi.org/10.1521/aeap.2011.23.4.382
.2003.08.008 U.S. Census Bureau. (2011). Age and sex composition: 2010. Retrieved
Ritsher, J. B., & Phelan, J. C. (2004). Internalized stigma predicts erosion from http://www.census.gov/prod/cen2010/briefs/c2010br-03.pdf
of morale among psychiatric outpatients. Psychiatry Research, 129, Wiley-Exley, E. (2007). Evaluations of community mental health care in
257–265. http://dx.doi.org/10.1016/j.psychres.2004.08.003 low- and middle-income countries: A 10-year review of the literature.
Robitaille, A., Orpana, H., & McIntosh, C. N. (2011). Psychometric Social Science & Medicine, 64, 1231–1241. http://dx.doi.org/10.1016/j
properties, factorial structure, and measurement invariance of the Eng- .socscimed.2006.11.009