Journal of Affective Disorders 189 (2016) 3642

Contents lists available at ScienceDirect

Journal of Affective Disorders

journal homepage: [Link]/locate/jad

Research report

Depression is who I am: Mental illness identity, stigma and wellbeing

Tegan Cruwys n, Sathiavaani Gunaseelan
University of Queensland, Australia

art ic l e i nf o a b s t r a c t

Article history: Background: Previous research has found that in the face of discrimination, people tend to identify more
Received 4 June 2015 strongly with stigmatized groups. Social identication can, in turn, buffer wellbeing against the negative
Received in revised form consequences of discrimination. However, this rejection identication model has never been tested in the
23 August 2015
context of mental illness identity.
Accepted 5 September 2015
Available online 10 September 2015
Methods: A survey was conducted with 250 people with diagnosed depression or current symptoms of at
least moderate clinical severity.
Keywords: Results: Experiencing mental illness stigma was associated with poorer wellbeing. Furthermore, people
Mood disorders who had experienced such stigma were more likely to identify as a depressed person. Social identi-
Mental health
cation as depressed magnied, rather than buffered, the relationship between stigma and reduced
Social identity
wellbeing. This relationship was moderated by perceived social norms of the depressed group for en-
Discrimination
Normative inuence gaging in depressive thoughts and behaviors.
Health behavior Conclusions: These ndings suggest that mental illness stigma is a double-edged sword: as well as the
direct harms for wellbeing, by increasing identication with other mental illness sufferers, stigma might
expose sufferers to harmful social inuence processes.
Crown Copyright & 2015 Published by Elsevier B.V. All rights reserved.

1. Introduction psychology for over 30 years, albeit with social groups other than
people with mental illness. This research has revealed a variety of
People who experience compromised mental health are reg- strategies that are used to manage low status group memberships
ularly confronted with discrimination in the workplace, in (Tajfel and Turner, 1979). Chief among them is collective action,
healthcare settings, and in the community more broadly (Corrigan whereby members of a stigmatized group band together to chal-
et al., 2014; Link et al., 1997, 1999). Research has established that lenge the status quo. However, this social change strategy is only
the stigma of mental illness actually constitutes a stressor in its likely to emerge when people perceive the low status of their
own right, with the potential to compound mental illness and group to be illegitimate (Ellemers et al., 1993; Hansen and Sas-
reduce the likelihood of recovery (Rsch et al., 2014; Pascoe and senberg, 2011). When the low status of the group is seen to be
Smart Richman, 2009; Yanos et al., 2010). Although these harmful legitimate, people are more likely to attempt to leave the group
effects have been established, relatively little research has in- through individual mobility strategies, particularly to the extent
that the boundaries between low and high status groups are
vestigated how people with mental illness negotiate their stig-
perceived to be permeable (Ellemers et al., 1990). These strategies
matized identity. Those (mainly qualitative) studies that do exist
have implications not only for social and political structure, but
suggest that patients often experience depression as self-dening
also substantial and long-lasting consequences for health (Pascoe
(Karp, 1994), and hold self-stigmatizing beliefs, for example that
and Smart Richman, 2009; Schmitt et al., 2014)
depression reects a weakness of character (Cornford et al., 2007;
The rejection identication model (RIM; Branscombe et al.,
Boardman et al., 2011). Studies have also established that many 1999) brought attention to the fact that being subject to dis-
patients hold a biochemical causal model of depression (Lebowitz crimination due to one's membership of a social group increases
et al., 2013; Kvaale et al., 2013), and experience difculties in the salience of the social identity associated with that social group.
transitioning away from a mental illness identity (Howard, 2008). For example, a woman who experiences workplace discrimination
The question of how people negotiate low-status group mem- because of her gender is more likely to think of herself as a wo-
bership has been the focus of research attention in social man, particularly when at work. This social identication, or sub-
jective self-denition in terms of a social category, has a number of
n
Correspondence to: School of Psychology, University of Queensland, St Lucia,
well-established consequences, including for wellbeing. RIM
QLD 4072, Australia. Fax: 61 7 3365 4466. brought attention to the fact that when discrimination encourages
E-mail address: [Link]@[Link] (T. Cruwys). and enables social identication with a stigmatized group, some of

[Link]
0165-0327/Crown Copyright & 2015 Published by Elsevier B.V. All rights reserved.
 T. Cruwys, S. Gunaseelan / Journal of Affective Disorders 189 (2016) 3642 37

the harms of discrimination may be ameliorated. That is, social depression identity describes a person's lack of wellbeing. That is,
identication buffers individuals against the negative con- perhaps more so than any other social category, what it means to
sequences of stigma. RIM has been investigated in the context of be a member of the depressed group is to experience poor
identities including ethnic minorities (Branscombe et al., 1999; wellbeing. Thus, even if wellbeing is operationalized in terms of
Latrofa et al., 2009; Schmitt et al., 2003), older people (Garstka constructs distinct from depression (such as life satisfaction, an-
et al., 2004), women (Schmitt et al., 2002) and people with phy- xiety, stress), it seems likely that seeing depression as central to
sical disabilities (Fernndez et al., 2012), but not, to our knowl- one's identity is incompatible with wellbeing.
edge, in the context of mental illness groups.
RIM is part of a burgeoning body of research demonstrating
that social identication is the psychological mechanism through 1.1. The current study
which our social reality impacts our individual perception, and,
ultimately, our wellbeing (Haslam et al., 2009). For instance, This study, informed by the social identity approach to health,
people with depression who engage in recreational social groups, aimed to investigate the relationship between the experience of
or group psychotherapy, only experience improvement in their discrimination, mental illness identity, normative inuence, and
symptoms over time to the extent that they identify with the group wellbeing among people experiencing depression symptomatol-
in question (Cruwys et al., 2014b). The benets of social identi- ogy. Hypotheses one and two were derived from RIM (Branscombe
cation for health have been demonstrated for a wide array of po- et al., 1999), which has been established for a wide variety of
pulations and specic conditions, including in pilgrims (Tewari stigmatized group memberships, although never with mental ill-
et al., 2012), school children (Turner et al., 2014), people with ness identity:
multiple sclerosis (Wakeeld et al., 2013), or people who have (H1) Personal experience of discrimination towards depressed
experienced a traumatic event (Jones et al., 2011). people will be associated with reduced wellbeing (measured as
However, there has been increasing acknowledgment that we life satisfaction, stress, anxiety symptoms and depression
should expect some exceptions to this "rule"; that is, some social symptoms).
identities may do more harm than good. A recent theoretical re- (H2) Personal experience of discrimination towards depressed
view of social identity and depression (Cruwys et al., 2014a) hy- people will positively predict social identication as a depressed
pothesized that the exception to the protective benets of social person.
identity might be stigmatized groups, or those group where social Two competing hypotheses were proposed regarding the re-
inuence by other group members might prove harmful to mental lationship between social identication as a depressed person and
health. Specically, some groups have norms (the content of a wellbeing. First, also derived from RIM, we proposed a positive
social identity) that endorse thoughts or behaviors that are likely relationship between these variables, specically:
to worsen depression symptoms. In this case, we might predict (H3) Social identication as a depressed person will buffer
that identication would be associated with increased norm con- wellbeing against the negative effects of stigma.
formity (Turner, 1991), and thus poorer wellbeing. Very little re- In competition with this, however, and derived from a referent
search has found evidence of this to date, however, one recent informational inuence model (Turner, 1991; also H3 of Cruwys
study found that people in treatment for substance use disorder et al. (2014a)), we proposed a negative relationship between these
had better outcomes to the extent that they disidentied with variables, specically:
previous substance using identities (Dingle et al., 2014). One par- (H4) Social identication as a depressed person will predict
ticularly relevant study (Crabtree et al., 2010) explored social reduced wellbeing.
identication with mental health support groups and the con- Finally, and also consistent with a referent informational in-
sequences for social support, stigma-resistance, and self-esteem. uence account, we predicted a negative relationship between
Interestingly, while social identication was associated with social identication and wellbeing only among those who per-
greater perceived support and stigma resistance, this had the ef- ceived a strong norm of depressogenic thoughts and behaviors
fect of suppressing an otherwise negative relationship between among the depressed group. Specically:
identication and self-esteem. Although not investigating mental (H5) Social identication as a depressed person will interact
illness identity per se, these studies suggest the possibility that with perceived symptom norms of depressed people, such that
identication, particularly in the context of stigmatized groups, social identication will be associated with reduced wellbeing
need not always be benecial for wellbeing. only among those who characterize people with depression in
People with depression are part of a social category that is both terms of thoughts and behaviors that are consistent with depres-
(a) subject to stigma and discrimination, and (b) associated with sive symptomatology.
specic normative thoughts (e.g., hopelessness, negativity) and
behaviors (e.g., self-harm) that, if internalized by group members,
are likely to worsen mental health. Therefore this social category 2. Method
represents an ideal test case for the theoretical question of whe-
ther there are social categories where social identication is ne- 2.1. Participants
gatively associated with wellbeing. Furthermore, depression
identity differs from the social categories where the RIM has ty- Participants were recruited widely from psychology clinics,
pically been investigated in at least three important ways. First, health centers, research networks, online depression forums and
being a depressed person is usually concealable, and many paid participant pools. People were invited to participate if they
people choose not to disclose their diagnosis to others. Previous met the following criteria: experiencing persistent and intense
research has suggested that concealing stigmatized identities can feelings of sadness, people who feel depressed, and people who
be associated with reduced wellbeing (Barreto et al., 2006). Sec- have been formally diagnosed with depression by a health prac-
ond, the boundaries between mental illness and mental health are titioner. The study was not limited only to those diagnosed with
permeable, in that it is possible for a person to leave the stigma- depression or currently undergoing treatment because previous
tized group via recovery although perhaps not easily or quickly. evidence suggests that only a minority of people with mental ill-
Permeable group boundaries may reduce social identication (El- ness seek treatment (Goldman et al., 1999), and that these are not
lemers et al., 1990). Third, and perhaps most importantly, representative of the mentally ill population (Saxena et al., 2007).
38 T. Cruwys, S. Gunaseelan / Journal of Affective Disorders 189 (2016) 3642

The nal sample included 250 people with depression, 61.6% of most of the time. For each subscale, responses were summed and
which reported that they had received a formal mood disorder multiplied by two in accordance with recommended practice
diagnosis from a health professional, while the remaining parti- (Lovibond and Lovibond, 1995). The mean scores were in the
cipants reported current depression symptoms of at least moder- severe range for depression (M 24.92; SD 9.82), the severe
ate severity (according to published cut-off criteria; Lovibond and range for anxiety (M16.50; SD 9.31) and the moderate range
Lovibond, 1995). Participants without a formal diagnosis whose for stress (M 23.45; SD 9.02).
current symptoms were mild or subclinical were excluded from
the survey (an additional N 60). This was a conservative ap- [Link]. Life satisfaction
proach in order to increase our condence that all participants had
experienced, or were experiencing, clinical depression symptoms. Life satisfaction was measured using the well-validated Sa-
The mean age was 27.37 years (SD 11.64; range 1770). Partici- tisfaction with Life Scale (Diener et al., 1985). Participants rated
pants were 64% female and recruited from across the world, with ve items such as If I could live my life over, I would change al-
the most common nationalities being United States (38.8%), Aus- most nothing on a seven-point scale from Strong Disagree to
tralia (36.8%), and India (5.6%), with small numbers of participants Strong Agree ( .90). This sample reported low wellbeing on
from 20 other countries. average (M 15.83, SD 7.03), approximately equivalent to a clin-
ical outpatient sample at intake (Pavot and Diener, 1993).

2.2. Measures 2.3. Social identication

2.2.1. Wellbeing In recent years, there has been a divergence between the
measurement of social identication in terms of simple, one-item
Wellbeing was assessed using four overlapping but distinct scales (e.g., I identify as a depressed person, Postmes et al., 2012)
constructs: depression, anxiety, stress, and life satisfaction. This or in terms of complex, multidimensional constructs (Leach et al.,
was to address the potential restriction of range that may have 2008). Our view is that, while a single construct is likely to capture
resulted from using depression symptoms as both a criteria for social identity phenomena in many contexts, it is less appropriate
eligibility as well as a dependent variable. However, as can be seen for social categories that have negative and stigmatized content
from Table 1, almost the full range of the scale was utilized for all (see also Mlicki and Ellemers, 1996; Ouwerkerk et al., 1999). More
four wellbeing variables. specically, we hypothesize that the identity centrality dimension
(Being depressed is an important part of how I see myself) might
[Link]. Depression Anxiety Stress Scales (DASS-21) best approximate the theoretical component of self-denition, or
self-categorization, that is relevant to the hypotheses of the present
The DASS-21 is a well-validated short form of the Depression study. Therefore we measured all ve dimensions of social identi-
Anxiety Stress Scales (Lovibond and Lovibond, 1995). The measure cation in accordance with the recommendations of Leach and
includes three seven-item subscales, assessing depression, anxiety, colleagues (2008) and made a tentative hypothesis that our hy-
and stress symptoms respectively. The DASS has excellent validity pothesized effects would be most apparent for identity centrality.
in both clinical and non-clinical samples and reliability of at least Participants responded to 11 items such as I feel a bond with
.88 (Crawford et al., 2009; Henry and Crawford, 2005). For other people who have depression (identity solidarity, .86),
instance, one study with a clinical sample found that the DASS-21 Being part of a group of people who have depression gives me a
accurately distinguishes between individuals with mood disorder, good feeling (identity satisfaction, one item only), The fact that I
panic disorder, or obsessivecompulsive disorder (Page et al., have depression is an important part of my identity (identity
2007). Participants were asked to indicate how frequently in the centrality, .83), I am similar to the average person who has
preceding week they had experienced symptoms such as I felt depression (identity self-stereotyping, r .78), and People who
like I wasn't worth much as a person, (depression subscale, have depression have a lot in common with each other (identity
.88), I felt I was close to panic (anxiety subscale, .81) and homogeneity, r .83) on a seven-point scale from Strongly Dis-
I tended to over-react to situations (stress subscale, .84) from agree to Strongly Agree. Only one item from the identity sa-
0 Did not apply to me at all to 3 Applied to me very much, or tisfaction subscale was included in the survey as the other items

Table 1
Descriptive statistics and correlations.

M SD Range Correlations

1 2 3 4 5 6 7 8 9 10

1. Depression 24.92 9.82 242

2. Anxiety 16.50 9.31 042 .48n
3. Stress 23.45 9.02 042 .57n .61n
4. Life satisfaction 15.83 7.03 532  .47n  .12 .24n
5. Identity solidarity 4.36 1.34 17 .15 .24n .18n .01
6. Identity satisfaction 3.61 1.67 17 .08 .28n .14n .03 .66n
7. Identity centrality 4.23 1.44 17 .29n .25n .30n  .18n .40n .34n
8. Identity self-stereotyping 4.24 1.38 17 .07 .14n .10  .05 .40n .33n .37n
9. Identity homogeneity 4.04 1.39 17 .07 .12 .02  .04 .40n .39n .31n .69n
10. Discrimination 3.71 1.55 17 .26n .27n .21n  .17n .29n .22n .41n .25n .22n 
11. Symptom norms 5.56 0.99 17 .24n .12 .11  .19n .20n .01 .16n .11 .14n .08

N 250.
n
p o .05.
 T. Cruwys, S. Gunaseelan / Journal of Affective Disorders 189 (2016) 3642 39

were judged to have poor t with the depression category (e.g. It notion that depression informed their identity. Often, this included
is pleasant to be a depressed person).1 reference to the idea that depression shaped their behavior and
Participants were also given the opportunity to provide an was preoccupying, or an emphasis that depression provided a
optional open-ended response to the question Is depression part cogent explanation for their subjective experience. Some examples
of how you see yourself as a person? are:

Unfortunately depression is now so rmly rooted with me that it

2.4. Perceived discrimination
is an indelible part of my persona. I wouldn't know how to act
without depression. (M, 58)
Personal experience with discrimination as a depressed person
[Depression] has control over every thing I do, each day. It de-
was measured using an adapted version of the Perceptions of
termines what I will and will not do (F, 22)
Discrimination Scale (Schmitt and Branscombe, 2002). Participants
Yes, I always see myself as a generally depressed and unhappy
responded to six items such as I regularly encounter discrimina-
person. I can't imagine myself any other way. (F, 43)
tion against people with depression on seven-point scale from
Depression becomes all I can think about myself. Depression
Strongly disagree to Strongly agree ( .91).
becomes all-consuming. (M, 42)
2.5. Symptom norms Approximately one third (35.6%) of participants who provided a
valid response felt ambivalent about identifying as a depressed
The measure of symptom norms consisted of 4 items adapted person, for example:
from various sources (e.g., Cruwys et al., 2015; Smith and Louis,
2008) that asked participants to rate how likely it would be for It's hard to see myself without depression, however I don't feel
other people with depression to engage in four kinds of symptom- that it denes me. (F, 20)
related cognitions (e.g. Keep thinking negative and unhelpful Depression permeates virtually every aspect of my life. At the
thoughts) and behaviors (e.g. Engage in self-harm behaviors) on same time, I don't want depression to dene how I see myself as a
a seven-point scale from Very unlikely to Very likely. Items person. I deal with it. It is a part of me, but it is not all that I am. I
were chosen to map directly onto diagnostic criteria. Norms for want to be dened more by the things that give me a reason to live
emotional symptoms, such as guilt and sadness, were not directly and bring fulllment into my life not by the things that make life
assessed for this scale, in order to minimize conceptual overlap hard. (F, 43)
with the wellbeing items. The reliability of the items was good
( .82). 3.2. Evaluation of the hypotheses

The primary quantitative results are summarized in Fig. 1. To

3. Results assess the rst hypothesis, concerning a negative relationship
between discrimination and wellbeing, four separate regression
Descriptive statistics and correlations are outlined in Table 1. analyses were conducted (one for each dependent variable). Per-
sonal experience of discrimination predicted lower life satisfac-
tion, F(1,248) 7.76, p .006, R2 .03, more depression symptoms,
3.1. Qualitative responses
F(1,248) 17.50, p o.001, R2 .07, more anxiety symptoms, F
(1,248) 18.95, p o.001, R2 .07, and greater stress, F(1,248)
121 participants (48.4%) chose to provide comments in re-
10.86, p .001, R2 .04. Therefore H1 was conrmed: the experi-
sponse to the open-ended optional question Is depression part of
ence of discrimination as a depressed person was associated with
how you see yourself as a person? Some participants (15.4% of
poorer wellbeing.
those who provided a valid response) rejected the notion that
Similarly, to assess the second hypothesis, concerning a positive
depression was or should form part of their self-denition.
relationship between discrimination and identication, ve sepa-
Often, this included an explicit comparison to a physical health
rate regression analyses were conducted (one for each component
condition (e.g., a headache), or an emphasis on the possibility of
of identication). Personal experience of discrimination predicted
change or recovery. Some examples are:
greater identity centrality, F(1,248) 51.07, p o.001, R2 .17,
Not at ally I strongly disagree with that phrase, depression greater identity solidarity, F(1,248) 22.70, p o.001, R2 .08,
should not be used as an identity. It is a mental illness but doesn't greater identity satisfaction, F(1, 247) 12.53, p o.001, R2 .05,
dene who a person is. Associating it with who you are would greater perceived identity homogeneity, F(1,248) 12.99, p o.001,
only make it that much difcult to overcome it. (Female, age 26) R2 .05, and greater identity self-stereotyping, F(1,248) 16.54,
Depression is an illness not a denition of who I amyI am not po .001, R2 .06. Therefore H2 was conrmed: the experience of
my illness no more than someone with cancer or heart problems discrimination as a depressed person was associated with stronger
are dened by their illness. (F, 22) identication as a depressed person. Discrimination was a parti-
No, I still see myself as a unique individual who is going through cularly strong predictor of the centrality of depression identity.
a tough time at the moment but it is not me in reality. (F, 57) In order to assess the third and fourth hypotheses, Hayes
PROCESS (2012, model 4) was used. The model included dis-
Contrary to this view, however, a much larger group of parti- crimination as the independent variable, the ve components of
cipants (49% of those who provided a valid response) endorsed the identication as mediating variables, and a global wellbeing
measure (created from the mean of the z-scores of the four se-
1
Factor analysis indicated a three-factor solution of social identication was parate wellbeing variables, such that high scores indicated greater
the best t for these data. The rst factor corresponded to the centrality subscale; wellbeing) as the dependent variable.2 There was no signicant
the second factor consisted of an entitativity dimension made up of self-stereo-
typing and homogeneity; and the third factor was made up of solidarity plus sa-
2
tisfaction. Given that our results focus on the identity centrality subscale, the co- Note that these ndings also held when the wellbeing measures were as-
herence of which is supported by both the factor analysis and the extant literature, sessed separately, such that for depression, stress, and life satisfaction the results
we retained the ve-factor model for our analyses in keeping with published re- were identical. For anxiety, identity satisfaction, rather than identity centrality, was
commendations (Leach et al., 2008). the signicant mediator.
40 T. Cruwys, S. Gunaseelan / Journal of Affective Disorders 189 (2016) 3642

Fig. 1. The centrality of depression to one's identity was associated with poorer wellbeing, particularly among those who thought that depression symptoms characterized
people with depression.
NB. Standardized beta values are included in the gure and in text to aid interpretability, however, signicance levels were calculated using unstandardized values in Hayes
PROCESS models 4 and 14, in accordance with recommendations (Hayes and Preacher, 2013).

indirect effect via ingroup homogeneity, ingroup self-stereotyping, nd a direct negative relationship between these two variables. It
identity satisfaction, or identity solidarity. That is, for these sub- has previously been theorized, however, that there are conditions
scales, neither H3 nor H4 were supported; these measures of under which social identication could harm wellbeing (Cruwys
identication were independent of wellbeing. However, identity et al., 2014a), specically in the context of stigmatized group
centrality was a signicant indirect mediator of the relationship membership or group norms that proscribe activities harmful to
between discrimination and wellbeing. The direction of this re- wellbeing. Therefore, rather than being a disconrmation of the
lationship was consistent with H4, such that identication as a social identity approach to health (including the rejection identi-
depressed person was associated with reduced wellbeing cation model), we see these results as providing evidence for
( .26, p o.001) and indirectly mediated the negative re- boundary conditions under which the psychological resources
lationship between discrimination and wellbeing (  .11, CI: provided by social identity become toxic to health.
.18,  .05). These data also have implications for the way in which we
A sensitivity analysis was conducted to investigate the most conceptualize social relationships and group membership. That is,
plausible alternative model, specically that people with more it is the subjective psychological connection with a group social
severe depression symptoms perceive themselves to be more identication and not one's ofcial category membership that
prototypical of the group (ingroup self-stereotyping) and subse- matters for mental health. All of the participants in the current
quently come to identify more strongly as depressed. However, study were, objectively speaking, part of the social category de-
this model was not supported by the data: depression severity did pressed people, and yet it was only those who self-dened as
not predict ingroup self-stereotyping ( .01, p .254) and in- such who were subject to potentially harmful social inuence
group self-stereotyping did not mediate the relationship between processes from fellow people with depression. This was borne out
depression severity and identication ( .00, CI: .01, .01). by the qualitative results, where those who accepted the depres-
Hypothesis ve, concerning the moderating role of symptom sion identity label emphasized how this identity shaped their
norms, was tested using Hayes PROCESS (2012, model 14). The behavior. Relatedly, these data provide further evidence for the
interaction between identity centrality and symptom norms in importance of identity content. Cruwys and colleagues (2014b)
predicting wellbeing was signicant, t(245) 3.19, p .002, .16. found that depressed people who socially identied with recrea-
Specically, among those who were below the median (and the tional groups or psychotherapy groups experienced reduced de-
mid-point of the scale) in their identication as depressed, there pression symptoms over time. While this might, on the surface,
was no relationship between symptom norms and one's own appear to be a comparable sample with contrary results to those
wellbeing (t(100) .79, p .430). That is, people who did not reported here, the norms of recreational or psychotherapy groups
identify as depressed were not inuenced by the symptom norms are critically different from those of depressed people, and thus
of depressed people. Among those above the median, however, their consequences for mental health diverge.
there was a signicant relationship between symptom norms and It is interesting to note that the reverse pathway, whereby the
wellbeing (t(146) 4.47, p o.001). This was such that among experience of depression symptoms leads to self-stereotyping and
people who identied as depressed, there was an association subsequent self-denition, was not supported by these data. This
where the more they characterized depressed people in terms of pathway does, however, remain a plausible means through which
depression symptomatology, the poorer their own wellbeing. a depression identity might develop over time in a sub-clinical
Therefore, H5 was conrmed identifying as depressed was sample. This might have historically occurred in the present
harmful to wellbeing because it facilitated convergence towards sample, however, this could not explain the observed differences
norms of depressogenic thoughts and behaviors. in wellbeing. It is worth noting also that previous research in-
vestigating the role of ingroup self-stereotyping also concluded
that it follows from, rather than precedes, identication and
4. Discussion wellbeing (Latrofa et al., 2009). Nevertheless, a limitation of this
study is its correlational nature, and no reverse causal pathway
This study is the rst to quantitatively demonstrate that social can be condently ruled out here.
identication as mentally ill is associated with the experience of In terms of applied implications, these data are relevant to the
discrimination and poorer wellbeing. Furthermore, out of dozens ongoing debate regarding the clinical impact of diagnosis. We do
of extant studies of social identity and wellbeing (for reviews, see not dispute the benets of diagnosis for guiding formulation and
Cruwys et al. (2014a) and Jetten et al. (2014)), this is the rst to treatment. Nevertheless, these ndings suggest that diagnosis, to
 T. Cruwys, S. Gunaseelan / Journal of Affective Disorders 189 (2016) 3642 41

the extent that it contributes to patients self-denition in terms of Ellemers, N., Wilke, H., Van Knippenberg, A., 1993. Effects of the legitimacy of low
their mental illness, may have harmful consequences for well- group or individual status on individual and collective status-enhancement
strategies. J. Personal. Soc. Psychol. 64 (5), 766778.
being. Going beyond our data somewhat, but in keeping with Fernndez, S., et al., 2012. Inuence of the social context on use of surgical-
social-psychological theory, we posit that when a mental illness lengthening and group-empowering coping strategies among people with
diagnosis is disclosed to a patient, it ought to be accompanied by dwarsm. Rehabil. Psychol. 57 (3), 224235.
Garstka, T.A., et al., 2004. How young and older adults differ in their responses to
information that (a) challenges the legitimacy of stigma towards perceived age discrimination. Psychol. Aging 19 (2), 326335.
people with mental illness, and (b) encourages patients to see Gleibs, I.H., Haslam, C., Jones, J.M., et al., 2011a. No country for old men? The role of
mental illness as a temporary state, rather than a permanent a Gentlemens Club in promoting social engagement and psychological well-
being in residential care. Aging Ment. Health 15 (4), 456466.
condition. This combination may limit negative effects and enable Gleibs, I.H., Haslam, C., Haslam, S.A., et al., 2011b. Water clubs in residential care: is
patients to engage in strategies of individual mobility, that is, re- it the water or the club that enhances health and well-being? Psychol. Health
covery efforts. 26 (10), 13611377.
Goldman, L.S., Nielsen, N.H., Champion, H.C., 1999. Awareness, diagnosis, and
A further implication for clinical practice is that, while inter-
treatment of depression. J. Gen. Intern. Med. 14 (9), 569580.
ventions to boost social engagement are very much indicated, Hansen, N., Sassenberg, K., 2011. Exploring the self-directed anger of the stigma-
attention must be paid to the normative content of these groups. tized: the interplay between perceived legitimacy and social identication.
Group Process. Intergroup Relat. 14, 807818.
Illness support groups, including online communities, have the
Haslam, S.A., Jetten, J., Postmes, T., Haslam, C., 2009. Social identity, health and
potential to be harmful if they encourage patients to adopt an well-being: An emerging agenda for applied psychology. Appl. Psychol. An Int.
illness identity that is framed in permanent, biochemical terms, or Rev. 58, 123.
Hayes, A.F., Preacher, K.J., 2013. Statistical mediation analysis with a multi-
to engage in collective rumination (see also Howard, 2006 and
categorical independent variable. Br. J. Math. Stat. Psychol., 120.
Kvaale et al., 2013). Instead, clinicians should seek to facilitate al- Henry, J.D., Crawford, J.R., 2005. The short-form version of the Depression Anxiety
ternative, positive sources of self-denition around which social Stress Scales (DASS-21): construct validity and normative data in a large non-
identities might cohere, including recreational or community clinical sample. Br. J. Clin. Psychol. 44 (2), 227239.
Howard, J., 2006. Expecting and accepting: the temporal ambiguity of recovery
groups (e.g. Cruwys et al., 2014b; Gleibs et al., 2011a,, 2011b). identities. Soc. Psychol. Q. 69 (4), 307324.
In sum, these data are the rst to show, in a large international Howard, J., 2008. Negotiating an exit: existential, interactional, and cultural ob-
sample of people with depression, that identifying as mentally ill stracles to disorder disidentication. Soc. Psychol. Q. 71, 177192.
Jetten, J., et al., 2014. How groups affect our health and well-being: the path from
is related to the experience of stigma and with poorer wellbeing. theory to policy. Soc. Issues Policy Rev. 8 (1), 103130.
These data are also the rst to provide evidence of a theorized, but Jones, J.M., et al., 2011. That which doesnt kill us can make us stronger (and more
to date hypothetical, boundary condition for the benets of social satised with life): the contribution of personal and social changes to well-
being after acquired brain injury. Psychol. Health 26 (3), 353369.
identication for wellbeing. Identifying as depressed enables a Karp, D.A., 1994. Living with depression: illness and identity turning points. Qual.
process of social inuence, whereby people are more likely to Health Res. 4 (1), 630.
converge upon the norms of the in-group: in this case, thoughts Kvaale, E.P., Haslam, N., Gottdiener, W.H., 2013. The side effects of medicalization:
a meta-analytic review of how biogenetic explanations affect stigma. Clin.
and behaviors that are depressogenic. In this way, the stigma of Psychol. Rev. 33 (6), 782794.
depression poses a double burden not only through its direct Latrofa, M., et al., 2009. United we stand, divided we fall! The protective function
harm to health, but also by bolstering mental illness identity and of self-stereotyping for stigmatised members psychological well-being. Appl.
Psychol. 58 (1), 84104.
its associated ills. New directions in depression treatment should Leach, C.W., van Zomeren, M., Zebel, S., Vliek, M.L.W., Pennekamp, S.F., Doosje, B.,
ideally establish methods of empowering people to recover from Ouwerkerk, J.W.,, 2008. Group-level self-denition and self-investment: A
depression without the need to accept or enact an illness label. hierarchical (multicomponent) model of in-group identication. J. Pers. Soc.
Psychol. 95, 144165.
Lebowitz, M.S., Ahn, W.-K., Nolen-Hoeksema, S., 2013. Fixable or fate? Perceptions
of the biology of depression. J. Consult. Clin. Psychol. 81 (3), 518527.
References Link, B.G., et al., 1997. On stigma and its consequences: evidence from a long-
itudinal study of men with dual diagnoses of mental illness and substance
abuse. J. Health Soc. Behav. 38 (2), 177190.
Barreto, M., Ellemers, N., Banal, S., 2006. Working under cover: Performance-re- Link, B.G., et al., 1999. Public conceptions of mental illness: labels, causes, dan-
lated self-condence among members of contextually devalued groups who try gerousness, and social distance. Am. J. Public Health 89 (9), 13281333.
to pass. Eur. J. Soc. Psychol. 36, 337352. [Link] Lovibond, P.F., Lovibond, S.H., 1995. The structure of negative emotional states:
Boardman, F., et al., 2011. Resilience as a response to the stigma of depression: a comparison of the Depression Anxiety Stress Scales (DASS) with the Beck De-
mixed methods analysis. J. Affect. Disord. 135, 267276. pression and Anxiety Inventories. Behav. Res. Ther. 33 (3), 335343.
Branscombe, N.R., Schmitt, M.T., Harvey, R.D., 1999. Perceiving pervasive dis- Mlicki, P.P., Ellemers, N., 1996. Being different or being better? National stereotypes
crimination among African Americans: Implications for group identication and identications of Polish and Dutch students. Eur. J. Soc. Psychol. 26 (No-
and well-being. J. Personal. Soc. Psychol. 77 (1), 135149. vember 1994), 97114.
Cornford, C.S., Hill, A., Reilly, J., 2007. How patients with depressive symptoms view Ouwerkerk, J.W., Ellemers, N., Kortekaas, P., 1999. Self-categorisation, commitment
their condition: a qualitative study. Fam. Pract. 24, 358364. to the group and group self-esteem as related but distinct aspects of social
Corrigan, P.W., Druss, B.G., Perlick, D.A., 2014. The impact of mental illness stigma identity. Eur. J. Soc. Psychol. 29, 371389.
on seeking and participating in mental health care. Psychol. Sci. Public Interest Page, A.C., Hooke, G.R., Morrison, D.L., 2007. Psychometric properties of the De-
15 (2), 3770. pression Anxiety Stress Scales (DASS) in depressed clinical samples. Br. J. Clin.
Crabtree, J.W., et al., 2010. Mental health support groups, stigma, and self-esteem: Psychol. 46 (3), 283297.
positive and negative implications of group identication. J. Soc. Issues 66 (3), Pascoe, E.A., Smart Richman, L., 2009. Perceived discrimination and health: a meta-
553569. analytic review. Psychol. Bull. 135 (4), 531554.
Crawford, J.R., et al., 2009. A convenient method of obtaining percentile norms and Pavot, W., Diener, E., 1993. Review of the satisfaction with life scale. Psychol. Assess.
accompanying interval estimates for self-report mood scales (DASS, DASS-21, 5 (2), 164172.
HADS, PANAS, and sAD). Br. J. Clin. Psychol. 48 (2), 163180. Postmes, T., Haslam, S.A., Jans, L., 2012. A single-item measure of social identica-
Cruwys, T., Haslam, S.A., et al., 2014a. Depression and social identity : an integrative tion: Reliability, validity, and utility. Br. J. Soc. Psychol. [Link]
review. Personal. Soc. Psychol. Rev. 18 (3), 215238. bjso.12006.
Cruwys, T., Haslam, S.A., Dingle, G.A., et al., 2014b. Feeling connected again: inter- Rsch, N., et al., 2014. Does the stigma of mental illness contribute to suicidality?
ventions that increase social identication reduce depression symptoms in Br. J. Psychiatry 205 (4), 257259.
community and clinical settings. J. Affect. Disord. 159, 139146. Saxena, S., et al., 2007. Resources for mental health: scarcity, inequity, and in-
Cruwys, T., et al., 2015. Thats not what we do: evidence that normative change is efciency. Lancet 370, 878889.
a mechanism of action in group interventions. Behav. Res. Ther. 65, 1117. Schmitt, M.T., et al., 2002. Perceiving discrimination against one's gender group has
Diener, E., et al., 1985. The Satisfaction With Life Scale. J. Personal. Assess. 49 (1), different implications for well-being in women and men. Personal. Soc. Psychol.
7175. Bull. 28 (2), 197210.
Dingle, G.A., et al., 2014. Breaking good: Breaking ties with social groups may be Schmitt, M.T., et al., 2014. The consequences of perceived discrimination for psy-
good for recovery from substance misuse. Br. J. Soc. Psychol., 119. chological well-being: a meta-analytic review. Psychol. Bull. 140 (4), 921948.
Ellemers, N., van Knippenberg, A., Wilke, H., 1990. The inuence of permeability of Schmitt, M.T., Branscombe, N.R., 2002. The meaning and consequences of perceived
group boundaries and stability of group status on strategies of individual mo- discrimination in disadvantaged and privileged social groups. Eur. Rev. Soc.
bility and social change. Br. J. Soc. Psychol. 29 (3), 233246. Psychol. 12, 167199.
42 T. Cruwys, S. Gunaseelan / Journal of Affective Disorders 189 (2016) 3642

Schmitt, M.T., Spears, R., Branscombe, N.R., 2003. Constructing a minority group Turner, J.C., 1991. Social Inuence. Open University Press, Milton Keynes.
identity out of shared rejection: the case of international students. Eur. J. Soc. Wakeeld, J.R.H., Bickley, S., Sani, F., 2013. The effects of identication with a
Psychol. 33 (1), 112. support group on the mental health of people with multiple sclerosis. J. Psy-
Smith, J.R., Louis, W.R., 2008. Do as we say and as we do: The interplay of de- chosom. Res. 74 (5), 420426.
scriptive and injunctive group norms in the attitude-behaviour relationship. Br. Yanos, P.T., Roe, D., Lysaker, P.H., 2010. The impact of illness identity on recovery
J. Soc. Psychol. 47, 647666. from severe mental illness. Am. J. Psychiatr. Rehabil. 13 (2), 7393.
Tajfel, H., Turner, J.C., 1979. An integrative theory of intergroup conict. In: Austin,
W.G., Worehel, S. (Eds.), The Social Psychology of Intergroup Relations. Brooks/
Cole, Monterey, pp. 3347.
Tewari, S., et al., 2012. Participation in mass gatherings can benet well-being:
longitudinal and control data from a North Indian Hindu pilgrimage event. PLoS
One 7 (10), e47291.
Turner, I., et al., 2014. Well-being, school climate, and the social identity process: a
latent growth model study of bullying perpetration and peer victimization. Sch.
Psychol. Q. 29 (3), 320335.