Community Ment Health J (2010) 46:517–522 DOI 10.1007/s10597-009-9287-3
ORIGINAL PAPER
Changing Stigmatizing Perceptions and Recollections About Mental Illness: The Effects of NAMI’s In Our Own Voice Patrick W. Corrigan • Jennifer D. Rafacz • Julie Hautamaki • Jessica Walton • Nicolas Ru¨sch • Deepa Rao • Patricia Doyle • Sarah O’Brien • John Pryor • Glenn Reeder
Received: 25 July 2009 / Accepted: 28 December 2009 / Published online: 14 January 2010 Ó Springer Science+Business Media, LLC 2010
Abstract In Our Own Voice (IOOV) is a 90-min antistigma program that comprises face-to-face stories of challenges of mental illness and hopes and dreams commensurate with recovery. We pared down IOOV to a 30min version, using information from two focus groups. In this study, effects of 90- versus 30-min IOOV are contrasted with 30 min of education. Two hundred research participants were randomly assigned to one of these three conditions and completed a measure of stigmatizing perceptions and recollections. People in the education group remembered more negatives than the two IOOV groups. To control for overall response rate, a difference ratio was determined (difference in positive and negative recollection divided by overall recollections). Results showed the two IOOV conditions had significantly better ratios than education. These findings suggest the 30 min version of IOOV is as effective as the 90 min standard.
P. W. Corrigan (&) J. D. Rafacz J. Hautamaki J. Walton N. Ru¨sch Institute of Psychology, Illinois Institute of Technology, 3424 S State Street, Chicago, IL 60616, USA e-mail:
[email protected] D. Rao University of Washington, Seattle, WA, USA P. Doyle NAMI, DuPage County, IL, USA S. O’Brien NAMI-National, Arlington, VA, USA J. Pryor G. Reeder Illinois State University, Normal, IL, USA
Keywords Stigma Discrimination In Our Own Voice Anti-stigma intervention Contact Education Recollection Memory
Introduction The stigma of mental illness impedes the lives and goals of people with serious mental illness in several ways. Antistigma programs have been developed to diminish these egregious effects. In this paper, we examine the impact of one such program—In Our Own Voice—on the public stigma of mental illness. Frequently, research of this kind focuses on changing attitudes which are likely to be biased by social desirability. We sought to expand measurement of stigma by using a life story perception and recollection task as outcome index; it provides a more dynamic and intervention-sensitive assessment than typical attitudes assessments. The stigma of mental illness has been described in various ways with two models being especially common (Hinshaw 2007; Link and Phelan 2001; Thornicroft 2006). First public stigma: the prejudice and discrimination that occur when the general population endorses the stigma; and second self-stigma: the threat to self-esteem and selfefficacy that occurs when the person internalizes stigma. Public stigma is the focus of this study. Three omnibus anti-stigma approaches have been identified as relevant to public stigma change (Corrigan and Penn 1999). Protest involves review of heinous ways in which people with mental illness are represented in the media and more broadly in the population; e.g., tabloid headlines that read ‘‘Get the Violent Crazies Off Our Streets,’’ from a November 1999 front page of New York’s Daily News. Protest yields an appeal to moral authority, directing the
123
518
public to suppress such behavior and thoughts. Education involves comparison of the misunderstandings about mental illness with the facts. Contact is face-to-face interactions of the public and people with mental illness. Research has shown that protest leads to no change in stigmatizing attitudes and may actually worsen it (Corrigan et al. 2001). Education and contact have been shown to yield positive gains in attitudes about mental illness though research on which condition yields better outcomes— education or contact—has varied by study (Corrigan et al. 2002). Much of the rigorous research to date has used randomized trials comparing the intervention of interest and a wait list control. Given that education and contact both seem to yield positive effects, both are included in the study described herein. Prior research has principally examined attitudinal measures of outcome. Attitude measures are limited because they are significantly affected by social desirability; improving attitudes that help people to appear unbiased, regardless of what actual stigma they might endorse (Hinshaw 2007). Social desirability at baseline may yield ceiling effects which diminish sensitivity to change. Researchers have looked to information processing measures which are influenced less by social desirability and more sensitive to change (Johnston and Macrae 1994; Lincoln et al. 2008; Macrae et al. 1994; Ru¨sch et al. 2010a, b, c). Research has used perception and recollection of positive and negative characteristics described in a person’s statement about his or her life and found that members of the general public who recalled more negatives are likely to be more stigmatizing than a comparison group (Corrigan et al. 2001). Perception and subsequent recall of this kind of information is influenced by existing schemas; these are knowledge structures that are influenced or activated by contemporaneous actions like education or contact. The direction of the intervention effect depends on the ratio of positive to negative recollections. The National Alliance on Mental Illness (NAMI) developed In Our Own Voice (IOOV) as an anti-stigma program. It is largely a contact program that was developed and implemented by NAMI consumers. Research has shown that participation in a group presentation of IOOV compared to control groups led to significantly less social avoidance and less endorsement of other stigmatizing attitudes (Rusch et al. 2008; Wood and Wahl 2006). The original IOOV was 90 min which seemed to tax its feasibility for many audiences. For example, employers are often targeted for change in order to help more people with mental illness get back to work. They might be engaged in IOOV as part of a civic group, weekly luncheon. Most luncheons of this kind, however, have no more than a half hour for outside speakers. Police officers are also an important group and morning roll call is a suitable venue
123
Community Ment Health J (2010) 46:517–522
for IOOV. Most roll calls will have no more than 30 min, so again a shorter version is needed. One of the goals of this study was to contrast a 30 min IOOV to the original 90 min version. The way in which the 30 min version was pared down from the 90 is summarized in the ‘‘Methods’’ section. The goal of this study is to determine how perception and recall of person characteristics varies across the 30 and 90 min IOOV, and education. We expect research participants in the education to recall more negative, and less positive, characteristics compared to the two IOOV groups. We also seek to show that a 30 min IOOV will be as successful as the 90 min condition. We recognize this as searching for support of the null, but do so viewing the 90 min version as the stigma change standard.
Methods Research participants were recruited from the student body of four colleges in the Midwest (N = 200). Demographics of the sample are summarized in Table 1 and discussed in the ‘‘Results’’ section. Only 10 out of 210 informed subjects decided not to participate in the study. Research participants who consented to participate were then randomized to one of three anti-stigma approaches: the original 90 min IOOV, a 30 min version of IOOV, or a 30 min education comparison group. Research participants completed the Life Story Memory Test (LSMT) immediately after completing their assigned condition. The LSMT was only administered at post-test because memory tests scores are typically influenced by practice effects especially in a study with multiple trials. Anti-Stigma Interventions In Our Own Voice (IOOV) was developed by people with serious mental illness associated with the National Alliance on Mental illness (NAMI). As originally developed, IOOV is a 90-min group interaction (hence IOOV-90) provided by two group facilitators with serious mental illness in recovery. ‘‘Recovery’’ means the person is able to control distress, approach hope, and achieve goals. IOOV-90 comprises five components which facilitators review in order: (1) Dark Days, a discussion of first experiences of mental illness and its distress; (2) Acceptance, a review of how the person has come to terms with his or her illness; (3) Treatment, what kind of therapies work best for the individual; (4) Coping Mechanisms, day-to-day strategies used to meet stressors; and (5) Successes, Hopes, and Dreams, people do overcome mental illness and are able to move ahead on their goals. IOOV-90 also includes a videotape that corresponds with each of the five components.
Community Ment Health J (2010) 46:517–522
519
Table 1 Demographics for overall sample and for each assigned condition
Age
Overall (N = 200) M (SD)
30 min IOOV group (N = 67) M (SD)
90 min IOOV group (N = 66) M (SD)
Education group (N = 67) M (SD)
20.2 (2.96)
20.1 (2.48)
20.4 (2.15)
20.1 (3.75)
Group differences
N.S. P [ .85
Gender % female
66.7%
59.7%
70.8%
70.6%
N.S. P [ .30
Ethnicity % Euro American
80.4%
87.3%
65.0%
88.5%
v2(8) = 15.7
Afr American
8.2
7.9
14.3
3.3
P \ .05
Asian American
8.2
3.2
15.0
6.6
Other
2.7
1.6
5.0
1.6
10.5%
13.4%
10.8%
7.4%
Percent hispanic
a
N.S. P [ .53
Marital status % Single
98.5%
98.5%
98.5%
98.5%
N.S.
Married
1.0
0
1.5
1.5
P [ .56
Separated/divorced
0
0
0
0
Widowed
.5
1.5
0
0
H.S. grad
2.5%
3.0%
1.5%
3.0%
N.S.
Some college
93.5
95.5
92.3
92.5
P [ .87
College degree Master’s degree
3.0 1.0
1.5 0
4.6 1.5
3.0 1.5
% $0–20 K
21.7%
22.7%
24.6%
17.9%
N.S.
$20–40 K
12.6
7.6
16.9
13.4
P [ .65
$40–60 K
15.7
15.2
18.5
13.4
$60–80 K
15.2
15.2
13.8
16.4
34.8
39.4
26.2
38.8
96%
97%
100%
Education
Household income
More than $80 K Fidelity to manual scores
Ratio of intervention behaviors to overall behaviors Percent demonstrated behaviors
N.S.
Fidelity scores are also provided here a
Consistent with the US Census, research participants are queried first about ethnicity (e.g., African American, European American, or Asian American) and then Hispanic background. The questions are asked separately to specifically examine interactions of these two constructs
Video segments are played at the beginning of each component and then two facilitators discuss corresponding personal experiences. Facilitators also have discussion questions which they use to stimulate group interactions. Given that the 90 min IOOV is not always feasible for important targets such as employers, an ad hoc group of NAMI consumers and services researchers convened to consider reducing time for IOOV from 90 to 30 min. To inform these decisions, data were collected from two focus groups about perceptions of and experiences with IOOV90. The first group were people from the community at large who had in the past month participated in the 90 min
version of IOOV (two men and four women). The second comprised trained and experienced IOOV facilitators (also two men and four women) including the State of Illinois coordinator of IOOV. We developed an interview guide that asked focus group participants to evaluate the benefits and limitations of the five IOOV components. They were also asked to reflect on more general structural issues (e.g., time and place of IOOV sessions) as well as those more pedagogic in nature (e.g., the number of facilitators and the videotape). Groups were conducted by a research investigator along with a research associate who took verbatim notes. Results of the
123
520
focus groups were coded with the goal of identifying aspects of IOOV-90 that might be omitted thereby reducing it to a 30 min version. In particular, our goal was to determine whether any of the five components, or perhaps structural and pedagogic issues, should be cut. Consensus arose on two sets of decisions. First, we agreed that the five components of IOOV tell a complete and compelling story describing the challenges of mental illness, efforts to overcome these challenges, and successes thereafter. In addition, we agreed that the guided discussion was essential to personalizing its effects. Hence, coders concurred to not exclude components or discussions. Instead, pedagogic changes were recommended. Focus group members thought there was significant overlap in stories so that only one of the two facilitators is needed. To further streamline our revised approach, we omitted the videotapes. Hence, the 30 min IOOV had one facilitator talk about her or his experiences across the five IOOV components and then lead relevant group discussion. One of the strengths of the IOOV-90 is an already existing manual and fidelity measure, plus training program to help a prospective facilitator learn and master the program laid out in the manual and meet fidelity criteria. We decided that development of a 30 min IOOV should largely focus on cutting discrete aspects of the 90 min version. In this way, development would entail cleanly omitting chapters of the manual and corresponding training program and fidelity test. Education was the comparison group in this study. Participants in this condition were taught about the misunderstandings of mental illness and corresponding facts. For example, one misunderstanding is the notion that people with mental illness are unable to recover. Fact is that a large number of people with serious mental illness are able to get a job and live independently. The education condition was presented as a 30 min power point presentation with provocative questions meant to stimulate discussion at the end. Individuals randomized to all three conditions participated in groups of 6–15. Memory Change We used the Life Story Memory Test (LSMT), a measure designed by Macrae and colleagues (Johnston and Macrae 1994; Macrae et al. 1994), to assess the impact of stigma changing strategies on the perception and recollection of persons with severe mental illness. The LSMT has subsequently been used in measuring stigma change after participating in anti-stigma programs targeting mental illness stigma (Corrigan et al. 2001). Research participants viewed two videotapes, each about 3 min long, of an actor labeled ‘‘mentally ill’’ who is telling his or her life story. Actors for the two separate vignettes were male and female. The
123
Community Ment Health J (2010) 46:517–522
narrative contained 20 items, 10 that were reliably rated by a pilot group (n = 29) as negative and stereotypic statements about mental illness (e.g., ‘‘Sometimes I believe I’m George Washington.’’) and 10 that were rated as positive statements (e.g., ‘‘I work as an engineer.’’). These items were randomly ordered and then written into a coherent narrative that the actor read on the videotape. One minute after viewing each videotape (during which time participants were instructed to complete an interference task: to draw a map of their childhood home or elementary school), participants were asked to write down as many of the statements as they could remember. Raters then counted the number of positive and negative statements in each participant’s list.
Results Descriptive statistics of research participants are summarized in Table 1. Overall, age, education, and marital status are as expected for relatively young adults in college. The sample did not have the kind of 50–50 split in gender one might expect of college students. A more varied range of values are evident for ethnicity and income. Table 1 summarizes demographics by anti-stigma intervention. No differences were found across groups for all but ethnicity, with the P values higher than .30. To then analyze ethnicity effects, the sample was broken into white (n = 161) and non white (n = 39). Results of chi-squared analyses examining ethnicity by condition found significant differences (P \ .05). We examined whether this interaction affected memory scores across condition and found significant interactions for some memory indices. But it was difficult to infer meaning from these findings because total nonwhites per condition was quite small (n = 6) for the 30 min IOOV and education interventions. Hence, we did not break out ethnicity effects in the remainder of the analyses. Table 1 also includes fidelity scores for each condition. Fidelity was determined by an independent rater who sat inconspicuously in the intervention room and checked off behaviors that comprise the intervention as they were demonstrated by the session facilitator. The fidelity score was the ratio of number of demonstrated behaviors and total composite behaviors. The Table 1 score represents fidelity as averaged across cohorts per anti-stigma intervention. A chi-squared test did not yield significant differences in fidelity across the three conditions (P [ .50). Negative and Positive Perceptions and Recollections Positive and negative perception recollection scores are summarized by condition in Fig. 1. Independent raters
Community Ment Health J (2010) 46:517–522
521
coded positive and negative recollections separately for the male and female vignette. Inter-rater reliability for negative and positive responses representing these codes was quite high and varied from .89 to .97. Previous research has failed to show responses differ by the gender of the actor which our data mirrored (P-values of main and interaction effects for gender all greater than .25). Hence, we collapsed perceptions and recollections across actor gender. The top histogram in Fig. 1 shows findings in terms of negative and positive recollections. Subsequent 3 (condition) by 2 (positive or negative recollection) mixed model ANOVAs showed significant differences between positive and negative statements with positive statements higher across the three conditions (F(1,181) = 318.1, P \ .001). Research participants are remembering more positives about a person with mental illness compared to negatives. The interaction between recollections and condition was not significant (F(2,181) = 6.98, N.S.) but there was a significant main effect for condition (F(2,181) = 10.61, P \ .001). Oneway ANOVAs were then conducted to examine positive or negative recollections across the three anti-stigma interventions. Significant differences across conditions were found for positive recollections (F(2,181) = 4.84, P \ .01) and negative recollections (F(2,181) = 13.1, P \ .001). Post hoc Tukey’s tests are summarized under the top histogram in Fig. 1. Research
positive statements negative statements
9.8 8.8 7.8 6.8 5.8 4.8 3.8 IOOV30
IOOV90
Educ
Post hoc tests Positive Statements (IOOV30 = Educ) > IOOV90 Negative Statements (IOOV30 = IOOV90) < Educ)
Memory Scores Represented by Ratios ratios
0.3
0.25
0.2
0.15 IOOV 30
IOOV 90 Post hoc tests
Education
(IOOV30 = IOOV90) > Educ
Fig. 1 Recall of positive and negative statements
participants in the IOOV-30 and education conditions recalled more positives than those in the IOOV-90 condition. Those in the education condition remembered significantly more negative statements than those in either of the two IOOV conditions. In some ways, these findings seem contrary; that education does worst on negative recollections but also best on positive recollections. One reason for this pattern may be education group participants were overall generating more responses—be they positive or negative—compared to the other groups. For this reason, we converted positive and negative recollection scores to ratios, the difference between positive and negative scores divided by total responses. These data are summarized in the histogram at the bottom of Fig. 1. Significant difference in ratios were found across conditions (F(2,181) = 5.7, P \ .01). Post hoc Tukey’s tests showed people completing the education condition had lower ratios than those in the two IOOV conditions. No significant difference was found in the 30 versus the 90-min IOOV interventions.
Discussion This study sought to contrast contact versus education approaches to stigma change using perception and recollection of actor statements as an outcome variable. Results showed people completing the education condition recalled significantly more negative statements compared to those in the two IOOV conditions. We need to be especially precise here because we may be conflating reduction of negative statements with stigma change. In fact, changing negative statements is not altering the stuff of mental illness stigma, not the content of stereotypes. Rather, findings in this study showed IOOV conditions yield better changes in cognitive processes, events which lead to the creation and maintenance of stereotypes. Results also suggested that people in the education group recalled more positives than in the IOOV-90 group. Data were transformed into difference ratios in order to control for an over-responding effect; namely, that these findings occurred because participants from the education condition generated more overall responses. The two IOOV conditions yielded significantly better ratios than education. Hence, findings from this study parallel those in previous research (Corrigan et al. 2001); namely, contact yields better impact on perceptions and recollections than education. Another goal of this study was to compare the effects of a 30 min IOOV to the 90 min original. Using focus group feedback, IOOV-90 was reduced to 30 min by omitting the videotape and one of the two speakers. Viewing IOOV-90 as the standard, we thought IOOV-30 would yield better anti-stigma effects than education, but not significantly
123
522
differ from IOOV-90. This was mostly found except for positive recollections where research participants in IOOV30 actually performed better than the 90 min condition. Hence, evidence shows that the IOOV-30 version has similar, or even better, effects on perception and recollection. This might represent a fatigue effect after the 90 min IOOV. For example, research participants in the IOOV-90 provided less positive and negative responses overall. We proffered the memory task as a more dynamic and less biased measure of stigma and stigma change compared to attitude measures. Significant differences found in this paper suggest IOOV and contact conditions yield a more plastic impact with subsequent effects on the public stigma of mental illness. Unfortunately, data from this study do not reflect the nature of this dynamism. We did not assess recollection over time but only at post-test and are therefore unable to report on stability findings. We did not assess whether enhanced perceptions and recollections moderate stereotypes and discriminatory behaviors. Impact variables in the study represent attitudinal responses to videotaped vignettes which is not a direct representation of cognitive or behavioral response to real interactions with people with mental illness. These are all important directions for future research. Representativeness of the sample is a second limitation of the study. Our findings are based on a college student population and the restrictions samples like these entail. One solution is seeking a stratified and broader sample. More importantly however, is examining stigma change on targeted samples, people from roles that have significant impact on the lives of people with serious mental illness; e.g., employers, landlords, and police officers. Although this is more difficult research to conduct, it shows how enhanced perception and recollection influence, for example, the hiring behaviors of employers or renting activities of landlords. Findings from this study have important implications for stigma change. The results add to the body of knowledge examining the efficacy of education versus contact for changing stigma. In addition, these conclusions add to the foundation of effectiveness research in the stigma arena. The study provided information about the impact of IOOV, a program that is being widely disseminated by NAMI. Our paper also examined feasibility concerns, demonstrating that efforts to make interventions more efficient do not hamper their effects.
123
Community Ment Health J (2010) 46:517–522 Disclosure of Interest and Funding Patricia Doyle is an employee of the National Alliance on Mental Illness (NAMI) in DuPage County, and Sarah O’Brien is an employee of NAMI National. Nicolas Ru¨sch was supported by a Marie Curie Outgoing International Fellowship of the European Union.
References Corrigan, P. W., & Penn, D. (1999). Lessons from social psychology on discrediting psychiatric stigma. American Psychologist, 54, 765–776. Corrigan, P. W., River, L., Lundin, R. K., Penn, D. L., UphoffWasowski, K., Campion, J., et al. (2001). Three strategies for changing attributions about severe mental illness. Schizophrenia Bulletin, 27, 187–195. Corrigan, P. W., Rowan, D., Green, A., Lundin, R., River, P., UphoffWasowski, K., et al. (2002). Challenging two mental illness stigmas: Personal responsibility and dangerousness. Schizophrenia Bulletin, 28, 293–310. Hinshaw, S. P. (2007). The mark of shame: Stigma of mental illness and an agenda for change. Oxford: Oxford University Press. Johnston, L. C., & Macrae, C. N. (1994). Changing social stereotypes: The case of the information seeker. European Journal of Social Psychology, 24, 581–592. Lincoln, T. M., Arens, E., Berger, C., & Rief, W. (2008). Can antistigma campaigns be improved? A test of the impact of biogenetic versus psychosocial causal explanations on implicit and explicit attitudes to schizophrenia. Schizophrenia Bulletin, 34, 984–994. Link, B. G., & Phelan, J. C. (2001). Conceptualizing stigma. Annual Review of Sociology, 27, 363–385. Macrae, C. N., Bodenhausen, G. V., Milne, A. B., & Jetten, J. (1994). Out of mind but back in sight: Stereotypes on the rebound. Journal of Personality and Social Psychology, 67, 808–817. Rusch, L. C., Kanter, J. W., Angelone, A. F., & Ridley, R. C. (2008). The impact of In Our Own Voice on stigma. American Journal of Psychiatric Rehabilitation, 11, 373–389. Ru¨sch, N., Corrigan, P. W., Todd, A. R., & Bodenhausen, G. V. (2010a). Implicit self-stigma in people with mental illness. Journal of Nervous and Mental Disease. Ru¨sch, N., Todd, A. R., Bodenhausen, G. V., & Corrigan, P. W. (2010b). Biogenetic models of psychopathology, implicit guilt and mental illness stigma. Psychiatry Research. Ru¨sch, N., Todd, A. R., Bodenhausen, G. V., Olschewski, M., & Corrigan, P. W. (2010c). Automatically activated shame reactions and perceived legitimacy of discrimination: A longitudinal study among people with mental illness. Journal of Behavior Therapy and Experimental Psychiatry, 41, 60–63. Thornicroft, G. (2006). Shunned: Discrimination against people with mental illness. Oxford: Oxford University Press. Wood, A. L., & Wahl, F. (2006). Evaluating the effectiveness of a consumer-provided mental health recovery education presentation. Psychiatric Rehabilitation Journal, 30, 46–53.
