Attitudes of Sheltered Care Residents Toward Others With Mental Illness

Abstract

The social acceptance expressed by 234 former mental patients and by the general public toward persons with serious mental illness was compared. Factors that may affect social acceptance of such persons, including personal characteristics and experiences that promote identification with mentally ill persons and the subject’s level of psychological distress, were examined. Former patients expressed attitudes that were much more accepting than those of the general public. As hypothesized, individual characteristics and experiences likely to increase former patients’ identification with their peers (that is, visible deviant appearance, a high level of involvement within the sheltered care community, and the experience of negative community reactions to a resident’s facility) were related to a higher score on a Guttman scale of social acceptance. An increased level of self-reported psychological distress tended to moderate such supportive attitudes.

Numerous studies have documented the caution and fear that people in the general population feel toward persons with mental illness (1,2). Events such as the intense public reaction during the last presidential campaign to Michael Dukakis’ use of psychiatric help or the often vociferous community objections to siting of residential care facilities attest to the continuation of these fears and concerns (3).

Do former patients share these perceptions? What attitudes do they have toward other mentally ill persons? Do former patients mirror the generally negative images, or do they perceive other patients more positively? The topic is important because of the key role that social support plays in the adaptation of former patients to community life (4–6). Because mental health service providers have come to expect that an increasing amount of this support will come from other former patients (7), it is crucial that we learn more about former patients’ attitudes toward one another.

This paper reports the results of a study in which the attitudes of persons with serious mental illness toward one another were examined. Their attitudes were compared to those of the general population. In addition, we examined two factors that may affect former patients’ attitudes—individual characteristics and experiences likely to promote identification with the mentally ill and self-reported psychological distress.

Experiences and individual characteristics likely to promote identification with, and thus more positive attitudes toward, others who are mentally ill include the direct experience of discrimination, personal characteristics that increase the probability of being a target of such discrimination, and a social life dependent on one’s status as a former patient. People who choose or are forced to interact together on the basis of some shared characteristic or experience tend to develop a mutually supportive identity based on that characteristic or experience (8).

Previous research demonstrated that mentally ill sheltered care residents who reported greater psychological distress also reported more limited access to and participation in social relations (9). We therefore expected that former patients’ increased psychological distress would be associated with increased rejection of opportunities for social interaction as part of a general tendency to withdraw from such interaction. We did not negate the possibility that previous findings of reduced interaction by more disturbed people were due to their rejection by others. We simply hypothesized that part of this reduced interaction may have resulted from a tendency to withdraw from interaction, to become less able to engage with others and to cope with interpersonal relationships and thus to become less accepting of others as one’s level of psychological distress increases.

Acceptance by others is an important element of a person’s social support system. Characteristics and experiences of former patients that promote identification with mentally ill persons and increases in psychological distress are likely to act as countervailing influences on the degree of acceptance of former patients by their peers. By determining the extent of these influences, we assessed one important aspect of how socially supportive this population can be of their fellow users of mental health services.

Two hypotheses were tested in this investigation. First, it was expected that, compared with the general population, former mental patients would be more accepting of other former patients. Second, the amount of acceptance or rejection was expected to be determined by personal characteristics and experiences that promote identification with the mentally ill and by psychological distress. The former was expected to lead to a more positive view of former patients and the latter to more negative attitudes.

Methods

Sample.

This study examined the ten-year experience of a probability sample of 393 residents of 211 California sheltered care facilities (board- and-care homes, family care, halfway houses, and psychosocial rehabilitation facilities). Initial interviews with the residents were conducted in the summer of 1973, and follow-up interviews were completed between 1983 and 1985. The 1973 data provided information on the characteristics and experiences of residents that were likely to influence their attitudes assessed at follow-up. The original sample was representative of all former psychiatric inpatients between the ages of 18 and 65 who in 1973 were living in sheltered care facilities in California that provided supportive living arrangements to residents who qualified for the higher rate under the Aid to the Totally Disabled Program (now Supplemental Security Income [SSI]).

Between 1983 and 1985, a total of 360 sample members (91.6 percent) were relocated, 90 (23 percent) of whom had died. Of the 270 living sample members, 253 (93.7 percent) were reinterviewed. The study reported here focused solely on the experiences of the 234 individuals who completed valid interviews both in 1973 and at follow-up. The survey methodology used in the 1973 study and at follow-up has been described elsewhere (10,11).

Analysis.

To test the first hypothesis, we compared responses to standard questions about social acceptance by sample members at follow-up with responses by the general public. Social acceptance scales have a long tradition in survey research on attitudes (12,13), particularly in measuring prejudice toward racial groups. During the 1960s, several studies measured the social acceptance of mentally ill persons by the general public (2,14,15). These studies frequently used the same social acceptance items.

In the study reported here, five frequently used questions about social acceptance (Table 1) were used as the basis for comparing the attitudes of the general public and the sample toward persons with mental illness. Before the questions were asked, the interviewer stated, “I would like you to think now of a former patient who is just being discharged from the mental hospital. He or she has been very ill and has been in the hospital for quite a long time.”

Table 1.

Percentages of persons in a sample of 234 former mental patients and in various general population samples (average percentages) who responded positively to questions about social acceptance of mentally ill persons

Question	Sample
	Former patients	General population1
Would you be willing to allow your child to marry a former patient?	57	31
If this person lived next door and you needed a babysitter, do you think you would ask him or her to babysit?	47	32
If you had a room to rent in your house, would you be willing to rent it to him or her?	67	55
Would you be willing to work in a regular job with him or her?	75	82
Would you be willing to have him or her as a member in your favorite club or organization?	81	79

¹Responses were averaged from samples indicated by references numbers:

Question 1: 16–18, 20–22, 24–27

Question 2: 13, 19 (two samples)

Question 3: 14, 17, 18, 20–22, 24, 26, 27, 29

Question 4: 14, 16, 18, 20–23, 25–29

Question 5: 18, 19 (two samples), 23, 28

We compared the percentage of our subjects who responded positively with the average percentage of positive responses to these questions by members of the general population as reported for 17 general population samples (13,14,16–29).

To test the second hypothesis, the study used an ordinary least-squares regression model. The dependent variable was our sample member’s total score on a social acceptance scale derived from the follow-up data. We constructed a five-item Guttman scale of social acceptance from a pool of ten items (Table 2). Guttman scaling makes use of a series of items that measure increasing tolerance for a particular behavior. The Guttman scale has two particular properties—unidimensionality and cumulativeness (30). Unidimensionality means that the scale measures a single concept—for example, acceptance. Each scale item is given a score of 1. In Guttman scaling, cumulativeness means that a respondent’s total score indicates that he or she has answered positively all items reflecting lesser acceptance. Thus higher scores on this scale indicated more accepting attitudes.

Table 2.

Percentages of a sample of 234 former mental patients who responded positively to questions about social acceptance of mentally ill persons

Question	N	%
Would you be willing to have a former patient as a member in your favorite club or organization?	182	80.8
If you owned a house or an apartment next door to you, would you be willing to rent it to him or her?1	181	77.9
Would you be willing to work with someone like this as a partner in a project at school or in the neighborhood?	184	76.6
Would you be willing to work in a regular job with him or her?1	187	75.4
If you owned a small store and you need to hire someone, would you hire this person?	174	67.2
If you had a room to rent in your house, would you be willing to rent it to him her?1	179	66.5
If you were working for him or her, do you think he or she would be a good boss?1	158	63.3
Would you be willing to allow your child to marry someone like this?2	161	57.1
If she or he were running for a local public office (for example city council), do you think you would vote for him or her?1	160	48.1
If this person lived next door and you needed a babysitter, do you think you would ask him or her to babysit?	169	46.1

¹Items used in the formation of a Guttman scale of social acceptance

²Wording modified to reflect a positive response direction.

Only items that produced the least number of errors in ordering a respondent along the acceptance dimension were included in the final scale. To avoid ordering errors, only one of two items measuring the same degree of acceptance was likely to be included in the scale. In this case, if the respondent agreed with an item that assumed a certain level of acceptance, such as living next door to someone, then logically he or she would have also agreed with items that assumed less social acceptance, such as working in the same city with someone.

In order of decreasing acceptance, the five items that formed the Guttman scale were as follows: “If you owned a house or an apartment next door to you, would you be willing to rent to a former patient?” “Would you be willing to work in a regular job with him or her?” “If you had a room in your house, would you be willing to rent it to him or her?” “If you were working for him or her, do you think he or she would be a good boss?” “If he or she were running for a local public office (for example, city council), do you think you would vote for him or her?”

The five items had the fewest ordering errors and formed a Guttman scale with a coefficient of reproducibility of .91 and a minimum marginal reproducibility of .67. Coefficient alpha for this scale was .81.

Each person was assigned a score representing the item he or she agreed with that reflected the greatest degree of acceptance, regardless of whether items representing less acceptance were agreed with. This strategy assumed a perfect Guttman scale; in fact, only 28 percent (N = 55) of the respondents did not fit a perfect pattern, and 66 percent of these (N = 36) missed by only one item. For example, a person who scored 5 because he or she agreed to vote for a former patient for public office agreed with all but one of the statements that implied less accepting attitudes.

The independent variables, taken from 1973 data, included four indicators of factors thought to increase the likelihood of identification with mentally ill persons as well as an assessment of psychological distress. They are described below.

Factors likely to increase identification.

Experiences and individual characteristics that were likely to have confined a former patient’s interactions within the group of former patients and that made his or her life situation dependent on the status of former patient were taken as indicators of experiences likely to increase identification with mentally ill persons by the mechanism of subculture formation and isolation (8). Behaviors that indicated an increased focus of one’s life within the sheltered care facility were thought especially to increase identification.

The first measure, visible deviant appearance, was based on the interviewer’s impression of the respondent at the first meeting in 1973. The interviewer chose one of three statements to describe the respondent’s appearance: “Having seen him or her in any other place, I would have no idea that there was anything unusual about him or her,” “The respondent looks a little odd but not noticeably so,” and “The respondent looks disturbingly odd.” This measure was chosen because it was believed to be an indicator of the amount of isolation a person would experience as a result of an odd appearance or deviant persona. Interviewer instructions defined a “disturbingly odd person” as one who “if observed in the aisle of a supermarket would cause you to choose a different aisle in which to shop so as to avoid the possible encounter.” Categories were collapsed to “not odd” and “very odd.”

The second indicator, community reaction to facilities and residents, was measured by facility managers’ reports of having received complaints from neighbors. Segal and Aviram (9) reported that complaints, while not specifically defined, were associated with reductions in independent social functioning of residents outside the sheltered care facility. Thus these nonspecific complaints seem to have the effect of limiting or confining former patients’ interactions to their sheltered care environment. For this reason, whether or not managers received complaints from neighbors was used as an indicator of identification with mentally ill persons.

The Internal Social Integration Scale, a third indicator of identification with mentally ill persons, measured the extent to which the resident was involved in the life of the sheltered care facility and had his or her life situation assisted by service providers. The scale thus measured the extent to which the former patient’s life situation was dependent on his or her disability status. The scale has five subscales that measure the extent to which a resident has access to community resources with transportation provided by the facility operator, has access to activities sponsored by the operator (for example, social activities, vocational training, and religious services), has access to basic necessities (for example, laundry service and toilet supplies) at the house, socializes with other residents, and purchases laundry services, clothing, and other basics at the house. Higher scores indicated a greater extent of involvement in activities either within or sponsored by the residential care facility. Further information concerning the development of this scale is published elsewhere (9).

The fourth indicator of identification was the question “How often do you join in the activities of social or political groups outside the house for people who are considered former patients?” The five scaled-response categories ranged from “very often” to “never.” For the analysis, categories were collapsed to “rarely” and “often.”

Measurement of psychological distress.

Psychological distress was measured by the Langner 22-Item Stress Symptom Scale, originally developed in the Midtown Manhattan Study (31). This scale is generally perceived as a measure of self-reported distress (32).

Results

Table 1 compares the responses of former patients to their peers with the responses of the general public to former patients. Compared with the general public, former patients were more willing to have more personal contact with other former patients. The level of acceptance of former patients by the two groups differed less for public activities, such as work and participation in community organizations, than for personal ones, such as marriage.

Table 2 presents ten questions traditionally used as social acceptance indicators and the percentage of the former patients in our sample who gave positive responses to these questions. (The questions were prefaced by a similar interviewer statement: “I would like you to think now of another former patient who is just being discharged from the mental hospital. He or she has been very ill and has been in the hospital for quite a long time.”)

The respondents’ mean±SD score on the Guttman scale was 3.34±1.78, indicating that they would be willing to rent a room in their house to a former patient but did not think that a former patient would make a good boss.

A series of analyses tested for significant relationships between demographic variables and other variables of interest and the dependent variable, level of social acceptance as measured by the Guttman scale. Several variables were considered: age, gender, marital status, race or ethnicity, education, number of times in a psychiatric hospital, total length of time in a psychiatric hospital, and whether the subject had experienced two years or more of continuous psychiatric hospitalization. None of these variables were significantly related to the level of social acceptance. In addition, where people lived at follow-up, that is, in sheltered care, in an institution, or in the community, also was unrelated to social acceptance, Since these variables had neither theoretical nor bivariate statistically significant associations with social acceptance, they were not included in the proposed multivariate model.

The results of the ordinary least-squares regression that represents the multivariate model of social acceptance are shown in Table 3. The analysis showed that identification-enhancing experiences, which were represented by independent variables in the model, generally tended to be associated with a greater acceptance of peers. Contrary to the stated hypothesis, independent participation in social and political groups for former patients led to less accepting attitudes. As hypothesized, psychological distress was associated with less acceptance. The model was significant (F = 5.71, df = 5,190, p = .0001) and explained 13 percent of the variance in the outcome measure.

Table 3.

Ordinary least-squares regression of factors related to social acceptance of mentally ill persons by a sample of former mental patients¹

Variable	Unstandardized coefficient	standardized coefficient	t	p2
Indicators of identification with mentally ill persons
Visible deviant appearance (not odd = 1, vey odd = 2)	1.14	.18	.2.58	.01
Complaints from neighbors about sheltered care facility (no = 1, yes = 20)	.82	.20	2.95	.00
Internet Social Integration Scale score in 19733	.26	.14	1.96	.05
Participation in social or political groups for former patients (rarely = 1, often = 2)	−1.07	−.16	−2.30	.02
Measure of psychological distress: Langner 22-item scale score4	−.08	−.20	−2.86	.00

¹Complete data were available for 196 subjects (R² = .13, F = 5.71, df = 5, 190, p = .0001).

²Significance levels are based on two-tailed tests.

³A higher score indicates greater involvement in sheltered care facility activities.

⁴A higher score indicates increased psychological distress.

Discussion

Former patients were highly accepting of physical and social proximity to other former patients. They were especially willing to tolerate more personal involvements with former patients than was the general public. Thus they warrant the faith mental health professionals have in their potential as a source of social support for other former patients (33).

Three phenomena seemed to be at work in the increased acceptance of former patients by one another. First, people who were more likely to be identified as former patients or to be involved in activities of former patients that were isolated from the larger community were more accepting of persons with mental illness. Those who had a visible deviant appearance or whose activities were more focused within the sheltered care setting and who received more assistance from mental health service providers were more likely to be accepting of other former patients.

These results appear to confirm the hypothesis that mutually supportive attitudes develop within a subgroup formed around a particular shared characteristic—especially when this characteristic leads to increased isolation from the general community. This outsider hypothesis (8) is strengthened by the observation that people living in facilities that had received complaints from their neighbors were also likely to be more accepting of other former mental patients.

The second phenomenon was the impact of psychological distress on personal social functioning and, subsequently, on residents’ levels of social acceptance. Respondents who reported higher levels of psychological distress also reported less accepting attitudes. This lower level of acceptance may result from the respondents’ withdrawal from social interaction and their reduced ability to engage with others and to cope with interpersonal relationships. Consequently, they may become less accepting of others.

The third and perhaps most interesting phenomenon was the finding that individuals who independently join social and political groups for former patients independent of their residential care facilities were less likely to be accepting of former patients. As this finding was unexpected and contradicted the stated hypothesis, the authors considered an alternative hypothesis—that is, that participation in these groups might be a normalizing experience for sheltered care residents who then move toward identification with persons who are not mentally ill and away from identification with lower-functioning and stigmatized former patients. It is possible that this shift in identification is accompanied by a shift in attitudes toward persons with mental illness.

Crocker and Major (34) discussed ways that members of stigmatized groups protect their self-esteem. In this paper, we propose that persons with mental illness who are more successful functionally are the ones who become regular members of groups for former patients. These individuals may need to distance themselves from their less successful peers as a mechanism for enhancing their own self-esteem. Frequent participants in groups for former patients may therefore adopt attitudes toward their peers that are more similar to those held by the general public.

To partially validate this hypothesis, it was necessary to determine whether frequent participants in groups for former patients were higher functioning socially than those who participated rarely. The External Social Integration Scale used in an earlier study (9) measures the extent to which residents are involved independently in the life of the community-at-large (independent of the facility).

A t test was used to determine if there was a significant difference in the level of external social integration between sample members who rarely participated in “social or political groups outside the house” for former patients and those who often participated in such groups. The results indicated that those who participated often had significantly higher scores on the External Social Integration Scale than those who participated rarely (t = 2.26, df = 228, p = .02).

These findings offer some support for the alternative hypothesis that individuals who frequently participate in groups for former patients have higher levels of external social integration. Such participants also may identify less with lower-functioning sheltered care residents and therefore become less accepting of former patients in general as a means of enhancing their self-esteem. However, we did not test the latter processes stated in this alternative hypothesis. These processes should be the subject of future research on groups for former patients.

Conclusions

In an effort to promote better community care, mental health professionals need to consider how the concept of social acceptance may be critical to the potential for peer support. Mutual support among former patients holds incredible promise, and therefore it is crucial that we learn more about former patients’ attitudes toward one another. The social support potential of former patients clearly is established by our data; former patients express very accepting attitudes toward their peers, attitudes much more accepting than those expressed by the general public. However, the role of increased psychological distress and the possible effect of increased identification with non-mentally-ill persons in moderating such supportive attitudes should be a major concern in developing mutual-support programs.

The potential that former patients have for providing social support is clearly established. Former patients express much more accepting attitudes toward their peers than does the general public.