You are on page 1of 12

Swedish attitudes towards persons with

mental illness
TORBJÖRN HÖGBERG, ANNABELLA MAGNUSSON, KIM LÜTZÉN,
BÉATRICE EWALDS-KVIST

Högberg T, Magnusson A, Lützén K, Ewalds-Kvist B. Swedish attitudes towards persons with


mental illness. Nord J Psychiatry 2012;66:86–96.

Background: Negative and stigmatizing attitudes towards persons with mental illness must be
dealt with to facilitate the sufferers’ social acceptance. Aim: The present study aimed at survey
Swedish attitudes towards persons with mental illness related to factors impacting these atti-
tudes. Material and Methods: New CAMI-S based on the questionnaire “Community Attitudes
to Mental Illness in Sweden” ([CAMI] Taylor & Dear, 1981) was developed with nine behavio-
ral–intention items and thus comprised a total of 29 items. Of 5000 Swedish people, 2391
agreed to complete the questionnaire. Principal component analysis rendered four factors reflect-
ing attitudes towards the mentally ill: Intention to Interact, Fearful and Avoidant, Open-minded
and Pro-Integration, as well as Community Mental Health Ideology. The factors were analyzed
for trends in attitudes. By MANOVA, the experience of mental illness effects on mind-set
towards the sufferers was assessed. By means of logistic regression, demographic factors con-
tributing to positive attitudes towards persons with mental illness residing in the neighborhood
were assessed. Results: By New CAMI-S, the Swedish attitudes towards the mentally ill were
surveyed and trends in agreement with living next to a person with mental illness were revealed
in three out of four factors derived by principal component analysis. Aspects impacting the
Swedish attitudes towards persons with mental illness and willingness to have him/her residing
in the neighborhood comprised experience of mental illness, female gender, age (31–50 years),
born in Scandinavia or outside Europe, only 9 years of compulsory school and accommodation
in flat. Conclusion: The New CAMI-S came out as a useful tool to screen Swedish attitudes
towards persons with mental illness. Most Swedes were prepared to live next to the mentally ill.
• Mental illness, New CAMI-S, Swedish attitudes towards persons with mental illness.

Torbjörn Högberg, Karolinska Institutet, Institution of Clinical Neuroscience, Stockholm Centre


for Psychiatric Research and Education, Stockholm, Sweden. E-mail: torbjorn.hogberg@sll.se;
Accepted 27 May 2011.

A ll people are of equal value and have equal rights.


Yet, a substantial part of the Swedish population
perceives persons with mental illness as unpredictable
attitudes positively towards persons with mental illness
and increase the general awareness about mental disor-
ders (10). Consequently, negative and stigmatizing atti-
and dangerous (1–3). Namely, a person with a mental ill- tudes towards persons with mental illness must be dealt
ness is portrayed as 10 times more likely to be a violent with to facilitate the sufferers’ social acceptance to hin-
criminal than a mentally healthy person in prime-time der their marginalization and stigmatizing (11–13).
television. As a result, viewers considered locating men- Stigma—“a stick” or a “mark” (gr.) originally denot-
tal health services in residential neighborhoods as endan- ing an unusual or defamatory sign in a person’s moral
gering the residents and were less likely to support living character—is nowadays referred to as a social construct
next to persons with mental illness (4). In contrast, meta- comprising four interrelated components: 1) people dis-
analysis indicated that most violent persons were not tinguish and label human differences; 2) labeled persons
psychotic and most people with a psychotic illness were are caused damage by cultural beliefs and negative ste-
not violent (5). Also culture shapes public attitudes reotypes; 3) labeled persons are placed in distinct cate-
towards mental illness (6) as well as lack of knowledge gories to separate “us” from “them” and 4) labeled
about mental disorders (7–9). Therefore, through infor- persons experience status loss and discrimination (14,
mation campaigns, the Agency for Disability Policy 15). Stigmatization comprises both acknowledgement of
Coordination has sought to influence the public’s individual differences based on specific characteristics

© 2012 Informa Healthcare DOI: 10.3109/08039488.2011.596947


SWEDISH ATTITUDES TOWARDS PERSONS WITH MENTAL ILLNESS

and a steadily continuing belittling of a person (16). A respondents had finished a 9-year compulsory school,
stigmatized socially rejected person with mental illness 36.6% (n ⫽ 833) completed upper secondary school,
enlarges his own feelings of alienation. In other words, 27.5% (n ⫽ 625) had a university degree but 14.9%
negative stereotypes are transposed to “own self ” caus- (n ⫽ 339) were subjected to other schooling. With refer-
ing internalization of stigma leading to degradation of ence to respondents’ (n ⫽ 2391) experience of mental ill-
the self and feelings of shame combined with lower self- ness in general, 57.7% (n ⫽ 1331) had no such
esteem and poorer self-confidence as well as with an experience. However, 3% (n ⫽ 72) had experience of their
inferior self-image (17). own mental illness, 28.6% (683) had experienced rela-
Educated people are presumed to display more posi- tive’s, friend’s or other’s mental illness, 8% (n ⫽ 192) had
tive attitudes towards persons with mental illness, occupational experience of such illness and 4.7 (n ⫽ 113)
although the “not in my backyard” (NIMBY)-phenome- had experiences of mental disorders in varying ways.
non might be at hand, i.e. knowledgeable persons do not
necessarily want to live next to them (18, 19). Further- Internal reliability of the New CAMI-S instument
more, personal experience of mental illness is presumed The “New CAMI-S” instrument was an improvement of
to affect intolerance towards mental disorders (20–22). In “CAMI-S” (Community Attitudes to Mental Illness in Swe-
addition, negative attitudes towards persons with mental den; 22) and Cronbach’s alpha of the “New CAMI-S ques-
illness links to older age, lower standard of living and tionnaire” was computed (α ⫽ 0.954). Items numbered 19,
lower education (19, 23). Moreover, the concept “atti- 21, 23, 27, 29, 31, 34, 36, 41, 42 and 45 were negatively
tude”, is tripartite: cognitive, affective and behavioral, i.e. worded and their scoring was therefore reversed. The behav-
the cognitive part includes beliefs, the affective part com- ioral–intention items were numbered: 19, 22, 25, 28, 31, 34,
prises emotions, and the behavioral part covers actions or 37, 40 and 43 (Table 2). In view of the fact that all loadings
intention to act or interact. Presently, a special focus is exceeded 0.44, no item was excluded. A principal compo-
placed on the behavioral part, i.e. on the intention to nent analysis was carried out on the 29 items with the Vari-
interact with the mentally ill, more correctly, on the max rotation method, applying the Kaiser rule to drop all
willingness to live next to persons with mental illness components with eigenvalues under 1.0. After a varimax
(24–26). Presently the definition of persons with mental rotation, each factor was presumed to have either large or
illness includes mental dysfunctions requiring long-term small loadings of any particular variable and thus yielded
treatment (27). results to make it easy to identify each variable with a sin-
gle factor. The eigenvalues for the four factors were: 12.90,
1.62, 1.35 and 1.20. Inspection of the factors’ underlying
Aim themes brought about the following names: 1) Intention to
The present study aims at screening Swedish people’s Interact, 2) Fearful and Avoidant, 3) Open-minded and Pro-
attitudes and to cluster recurrent themes in these mind- Integration and finally 4) Community Mental Health Ideol-
sets towards persons with mental illness related to per- ogy. These factors are largely consistent with those found by
sonal experience of mental illness and to demographic Högberg et al. (20) and Rudder-Baker (22). The principal
factors. component analysis for the 29 items and Cronbach’s alpha
for each factors are shown in Table 2.

Material and Methods


Demographics Statistical analysis
A drop-out analysis is given in Table 1. Furthermore, The results were computed by SPSS, versions 15 and 17,
Table 1 presents demographic factors: it elucidates that as follows: Principal Component Analysis was carried out
the respondents’ age and income approximated the nor- with the Varimax rotation method applying the Kaiser rule
mal distribution fairly well. Just over a quarter (25.9%) to drop all components with Eigenvalues under 1.0. Also
of the respondents had an annual income between 160,000 Cronbach’s Alpha, Logistic regression with analysis of
to 235,000 SEK, and slightly more than 28.6% resided in Maximum Likelihood estimates and Odds Ratio estimates
major cities. Most of the respondents were married or were computed. MANOVA, partial Eta squared (η2), χ2
cohabiting but a third of the cohabiting couples lived as well as Wallis kruskal-Wallis one-way analysis of vari-
at separate places. A total of 13% of the respondents’ ance by ranks analysis of variances were carried out. Also
were born in other countries and 4% did not possess a Jonckheere trend tests were calculated where the alterna-
Swedish citizenship. tive hypothesis is tested against a null hypothesis of no
Regarding respondents’ accommodation, 57.7% systematic trend across treatments. The test can be applied
(n ⫽ 1336) lived in own house as compared with 39.9 for data for k independent samples, when measurement is
(n ⫽ 923) who resided in a flat or lived under other con- at least ordinal, and when it is possible to specify a priori
ditions (2.4% [n ⫽ 56]). Altogether 21% (n ⫽ 477) of the the ordering of the groups (36).

NORD J PSYCHIATRY·VOL 66 NO 2·2012 87


T. HÖGBERG ET AL.

Table 1. Drop-out analysis: Respondents’, non-respondents’ as well as sample’s characteristics.


Respondents Non-respondents Sample

n % n % n %

Gender
Men 1037 43.4 1337 51.5 2374 47.6
Women 1354 56.6 1257 48.5 2611 52.4
Age classes (10 years.)
⬍ 19 years 73 3.1 100 3.9 173 3.5
20–29 290 12.1 498 19.2 788 15.8
30–39 380 15.9 497 19.2 877 17.6
40–49 402 16.8 452 17.4 854 17.1
50–59 451 18.9 386 14.9 837 16.8
60–69 440 18.4 334 12.9 774 15.5
70–79 268 11.2 208 8 476 9.5
80– 87 3.6 119 4.6 206 4.1
Country of birth
Sweden 2108 88.2 2081 80.2 4189 84
Other 283 11.8 513 19.8 796 16
Citizenship
Swedish 2293 95.9 2374 91.5 4667 93.6
Other 98 4.1 220 8.5 318 6.4
Marital status
Married 1177 49.2 1037 40 2214 44.4
Unmarried 796 33.3 1111 42.8 1907 38.3
Lives at separate place 306 12.8 312 12 618 12.4
Other 112 4.7 134 5.2 246 4.9
Income
None (0) 94 3.9 213 8.2 307 6.2
1–84,999 294 12.3 402 15.5 696 14
85,000–159,999 402 16.8 486 18.7 888 17.8
160,000–234,999 619 25.9 647 24.9 1266 25.4
235,000–309,999 510 21.3 428 16.5 938 18.8
310,000– 472 19.7 418 16.1 890 17.9
All 2391 100 2594 100 4985 100
Municipality
1. large cities 358 15 527 20.3 885 17.8
2. Suburban 365 15.3 421 16.2 786 15.8
3. Major cities 683 28.6 689 26.6 1372 27.5
4. Commuter municipalities 167 7 139 5.4 306 6.1
5. Rural municipalities 82 3.4 77 3 159 3.2
6. Productive municipalities 177 7.4 161 6.2 338 6.8
7. Municipalities (other) ⬎ 25,000 inhab. 323 13.5 328 12.6 651 13.1
8. Municipalities 12,500–25,000 inhab. 172 7.2 167 6.4 339 6.8
9. Municipalities ⬍ 12,500 inhabitants 64 2.7 85 3.3 149 3
All 2391 100 2594 100 4985 100

Results answer in a more positive way was the case in Factor 1.


Trends favors the mentally ill In other words, a preparedness to interact with a person
From Tables 3–6 it can be seen whether the Swedish peo- with a mental illness seemed to be at hand.
ple agreed, were neutral or disagreed to the statements in Regarding factor 2: Fearful and Avoidant, it was
each of the four factors derived from the principal com- shown by means of χ2 that there was a difference
ponent analysis. By means of χ2, it was indicated that (P ⬍ 0.01) between the response groups towards the neg-
there was a difference between the three response groups ative claims incorporated in this factor. Based on the
“totally disagree”, “neutral” and “totally agree” towards means of responses: “totally agree” (mean ⫽ 201), “neu-
the statement included in factor 1: Intention to Interact tral” (mean ⫽ 492) and “totally disagree” (mean ⫽ 1607),
and based on the mean numbers of n in each response a trend seemed likely to be at hand. By means of
group (means ⫽ 346, 673 and 1294) it seemed plausible to Jonkheere’s trend test, a significant (P ⬍ 0.01) tendency
assume that there was a response trend. By means of in the answers was found. Most people (mean ⫽ 1607)
Jonkheere’s trend test, a significant (P ⬍ 0.01) tendency to did not agree with a statement like “It is best to avoid

88 NORD J PSYCHIATRY·VOL 66 NO 2·2012


Table 2. Principal component analysis for 29 items with Cronbach’s alpha for each factors.
Cronbach’s α F1 II F2 FA F3 OP F4 ID

Factor 1 (F1): Intention to interact (II)


22.* I can consider working together with someone who has a mental illness. 0.907 546 272 186 349
25. I would invite someone to my home even if I know they had a mental illness. 630 295 258 297

NORD J PSYCHIATRY·VOL 66 NO 2·2012


28. I can consider being friends with someone who had been a patient in the psychiatric care. 666 245 111 336
30. Most persons who were once patients in a mental hospital can be trusted as babysitters 441 156 440 103
35. The mentally ill should not be treated as outcasts of society 482 211 019 435
37. If someone had been a patient in the psychiatric care became one of my neighbors, I would welcome 736 229 286 140
them into my home sometimes.
40. I would speak in a natural manner with neighbors who have had a mental illness. 671 271 154 220
43. If someone who had a mental illness in the past became my neighbor, I would visit him/her. 724 245 270 073
Factor 2 (F2): Fear and Avoidance (FA) 0.901
19. I am against that someone with mental illness lives in my neighborhood. 112 632 356 192
21. It is frightening to think of people with mental problems living in residential neighborhoods 192 625 420 184
23. I would not want to live next door to someone who has been mentally ill 233 615 274 124
27. It is best to avoid anyone who has mental problems 257 676 161 192
29. The best way to handle the mentally ill is to keep them behind locked doors 168 612 142 432
31. I would avoid talking with neighbors who have had a mental illness in the past. 427 603 044 145
34. I would be worried if I visited someone with a mental illness. 380 531 262 009
36. The mentally ill should be isolated from the rest of the community 228 612 111 351
Factor 3 (F3): Open-minded and Pro-integration (OP) 0.909
32. Residents should accept the location of mental health facilities in their neighborhood to serve the needs 350 135 606 405
of the local community
33. The mentally ill are far less of a danger than most people suppose 465 192 472 346
38. Locating mental health services in residential neighborhoods does not endanger local residents 370 178 640 348
39. Mental illness is an illness like any other 385 063 481 230
41. Mental health facilities should be kept out of residential neighborhoods 088 355 745 178
42. Local residents have good reason to resist the location of mental health services in their neighborhood 112 410 720 146
44. Less emphasis should be placed on protecting the public from the mentally ill 349 119 422 270
45. Having mental patients living within residential neighborhoods might be good therapy but the risks to 149 484 627 100
residents is too great
Factor 4 (F4): Community Mental Health Ideology (ID) 0.874
17. The best therapy for many mental patients is to be part of a normal community 143 227 160 732
18. As far as possible, mental health services should be provided through community based facilities 159 234 223 722
20. We need to adopt a far more tolerant attitude toward the mentally ill in our society 254 142 257 637
24. Residents have nothing to fear from people coming into their neighborhood to obtain mental health services 345 136 390 494
26. No one has the right to exclude the mentally ill from their neighborhood 422 213 292 518

*The numbering of the items refers to their placement in the New CAMI-S questionnaire.

89
SWEDISH ATTITUDES TOWARDS PERSONS WITH MENTAL ILLNESS
90
Table 3. Factor 1: Intention to interact comprising eight statements respondents had to agree to.
Totally
χ2 and Jonkheere’s
disagree Neutral Totally agree
T. HÖGBERG ET AL.

Factor 1: Intention to interact trend S1


Item Mean s % n % n % n χ2; S

22. I can consider working together with someone who has a mental illness. 4.20 1.63 18.0 410 31.3 713 50.7 1156 χ2(14) ⫽ 2808.8 ⬎ 36.12,
25. I would invite someone to my home even if I know they had a mental illness. 4.37 1.56 14.6 333 30.8 896 54.6 1250 P ⬍ 0.01; S ⫽ 160 ⬎ 90,
28. I can consider being friends with someone who had been a patient in the psychiatric care. 4.96 1.42 7.9 181 20.3 467 71.8 1652 P ⬍ 0.01
30. Most persons who were once patients in a mental hospital can be trusted as babysitters 2.99 1.54 41.0 922 39.8 895 19.2 434
35. The mentally ill should not be treated as outcasts of society 5.08 1.40 7.2 166 16.4 378 76.4 1756
37. If someone who had been a patient in the psychiatric care became one of my neighbors, 4.28 1.54 14.6 333 34.8 795 50.6 1156
I would welcome them into my home sometimes.
40. I would speak in a natural manner with neighbors who have had a mental illness. 5.06 1.27 5.5 126 19.2 440 75.3 1727
43. If someone who had a mental illness in the past became my neighbor, I would visit him/her. 4.30 1.50 13.2 300 35.0 798 51.8 1179
Mean 346.4 672.8 1294.4

s, standard deviation.
1Jonckheere trend test. The alternative hypothesis is tested against a null hypothesis of no systematic trend across treatments. The test can be applied when you have data for k independent samples,
when measurement is at least ordinal, and when it is possible to specify a priori the ordering of the groups, the test is onetailed (Field, 2004).

Table 4. Factor 2: Fear and avoidance comprising eight statements respondents had to agree to.
Totally agree Neutral Totally disagree χ2 and
Factor 2: Fear and Avoidance Jonkheere’s trend S1
Item Mean s % n % n % n χ2 ; S

19. I am against that someone with mental illness lives in my neighborhood. 4.58 1.51 12. 7 292 26. 3 606 61.0 1406 χ2(14) ⫽ 565.9 ⬎ 36.12, P ⬍ 0.01;
21. It is frightening to think of people with mental problems living in residential neighborhoods 4.60 1.50 12.0 277 26. 4 609 61.6 1418 S ⫽ 194 ⬎ 90, P ⬍ 0.01
23. I would not want to live next door to someone who has been mentally ill 4.74 1.48 11.1 254 23.6 543 65.3 1502
27. It is best to avoid anyone who has mental problems 4.84 1.44 10.2 234 20.7 475 69.1 1583
29. The best way to handle the mentally ill is to keep them behind locked doors 5.17 1.21 4.5 104 17.9 419 77.6 1783
31. I would avoid talking with neighbors who have had a mental illness in the past. 5.26 1.19 4.9 112 15.2 350 79.9 1838
34. I would be worried if I visited someone with a mental illness. 4.67 1.38 9.8 226 26.2 603 64.0 1471
36. The mentally ill should be isolated from the rest of the community 5.23 1.17 4.8 109 14.4 331 80.8 1851
Mean 201 492 1606.5

s, standard deviation
1Jonckheere trend test. The alternative hypothesis is tested against a null hypothesis of no systematic trend across treatments. The test can be applied when you have data for k independent samples,

when measurement is at least ordinal, and when it is possible to specify a priori the ordering of the groups, the test is onetailed (Field, 2004).

NORD J PSYCHIATRY·VOL 66 NO 2·2012


SWEDISH ATTITUDES TOWARDS PERSONS WITH MENTAL ILLNESS

anyone who has mental problems”, that is to say, most

trend test. The alternative hypothesis is tested against a null hypothesis of no systematic trend across treatments. The test can be applied when you have data for k independent samples,
Kruskal–Wallis H
and Jonkheere’s S

S ⫽ 42 ⬍ 64, n.s.
H ⫽ 1.12, n.s.;
people did not seem to be fearful and avoidant towards

H; S1
mentally ill individuals.
With reference to factor 3: Open-minded and Pro-
Integration, the respondents were not consistent in their
response to the claims because a difference between the
groups “totally disagree”, “neutral” and “totally agree”
was not found (n.s.) and no significant trend was revealed

812.3
1048

1025
981

1203
527
404

918

392
n
Totally agree
(n.s.) either. Consequently, no pattern was detected when
analyzing responses to this factor.
When analyzing factor 4: Community Mental Health

45.8

49.4
42.9

52.4
23.1
17.8

40.7

17.8
Ideology by means of χ2, a difference between groups

%
“totally disagree”, “neutral” and “totally agree” towards
the statement included in the factor was found (P ⬍ 0.01).
Furthermore, based on the inspection of the means for

808.1
866

854
846

581
754
736

911

917
n
each group (means ⫽ 253, 733 and 1269), a significant

when measurement is at least ordinal, and when it is possible to specify a priori the ordering of the groups, the test is onetailed (Field, 2004).
Neutral
(P ⬍ 0.01) trend was calculated. The trend was positive
(i.e. in agreement) towards statements like: “No one has

37.9

37.4
37.0

25.0
33.0
32.4

40.4

32.4
the right to exclude the mentally ill from their neighbor-

%
hood” or “We need to adopt a far more tolerant attitude
toward the mentally ill in our society”. Accordingly, the
present results based on responses to three out of four

661.8
1001
1134

1077
Totally disagree
Table 5. Factor 3: Open-minded and Pro-integration combining eight statements respondents had to agree to.

373

302
461

519

427
n
factors yielded that the respondents displayed trends
towards positive attitudes towards mental illness and
mentally ill persons.
16.3

13.2
20.1

22.6
43.9
49.8

18.9

49.8
%

Impact of experience of mental illness


on attitudes
1.55

1.47
1.53

1.77
1.65
1.57

1.53

1.51
s

To find out if the experience of mental illness was sig-


nificantly related to the factors: 1) Intention to Interact,
2) Fearful and Avoidant, 3) Open-minded and Pro-
Mean

4.09

4.23
3.94

4.19
3.95
4.20

3.91

4.13
Integration, as well as to 4) Community Mental Health
Ideology, a multivariate analysis of variance (MANOVA)
was computed with the factors as dependent variables by
45. Having mental patients living within residential neighborhoods might be
41. Mental health facilities should be kept out of residential neighborhoods
42. Local residents have good reason to resist the location of mental health
32. Residents should accept the location of mental health facilities in their

the source: six levels of experience of mental illness. The


38. Locating mental health services in residential neighborhoods does not

four factors by source yielded significant results


33. The mentally ill are far less of a danger than most people suppose

44. Less emphasis should be placed on protecting the public from the

(P ⬍ 0.001) and the partial eta-squared (η2) ranged from


0.22 to 0.45. No significant difference was found between
neighborhood to serve the needs of the local community

factors 1 and 2 (Tables 7 and 8).


good therapy but the risks to residents are too great

Logistic regression results on factors


Table 9 provides an overview on effects of demographics
39. Mental illness is an illness like any other
Factor 3: Open-minded and Pro-integration

on the principal component analysis as a whole. It can be


seen that female gender impacts the factors: Fearful and
Avoidant and Open-minded and Pro-Integration. Belong-
services in their neighborhood

ing to the age group 18–30 years affects Intention to


Interact and Open-minded and Pro-Integration, and age
endanger local residents

group 31–50 years affects all factors. Living in a flat as


s, standard deviation.

well as being born in Scandinavia influence Intention to


Interact with a person with mental illness. Only nine
mentally ill.

years of compulsory school affect three of the four fac-


1Jonckheere

tors: Intention to Interact, Fearful and Avoidant and


Open-minded and Pro-Integration, and finally to be
Mean
Items

employed has an effect on all four factors.

NORD J PSYCHIATRY·VOL 66 NO 2·2012 91


T. HÖGBERG ET AL.

trend test. The alternative hypothesis is tested against a null hypothesis of no systematic trend across treatments. The test can be applied when you have data for k independent samples,
χ2 and Jonkheere’s trend S1 Aspects contributing to the willingness to live in the

χ2(8) ⫽ 268.3 ⬎ 26.12,


P ⬍ 0.01; S ⫽ 75 ⬎ 45,
neighborhood of persons with mental illness were sub-
jected to a deeper analysis by means of direct logistic

P ⬍ 0.01
regression (Tables 10 and 11). The analysis was per-
χ2; S

formed to assess the impact of a number of demographic


factors on the likelihood that the respondents would or
would not like to have a person with mental illness in
their neighborhood (item 19). The binary dependent vari-
1269.4 able was coded as positive [coded on response alterna-
1097
1137

1367
1173

1573

tives 1–3] or negative [coded on response alternatives


Totally agree

4–6]) towards persons with mental illness in their neigh-


borhood. The full model containing all predictors was sta-
tistically significant, χ2 (df ⫽ 17, n ⫽ 2066: positive ⫽ 449:
47.8
49.3

55.1
51.6

64.2
%

negative ⫽ 1617) ⫽ 85.61, P ⬍ 0.0001, indicating that the


model was able to distinguish between respondents who
732.6
980
496

786
799

602

were or were not positive towards mentally ill individuals


when measurement is at least ordinal, and when it is possible to specify a priori the ordering of the groups, the test is onetailed (Field, 2004).
n
Neutral

in their neighborhood. Concordant association of


predicted probabilities and observed responses was 63.4%.
42.7
38.5

34.2
35.1

26.2
%

Based on odds ratio or effect sizes, women were twice


more likely compared with men to react positively
towards having a person with mental illness in their
Totally disagree

253.4
219
280

246
302

220
n

neighborhood. Besides female gender, factors linked to a


Table 6. Factor 4: Community Mental Health Ideology including five statements respondents had to agree to.

favorable attitude towards persons with mental illness in


the neighborhood are shown in Table 11.
9.5
12.2

10.7
13.3

9.6
%

Discussion
1.38
1.43

1.47
1.48

1.45
s

The aim of the present study was to survey Swedish atti-


tudes towards persons with mental illness related to fac-
Mean

4.32
4.30

4.46
4.28

4.69

tors impacting these attitudes by means of New CAMI-S


comprising a behavioral–intention element reflecting the
intent to interact with persons with mental illness.
20. We need to adopt a far more tolerant attitude toward the mentally ill in our society
18. As far as possible, mental health services should be provided through community

This was done by clustering the respondents’ recurrent


17. The best therapy for many mental patients is to be part of a normal community

24. Residents have nothing to fear from people coming into their neighborhood to

themes about mental illness by means of principal com-


26. No one has the right to exclude the mentally ill from their neighborhood

ponent analysis yielding four factors called: Intention to


Interact, Open-minded and Pro-Integration, Fearful and
Avoidant, Community Mental Health Ideology. These fac-
tors were subjected to trend analyses to reveal in which
direction, in agreement or disagreement, the respondents’
answers to the statements in each factor, went. The trend
analysis in Factor 1 revealed a preparedness to interact
with a mentally ill person. Most respondents were in
agreement with statements like: “I can consider working
Factor 4: Community Mental Health Ideology

together with someone who has a mental illness” or “I


would speak in a natural manner with neighbors who
have had a mental illness.” Most people did disagree
obtain mental health services

with statements like “I would not want to live next door


to someone who has been mentally ill or otherwise”;
“The best way to handle the mentally ill is to keep them
s, standard deviation.

behind locked doors”. To be precise, most people did not


based facilities

seem to be fearful and avoidant towards persons with


mental illness.
1Jonckheere

The present results are based on responses to three


Means

out of four factors yielding trends in agreement with


Item

positive attitudes towards mental illness and mentally ill

92 NORD J PSYCHIATRY·VOL 66 NO 2·2012


SWEDISH ATTITUDES TOWARDS PERSONS WITH MENTAL ILLNESS

Table 7. Impact of experience of mental illness on factors: Intention to interact, Fear and Avoidance, Open-minded and Pro-integration
and Community Mental Health Ideology.
Four factors retrieved from principal component analysis Number of experiences
as dependent variables in MANOVA by experiences of of mental illness t-test (two-tailed)
mental illness n ⫽ 2024 (five conditions) Mean s between factor means

Factor 1 (F1): Intention to interact by number of 132 0.00 40.55 9.36 F1 ⫽ F2,
conditions: Respondent’s own, other’s or professionally 1604 1.00 38.38 9.95 F1 ⬎ F3 ⫽ t(10) ⫽ 2.954,
experienced mental illness 205 2.00 43.04 7.96 P ⫽ 0.014,
62 3.00 46.18 7.31 F1 ⬎ F4 ⫽ t(10) ⫽ 9.929,
18 4.00 47.00 4.50 P ⫽ 0.0001
3 5.00 51.00 2.65
Factor 2 (F2): Fear and Avoidance by number of 132 0.00 40.07 7.72 F2 ⬎ F3 ⫽ t(10) ⫽ 2.577,
conditions: Respondent’s own, other’s or professionally 1604 1.00 38.42 7.98 P ⫽ 0.0276;
experienced mental illness 205 2.00 41.35 6.83 F2 ⬎ F4 ⫽ t(10) ⫽ 13.717,
62 3.00 43.39 6.30 P ⫽ 0.0001
18 4.00 43.89 4.70
3 5.00 45.00 3.00
Factor 3 (F3): Open-minded and Pro-integration by 132 0.00 33.82 8.84 F3 ⬎ F4 ⫽ t(10) ⫽ 6.029,
number of conditions: Respondent’s own, other’s or 1604 1.00 31.90 8.89 P ⫽ 0.0001
professionally experienced mental illness 205 2.00 35.46 8.88
62 3.00 36.84 8.82
18 4.00 35.33 7.75
3 5.00 45.33 2.52
Factor 4 (F4): Community Mental Health Ideology by 132 0.00 22.48 5.64
number of conditions: 1604 1.00 21.59 5.58
Respondent’s own, other’s or professionally experienced 205 2.00 23.27 4.87
mental illness 62 3.00 25.03 4.30
18 4.00 23.78 3.39
3 5.00 27.33 2.31

s, standard deviation.

persons. Even in the middle of the 1960s, Phillips (28) mentally ill person living in his neighborhood while
stated that laymen’s increased ability to recognize differ- 12.7% rejected a mentally ill as neighbor. This means
ent types of mental illnesses made it possible to estimate that the NIMBY phenomenon time and again is at hand,
their willingness to have as neighbor a paranoid schizo- implying that people with serious mental illnesses
phrenic (70%), simple schizophrenic or depressed neu- may be dangerous and unpredictable, which a part of
rotic (96.7%) and phobic–compulsive or normal (100%). the Swedish population may perceive as a fact (1, 2, 3,
Then with added knowledge that these persons had been 29, 30).
in mental hospital, the willingness to have as a neighbor It was presently observed that previous experience of
a paranoid schizophrenic (43.3%), simple schizophrenic mental illness significantly affected all four factors, of
(78.3%), a depressed neurotic (83.3%) and a phobic- which three comprised trends of positive attitudes
compulsive (93.3%) or a normal (96.7%) had changed towards mental disorder, in agreement with findings stat-
for the worse. Presently, most of the respondents (71.8%) ing that different kinds of personal experience correlate
agreed with the statement “I can consider being friends with positive attitudes towards persons with mental ill-
with someone who had been a patient in the psychiatric ness (1, 20). On the other hand, the experience of men-
care”. Currently, the concept of being mentally ill com- tal disorders may be intensely negative, and therefore
prised all kinds of disorders but 61% did not refuse a leads more often than not to a wish to keep a safe

Table 8. Multivariate analysis of variance: Tests of between-subjects effects with Factors 1–4 as dependent variables by experience of
mental illness.
Dependent Type III sum
Source variable of squares df Mean square F Sig. η2

Experience of mental illness Factor 1 8836.73 5 1767.35 19.089 0.000 0.045


Factor 2 3538.92 5 707.78 11.688 0.000 0.028
Factor 3 4385.40 5 877.08 11.144 0.000 0.027
Factor 4 1344.18 5 268.83 9.007 0.000 0.022

NORD J PSYCHIATRY·VOL 66 NO 2·2012 93


T. HÖGBERG ET AL.

Table 9. Overall results of logistic regression: Impact of demographics on principal components.


Age group Age group Birth country: Education: 9-year
Factor/item Gender female 18–30 31–50 Accommodation: flat Scandinavia compulsory school Employment

F1. Intention to Interact


22. *** *** ***

25. *** *** ***

28. *** *** ***

30. *** ***

35.
37. *** *** *** ***

40. *** ***

43. *** *** ***

F2. Fear and Avoidance


19. *** *** ***

21. *** *** ***

23. ***

27. ***

29.
31. ***

34. ***

36. ***

F3. Open-minded and


Pro-Integration
32. *** *** ***

33. *** ***

38. *** ***

39. *** *** ***

41. *** ***

42. ***

44. ***

45. ***

F4. Community Mental


Health Ideology
17. *** ***

18.
20. *** ***

24. ***

26. ***

***P ⬍ 0.0001.

distance from persons with mental illness (7, 18). The high education correlates with a positive attitude towards
stigmatized socially rejected person with mental illness persons with a serious mental illness. It is known that
internalizes then his/hers stigma leading to an inferior negative attitudes towards persons with mental illness
self-image (17). However it is also known that a rela- link to lower standard of living (19, 23). This was not
tive’s experience of mental illness may diffuse his/her confirmed in the present study, where 28.6% of the
mental health and may constitute such a heavy burden
that some relatives believed that a relative with a mental
illness would be better off dead, and/or wished that the Table 10. Logistic regression: Type 3 analysis of effects of
relative with a mental illness and the relative had never demographics on “not in my neighborhood” (item 19).
met, and that the relative with a mental illness had never
Effect df Wald χ2 Pr ⬎ χ2
been born (31).
Demographic factors impacting a person’s willingness Gender 1 41.70 ⬍0.0001
or reluctance to live in the same neighborhood as a per- Age group 3 9.90 0.019
Marital status 1 1.60 0.206
son with a mental illness was analyzed. Currently, per-
Children (n) 1 0.03 0.868
sons of the female gender, aged 31–50 years, born in Children (age group) 1 1.18 0.278
Scandinavia or outside Europe, educated 9 years of Country of birth 2 11.50 0.003
compulsory schooling and living in a flat were found to Education 3 6.61 0.086
be more sympathetic towards persons with a mental ill- Housing condition 2 0.17 0.921
Inhabitants (n) 3 2.90 0.41
ness as opposed to previous research (1, 7), claiming that

94 NORD J PSYCHIATRY·VOL 66 NO 2·2012


SWEDISH ATTITUDES TOWARDS PERSONS WITH MENTAL ILLNESS

Table 11. Analysis of maximum likelihood and odds ratio estimates for demographic data predicting a “not in my neighborhood attitude”
(item 19).
Maximum likelihood estimates Odds ratio estimates

Wald 95% CI
Parameter df Estimate s Wald χ2 Pr ⬎ χ2 Parameter Point estimates limits

Gender: woman 1 0.36 0.19 41.70 ⬍0.0001 vs. man 2.04 1.66 2.54
Age group
18–30 1 0.11 0.06 0.56 0.453 vs. age 66– 0.89 0.55 1.43
31–50 1 ⫺ 0.034 0.14 9.48 0.002 vs. age 66– 0.57 0.38 0.86
51–65 1 0.00 0.10 0.00 0.994 vs. age 66– 0.78 0.56 0.38
Marital status (single) 1 0.09 0.07 1.60 0.206 vs. cohabiting 1.19 0.91 1.55
Children (n) 1 0.03 0.18 0.03 0.868 1.03 0.73 1.46
Children (age group) 1 ⫺ 0.18 0.17 1.18 0.278 0.83 0.60 1.16
Country of birth other 1 ⫺ 0.46 0.20 5.49 0.019 vs. rest of Europe 0.60 0.30 1.31
than Europe
Scandinavia 1 0.41 0.13 10.38 0.001 vs. rest of Europe 1.43 0.89 2.31
Education other 1 ⫺ 0.09 0.12 0.56 0.453 vs. college/ 0.75 0.52 1.07
university
9 years compulsory 1 ⫺ 0.21 0.11 3.81 0.051 vs. college/ 0.66 0.47 0.93
school university
Upper secondary 1 0.10 0.09 1.19 0.275 vs. college/ 0.90 0.68 1.19
school university
Housing other 1 0.07 0.26 0.08 0773 vs. house/row 1.10 0.51 2.38
house
Flat 1 ⫺ 0.06 0.15 0.15 0.699 vs. house 0.96 0.68 1.19
Inhabitants
50–100,000 1 ⫺ 0.16 0.13 1.56 0.212 vs. rural area 0.80 0.42 1.51
⬍50,000 1 ⫺ 0.04 0.10 0.13 0.717 vs. rural area 0.90 0.50 1.62
⬎100,000 1 0.13 0.13 1.07 0.300 vs. rural area 1.07 0.56 2.01

s, standard deviation.

respondents resided in major cities. The connection implied that “persons with a serious mental illness” are
between families, friends and neighbors was surveyed by nowadays integrated in the community, thus helping the
Hilber (32), who consulted 30,000 people and found that respondent to recognize a “person with a serious men-
on average homeowners interact 30% more than renters tal illness” as a person with long-term mental distur-
with their immediate neighbors in developed neighbor- bance resulting in daily dysfunctions requiring
hoods. The flats are more likely than homes to be rented long-term treatment (13, 35).
and the interaction between neighbors living in flats may
be minimal, and the inhabitants do not necessary know
each other not to mention each other’s mental history. Conclusion
Regarding methods, 2391 (47.9%) agreed to partici- To sum up, the present study aimed at surveying Swedish
pate in the present study after two reminders. The drop-outs attitudes towards persons with mental illness related to
comprised object-loss and partial-loss but the partial factors impacting these attitudes. By New CAMI-S, the
loss was less than 5%, which was approved by the CSA Swedish attitudes towards persons with mental illness
and the response rate was considered satisfactory. were surveyed and trends showed in three out of four
The population consisted of Swedish people aged 18–85 factors derived by principal component analysis that the
years. Altogether 56.6% females and 43.4% males completed Swedes were rather in agreement with living next to a
the questionnaires, the numbers can be compared with person with mental illness. Aspects impacting the
1.03 men and women (aged 15–64 years) and 0.73 men Swedish attitudes towards persons with mental illness and
and women aged ⬎ 65 years (33). The gender balance their willingness to have him/her residing in their neigh-
was considered reasonable in the present study. The con- borhood comprised experience of mental illness, female
cept of “mental illness” included a variety of psychiatric gender, age (31–50 years), born in Scandinavia or outside
disorders such as e.g. depression, anxiety, alcoholism and Europe, only 9 years of compulsory education and accom-
schizophrenia (18, 34). In order to counteract ambiguity modation in a flat. The New CAMI-S came out as a use-
about the concept in question, the cover letter explained ful tool for screening Swedish attitudes towards persons
that the Swedish reform of psychiatric care (1995) with a mental illness.

NORD J PSYCHIATRY·VOL 66 NO 2·2012 95


T. HÖGBERG ET AL.

Acknowledgment This study was partially funded by research grants from 16. Brunt D, Hansson L. Att leva med psykiska funktionshinder: livs-
The Swedish National Board of Health and Welfare. situation och effektiva vård- och stödinsatser. [Living with mental
disabilities: life and effective care and support interventions.] Lund:
Studentlitteratur; 2005.
17. Erdner A. Stories about loneliness in everyday life; experienced by
Declaration of interest: The authors report no conflicts of people with serious mental illness. Doctoral Dissertation, Department
interest. The authors alone are responsible for the content of Clinical Neuroscience, Karolinska Institutet, Stockholm; 2006.
18. Angermeyer MC, Dietrich S. Public beliefs about and attitudes
and writing of the paper. towards people with mental illness: A review of publication studies.
Acta Psychiat Scand 2006;113:163–79.
19. Wolff G, Pathare S, Craig T, Leff J. Community attitudes to mental
Author contributions illness. Br J Psychiatry 1996;168:183–90.
20. Högberg T, Magnusson A, Ewertzon M, Lützén K. Attitudes towards
TH was responsible for the study conception and design, mental illness in Sweden: Adaptation and development of the
performed the data collection and drafted the manuscript. Community Attitudes towards Mental Illness Questionnaire. Int
KL and AM made critical revisions to the paper and J Ment Health Nurs 2008;17:302–10.
21. Kruglanski A, Higgins T. Social psychology: A general reader. New
supervised the study. BE-K carried out the statistical anal- York: Psychology Press; 2003.
ysis and was advisory in the drafting of the manuscript. 22. Rudder-Baker L. Explaining attitudes; A practical approach to the
mind. Cambridge: University Press; 1995.
23. Wolff G, Pathare S, Craig T, Leff J. Community knowledge of
mental Illness and reaction to mentally ill people. Br J Psychiatry
References 1996;168:191–8.
1. Brockington IF, Hall P, Levings J, Murphy C. The community’s 24. Rudder-Baker L. Explaining attitudes; A practical approach to the
tolerance of mentally ill. Br J Psychiatry 1993;162:93–9. mind. Cambridge: Cambridge University Press; 1995.
2. Leff, J, Warner, R. Social inclusion of people with mental 25. Taylor SM, Dear MJ. Scaling community attitudes toward the
illness. Cambridge: Cambridge University Press; 2006. p. 16. mentally ill. Schizophren Bull 1981;7:225–40.
3. Cowan, S. NIMBY syndrome and public consultation policy: The 26. European Commission, Green Paper. Improving the mental health of
implications of a discourse analysis of local responses to the the population, a strategy for the European Union. Brussels: Health
establishment of a community mental health facility. Health Soc and Consumer Protection, General Directorate; 2005.
Care Community 2003;11:379–86. 27. Statens offentliga utredningar; SOU 1998:16. När åsikter blir
4. Diefenbach DL, West MD. Television and attitudes toward mental handling. (Swedish public reports. Ministry of health and social
health issues: Cultivation analysis and the third person effect. affairs. Where opinions become deeds). Stockholm:
J Commun Psychol 2007;35:181–95. Socialdepartementet (in Swedish).
5. Douglas KS, Guy LS, Hart SD. Psychosis as a risk factor for 28. Phillips DL. Public identification and acceptance of the mentally ill.
violence to others: A metaanalysis. Psychol Bull 2009;135: Am J Public Health 1966;56:755–63.
679–706. 29. Oppenheim AN. Questionnaire design, interviewing and attitude
6. Foster K, Usher K, Baker JA, Gadai S, Ali S. Mental health measurement. New York: Continuum International; 1992.
workers’ attitudes toward mental illness in Fiji. Aust J Adv Nurs 30. Beaton DE, Guillemin F. Guidelines for the process of cross-cultural
2008;25:72–9. adaption of self-report measures. Spine 2000;25:3186–91.
7. Leff, J, Warner, R. Social inclusion of people with mental illness. 31. Östman M, Kjellin L. Stigma by association—Psychological factors
Cambridge: Cambridge University Press; 2006. in relatives of persons with mental illness. Br J Psychiatry
8. Warner, R. Community attitudes towards mental disorder. In: 2002;181:494–8.
Thornicroft G, Szmukler G, editors. Textbook of community 32. Hilber C. Homeowners in cities “make best neighbours”.
psychiatry. New York: Oxford University Press; 2001. p. 453–64. 2010. http://www.telegraph.co.uk/news/uknews/7774853/
9. Högberg T, Magnusson A, Lützén K. To be a nurse or a neigh- Homeowners-in-cities-make-best-neighbours.html.
bour? A moral concern for psychiatric nurses living next door to 33. Sveriges_demografi. Sveriges_demografi. http://sv.wikipedia.org/
individuals with a mental illness. Nurs Ethics 2005;12:468–78. wiki/Sveriges_demografi. 7 February 2011.
10. Regeringens skrivelse 2008/09:185. En politik för personer med 34. Link B, Phelan J, Bresnahan M, Stueve A, Pescosolido B. Public
psykisk sjukdom eller psykisk funktionsnedsättning. (Government conceptions of mental illness: Labels causes, dangerousness, and
Communication 2008/09: 185. A policy for people with mental social distance. Am J Public Health 1999;89:1328–33.
illness or mental impairment.) Stockholm: Government Offices 35. Statens offentliga utredningar; SOU 2006: 5. Vad är psykiskt
(in Swedish). funktionshinder? Nationell psykiatrisamordning ger sin definition av
11. Lundberg B, Hansson L, Wentz E, Björkman T. Stigma, begreppet psykiskt funktionshinder. (Swedish public reports.
discrimination, empowerment and social networks: A preliminary Ministry of health and social affairs. The significant meaning of
investigation of their influence on subjective quality of life in a serious mental illness? Definition given by The National Psychiatry
Swedish sample. Int J Soc Psychiatry 2008;54:47–55. Coordination). Stockholm: Socialdepartementet (in Swedish).
12. Lundberg B, Hansson L, Wentz E, Björkman T. Are stigma 36. Field, AP. (2004). Discovering statistics using SPSS (2nd edition).
experiences among persons with mental illness, related to London: Sage.
perceptions of self-esteem, empowerment and sense of coherence?
J Psychiat Ment Health Nurs 2009;16:516–22. Torbjörn Högberg, R.P.N., R.N.T., Ph.D., Karolinska Institutet, Institution
13. Socialstyrelsen följer upp och utvärderar; SoS 1999:1. Välfärd och of Clinical Neuroscience, Stockholm Centre for Psychiatric Research and
valfrihet? Slutrapport från utvärderingen av 1995 års psykiatrireform. Education, Stockholm, Sweden.
(The Swedish National Board of Health and Welfare. Welfare and Annabella Magnusson, R.P.N., M.Sc., Ph.D., Ersta Sköndal University
freedom of choise? The final report of the evaluation of the 1995 College, Department of Health Care Sciences, Stockholm, Sweden.
psychiatric care reform). Stockholm: Socialstyrelsen (in Swedish). Kim Lützén, R.P.N., R.N.T., Ph.D., Professor, Karolinska Institutet,
14. Link BG, Phelan JC. Conceptualizing stigma. Ann Rev Sociol Department of Neurobiology, Care Sciences and Society Division of
2001;27:363–85. Nursing Science, Stockholm, Sweden.
15. Link BG, Yang LH, Phelan JC, Collins PY. Measuring mental illness Béatrice Ewalds-Kvist, L.P., Ph.D., Associate Professor, Stockholm
stigma. Schizophren Bull 2004;30:511–41. University, Department of Psychology, Stockholm, Sweden.

96 NORD J PSYCHIATRY·VOL 66 NO 2·2012


Copyright of Nordic Journal of Psychiatry is the property of Taylor & Francis Ltd and its content may not be
copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written
permission. However, users may print, download, or email articles for individual use.

You might also like