We compared the inclusion-related attitudes of community agency staff in the United States and Israel. Samples of 147 (United States) and 74 (Israel) staff members were assessed with the Community Living Attitudes Scale (CLAS). Results showed that higher educational levels, regardless of nation, were associated with higher Empowerment scores, lower Exclusion scores, lower Sheltering scores, and higher Similarity scores. Empowerment attitudes were strongly related to differences in the agencies in which individuals were employed. Controlling for age, education, and agency, we found that staff in the United States had higher Empowerment scores than did Israeli staff. Israeli staff had higher Sheltering and Similarity scores and marginally higher Exclusion scores than did the United States staff.
The attitudes of staff affect the extent to which the philosophy of inclusion of people with intellectual disabilities is implemented. Attitudes of staff differ from those of the general public in that public attitudes favor empowerment less and exclusion more than do staff personnel (Henry, Keys, Balcazar, & Jopp, 1996). Early in their employment, staff members' attitudes may resemble those of the general public, but in the course of employment are shaped by numerous official and unofficial influences. These influences include training, attitudes of coworkers, overt agency values, and covert organizational norms and values (Katz & Kahn, 1978; Keys, Henry, & Schaumann, 1995; Schein, 1990) as well as their experiences with people who have intellectual disabilities. Staff members' attitudes affect their work motivation and, thus, may affect the quality of service that people with intellectual disabilities receive (Balcazar, MacKay-Murphy, Keys, Henry, & Bryant, 1998). Volunteers and staff in services for people with intellectual disabilities tend to have value systems that are distinct from those of the general population (Keys et al., 1995; Williams, 1986). Their values and attitudes motivate them to stay in the field despite low pay and benefits (Askvig & Vassiliou, 1991). Positive staff attitudes may also be associated with work of higher quality (Henry, Keys, & Factor, 2000).
Social psychological theory and research into the structure and cognitive components of attitudes suggests that attitudes toward people with intellectual disabilities, like other complex attitudes, may be multidimensional (e.g., Festinger, 1957; Heider, 1958; Osgood & Tannenbaum, 1955). Further, societal ambivalence toward persons with intellectual disabilities (Heal, 1988) may result from people simultaneously holding conflicting attitudes that are on two or more separate attitude dimensions. Accordingly, measures of attitudes toward people with mental illness—Community Attitudes Toward Mental Illness (Taylor & Dear, 1981) and intellectual disabilities—Mental Retardation Attitude Inventory (Antonak & Harth, 1994) have repeatedly been found to have multidimensional structures.
Attempting to measure inclusion-related attitudes in a manner consistent with current philosophy and grounded in issues important to persons with intellectual disabilities, Henry, Keys, Jopp, and Balcazar (1996) created the Community Living Attitudes Scale–CLAS. They found four factors in an analysis of items related to contemporary attitudes toward the inclusion of people with intellectual disabilities in society (Henry, Keys, Jopp, & Balcazar, 1996; Henry, Keys, Balcazar, & Jopp, 1996b; Henry, Duvdevany, & Keys, 2000; Horner-Johnson et al., 2002). These factors suggest four separate but related dimensions of inclusion attitudes, namely (a) Empowerment, the degree to which people with intellectual disabilities are permitted to make their own life decisions; (b) Exclusion, the degree to which respondents desire to segregate people with intellectual disabilities from community life; (c) Sheltering, the degree to which the respondent believes people with intellectual disabilities require protection from harm; and (d) Similarity, the degree to which the respondent believes that people with intellectual disabilities share a common humanity with others. This four-factor solution has been found to fit the data from community agency staff as well as students and community members in the United States (Henry, Keys, Balcazar, & Jopp, 1996; Henry, Keys, & Jopp, 1998; Henry, Keys, Jopp, & Balcazar, 1996). Cross-cultural validity is suggested by studies of Japanese (Horner-Johnson et al., 2002), and Israeli (Schwartz & Armony-Sivan, 2002) students.
Cross-cultural study of attitudes toward persons with intellectual disabilities can inform us as to the universality or particularity of attitude dimensions found in the United States studies. It can also help us understand the degree to which workplaces and job roles are associated with commonalities and differences in attitudes by comparing attitudes of persons in similar roles across cultures.
Historical and cultural factors are likely to contribute to differences in inclusion attitudes between Israel and the United States. Both nations have intellectual disabilities service delivery systems that include institutions and community arrangements. The two nations have very different histories of intellectual disabilities service delivery, however. In the United States, community agencies evolved after more than a century of institutionalization, exclusion, and prejudice against people with intellectual disabilities (Scheerenberger, 1983, p. 138; Wolfensberger, 1976). Although inclusion philosophy has made deeper inroads in the United States than in Israel, this philosophy affects public attitudes and policies in both nations. Because readers of this journal are likely to be more aware of historical and contemporary factors influencing inclusion attitudes in the United States, our primary but not exclusive focus here is on developments in Israel.
In Israel, the deinstitutionalization movement has played a less significant role in government policy than it has in the United States In the first 2 decades of the modern state of Israel, only two types of out-of-home placement were available: foster care and institutions (Hovav & Ramot, 1998). The first community house offering out-of-home placement was founded in 1969, and, since then, many new community-living arrangements have been opened. More than 90% of the present community agencies were opened during the 1980s (Rimmerman & Schwartz, 1998). Different agencies developed different models for community living. For example, agencies vary in the size of living arrangements, the degree to which staff make decisions for consumers, and the degree to which consumers engage in community employment. Individualized case management, accessing an array of services tailored to individual needs, is not frequently employed in Israel (Schwartz, 1993).
Prior to 1993, family choice in services was limited. In that year, the Israeli Ministry of Labour and Social Affairs changed policies related to referral to residential homes. This policy allowed families greater latitude in choosing the best out-of-home placement for their child. This policy of increased choice has improved the quality of community services and encouraged parents to begin to opt more frequently for long-term treatment in the community. Since 1993, the initiative for organization of community housing has been assumed mainly by local private and nonprofit parent organizations. These organizations provide about 75% of the community-living arrangements, whereas the public sector operates only about 25%. The trend today is to develop small programs with maximum abilities for inclusion and integration in the community, enabling greater independence and quality of life (Schwartz, 1993).
Traditionally, the Arab society in Israel has met individual needs through a large network of primarily familial ties (Haj-Yahia, 1995). Until recent years, out-of-home placement was unacceptable among most Israeli Arabs. However, young Arabs are gradually adopting new, more individualistic norms and values as the group and the extended family recede in importance as providers of services. Traditional arrangements are being replaced by government institutions as part of a transition from an informal support system to more formal services (Schwartz, Duvdevany, & Azaiza, 2002). In recent years, new community-living arrangements have been opened for Arab citizens with intellectual disabilities.
The differences between Israel and the United States are reflected in the funding of intellectual disabilities services. In Israel today, a high percentage (approximately 80%) of the intellectual disabilities services budget is allocated to institutions (Shnit, 1997), whereas in the United States these proportions are nearly reversed, with 25% spent on institutions and 75% on community arrangements in 2000 (Braddock, Hemp, Rizzolo, Parish, & Pomeranz, 2002).
There has been stronger emphasis on an inclusion philosophy of community living for persons with intellectual disabilities over a longer period of time in the United States than in Israel. Training and organizational culture over time should have resulted in greater diffusion of community-living philosophy among agency staff in the United States as compared to Israel. However, the emergence of community-living models and inclusion philosophy has been more rapid in Israel as compared to the United States These historical differences give reason to expect that individual staff members' attitudes may differ between the two nations.
In the present study we compare the inclusion-related attitudes of Israeli and United States staff who serve and support persons with intellectual disabilities. Using samples collected in the United States and Israel, we explored mean similarities and differences between the two nations. Based on the historical factors noted above, we expected that (a) United States staff members would show attitudes more favorable towards empowerment and less favorable toward exclusion, and (b) Israeli staff would display attitudes more favorable toward sheltering and similarity. In order to estimate attitudinal differences accurately, we controlled for other individual differences believed to be associated with attitudes. Because different agencies may have different degrees of emphasis on inclusion philosophy, we also used methods that could estimate the variability in attitudes due to agency differences. To our knowledge, this study is the first in which staff attitudes toward community living for people with intellectual disabilities in Israel and the United States has been compared.
The agencies represented in the United States sample were recruited to participate in a training needs study of community agencies in Illinois (Balcazar et al., 1998). All of the community agencies providing residential and other services to persons with intellectual disabilities in Illinois were solicited for participation. Those who responded represented 86% of the community agencies in Illinois (Balcazar et al., 1998). This state spends a smaller percentage on community services and supports for persons with intellectual disabilities than does the nation as a whole (Braddock et al., 2002). However, the agencies in the sample represented rural, suburban, and urban areas throughout the state, and the state's population represents the United States as a whole with respect to political affiliation, race/ethnicity, age distribution, and urbanization (U.S. Census Bureau, 2000).
The sample for the current study is a subset of that used in the Balcazar et al. (1998) study. In order to select a sample of United States agencies that was comparable to the sample of Israeli agencies, we selected individuals who reported serving persons with intellectual disabilities only and did not serve persons with mental illness or dual diagnosis. In addition, there was a clerical error in printing the CLAS scale in the original Balcazar et al. (1998) study that resulted in 45 cases having to be excluded because it could not be determined that the clerical error had not affected the data. Comparing the excluded cases with the included cases, we found no differences in age, educational level, race/ethnicity, or length of time they had worked in the field. These exclusions resulted in a United States sample of 147 cases from 46 agencies who had no missing data on the CLAS subscales, job type, age, and education. Between 1 and 7 staff members responded from each agency.
In Israel, the study was conducted in four different nonprofit agencies serving all areas of Israel: Jerusalem and the south, Tel-Aviv and the Mediterranean coast (Herzilia), and Haifa and the north. These agencies comprised four fifths of the agencies providing community services in Israel. Two of the agencies were divisions of a parents' organization and two were private, nonprofit organizations. Each agency operated several community facilities for persons with intellectual disabilities. All of the agencies offered residential living arrangements, support and training in skills necessary for living in the community, employment assistance, and other services. The agencies offered services to all Israeli citizens, Jewish, and nonJewish alike, but more than 90% of the workers and residents in the Israeli sample were Jewish.
Table 1 describes the individuals in the United States and Israeli samples and provides statistical comparisons of the samples' compositions on several variables. The United States sample was comprised of 147 staff working in 46 community residential agencies serving people with intellectual disabilities. Three quarters of this sample was female. The number of staff members selected for the study per agency who returned completed surveys varied between 1 and 7. The Israeli sample consisted of 74 staff members working in four agencies for people with intellectual disabilities. The number of staff members per agency who returned complete surveys varied between 4 and 24. Two thirds of the sample was female.
The samples did not differ significantly in gender composition or in their respective proportions of community support staff. The Israeli sample had a higher proportion of staff 55 years of age and over and a lower proportion of 21- to 34-year-old staff members, χ2(2, N = 2210) = 15.71, p < .001. The Israeli sample had a higher proportion of high school graduates than did the United States sample. More Israeli than United States staff also had postsecondary education.
Participants in both samples were assessed using the CLAS. As noted earlier, this 40-item measure has four subscales assessing facets of inclusion attitudes: Empowerment, Sheltering, Exclusion, and Similarity. The CLAS subscales have shown internal consistency reliabilities ranging from .75 to .86 in United States samples (Henry, Keys, Balcazar, & Jopp, 1996; Henry, Keys, Jopp, & Balcazar, 1996) and from .70 to .85 with an Israeli sample (Schwartz & Armony-Sivan, 2002). Validity studies show that they are relatively free from social desirability bias (Henry, Keys, Jopp, & Balcazar, 1996) and that they differentiated those who had training in inclusion philosophy from those who had an equal amount of training in another topic area (Henry, Keys, Balcazar, & Jopp, 1996). For both United States and Israeli samples, the CLAS items were answered on a 6-point Likert-type scale, with anchors ranging from strongly disagree (1) to strongly agree (6). The subscale scores were obtained by averaging the nonmissing items for cases in which at least 90% of the items were completed, a procedure consistent with the scale developers' instructions.
Level of education, age, and type of position held by each respondent were assessed by means of a demographic questionnaire. In order to compare the two samples on age and education, and use these variables in the analysis, we collapsed the United States data into the categories used by the Israeli investigators. For age, three age groups were used: 21 to 34, 35 to 54, and 55 to 64 years. For education, there were four groups: elementary school, partial high school, high school graduate, and academic (college/professional). The three job types found in the United States sample (community support, supervisor, manager) were collapsed into two categories, also to match those of the Israeli investigators: (a) community support staff, who provided direct service and support to persons with intellectual disabilities and neither supervised other workers nor were considered professionals such as psychologists or social workers and (b) supervisory/managerial/professional staff who held administrative or professional positions in agencies. As can be seen in Table 1, the United States and Israeli staff samples did not differ significantly regarding the proportions of staff in these categories.
Data from the United States sample were collected in the following manner. A package containing eight assessment measures were mailed to each of 152 community agencies serving people with intellectual disabilities, chronic mental illness, and people who were diagnosed with both intellectual disabilities and mental illness. A cover letter instructed agency personnel that 1 manager, at least 1 first-line supervisor, and at least 4 community support staff members were to complete the packets of measures. Packets were to be returned anonymously. Staff completed the CLAS scales, which were worded to reflect their attitudes toward the particular people they served (people with intellectual disabilities, chronic mental illness, or dual diagnosis). In addition to the CLAS scales, the packets also contained other measures, including a questionnaire asking for demographic information and an informed consent form. One hundred thirty agencies returned surveys usable for this study, representing an agency return rate of 85.5%. Staff in 48 of these agencies reported serving people with intellectual disabilities who were not dually diagnosed with mental illness.
Data from the Israeli sample were collected in the following manner. Letters were sent to the five agencies providing community services for persons with intellectual disabilities in Israel. Directors of 4 of these agencies agreed, by letter, to participate in the study, representing an agency participation rate of 80%. After receiving these responses, research staff members were dispatched to the agencies. All staff members at each agency who consented to participate completed measures with the research staff present. Measures were completed anonymously and were collected by the research staff rather than by agency administrators in order to preserve the anonymity of the staff respondents.
Table 2 contains the means, standard deviations (SDs), and correlations among variables used in the analysis. The four attitude subscales were correlated significantly and moderately in the same directions as have been reported in previous studies (Henry, Keys, Balcazar, & Jopp, 1996; Henry, Keys, Jopp, & Balcazar, 1996; Horner-Johnson et al., 2002). Age of the respondent correlated significantly and positively with scores on the Sheltering subscale. Educational level correlated significantly and positively with the Empowerment and Similarity subscale scores.
Because the individuals who completed measures of attitudes were selected by agency executives, it was possible that this selection process would affect the results. In order to compensate for this characteristic of the sample, we explored mean differences in attitudes by constructing hierarchical linear models (Bryk & Raudenbush, 1992) of the four CLAS subscales. Hierarchical linear models (also known as mixed models and random-effects regression models) are used to handle nested data structures by accounting for the clustering of individual observations within higher level units. We constructed hierarchical linear models of Empowerment, Exclusion, Sheltering, and Similarity scores. The predictors were nation (Israel, United States), and job type (community support, supervisory/managerial). Because members of the samples differed significantly on education levels and age, we included these variables as covariates in the analyses. We first fit models that included the interaction between nation and job type and then removed this interaction when it was not significant.
The results of these analyses are reported in Tables 3 and 4. Table 3 reports the parameters and significance tests for the hierarchical linear models and Table 4, the means and SDs of the CLAS subscales by the predictors in the equation. The hierarchical linear modeling allowed us to estimate the extent to which attitude scores were clustered within agencies. Empowerment was the only attitude subscale whose scores were significantly clustered by agency, with an intraclass correlation of .28. The intraclass correlation represents the proportion of variance accounted for by the clustering of individuals within agencies. Thus, 28% of the variance in Empowerment subscale scores was attributable to the clustering of individuals within agencies.
Controlling for individual age, nation, and agency, we found that higher levels of education were significantly associated with differences in attitude scores on all of the CLAS subscales. Higher educational levels were associated with higher scores on the Empowerment and Similarity subscales and lower scores on the Exclusion and Sheltering subscales. The effects of education were substantial as well as being statistically significant. The effect sizes on the Empowerment, Exclusion, and Sheltering subscales all accounted for between 6% and 7% of the variance in the scores. The effect of educational level on the Similarity subscale scores accounted for approximately 4% of the variance. The analyses produced no significant effects of either age or staff role on the CLAS subscale scores.
There also were significant effects for nation on all four attitude subscales. These effects are illustrated in Figure 1. United States staff had significantly higher scores on the Empowerment subscale, and the Israeli staff had significantly higher scores on the Exclusion, Sheltering, and Similarity subscales. The largest of these effects was on the Sheltering subscale, on which differences in nation accounted for 17% of the variance in the scores. Effects on the Empowerment, Exclusion, and Similarity subscales accounted for between 3% and 5% of the variance each.
In this study we compared United States and Israeli staff on attitude dimensions as represented by the CLAS. United States staff members had higher scores on the Empowerment subscale than did Israeli staff. This result was consistent with the importance of Empowerment to inclusion efforts in the United States. For over a decade in the United States and Canada, self-advocates with developmental disabilities have been instrumental in adding dimensions of self-advocacy and empowerment to contemporary philosophies of inclusion (Foster-Fishman & Keys, 1997; Longhurst, 1994; Miller & Keys, 1996). Self-advocates say that they and others with intellectual disabilities desire an identity in society defined by their humanity rather than their disabilities. Their efforts have contributed to several changes. Apartments and competitive and supported employment in communities are replacing institutions and sheltered workshops (Goldberg, McLean, LaVigne, Fratolillo, & Sullivan, 1990). People with intellectual disabilities are even becoming salaried professionals serving others with intellectual disabilities. They are joining the boards of directors of the organizations that provide services and supports (Braddock, 1994), such as the Association for Retarded Citizens of the United States and the American Association on Mental Retardation.
The inclusion movement is beginning to make inroads in Israel also. The first community residence was founded in 1969, and since then many new community homes and services have opened. The initiative for, and organization of, community-living arrangements has been assumed mainly by local private and nonprofit parents' organizations without integrated policies or centralized structures. Driven mainly by these organizations, there is a trend emerging toward developing small programs with maximum abilities for inclusion and community integration (Schwartz, 1993). The emergence of these community-living programs has not yet significantly affected the existing institutions, and the self-advocacy movement in Israel is embryonic. Unlike Canada and the United States, the self-advocacy movement has not yet had a significant impact on policy and public opinion. It is also unlikely that staff training in Israel has reflected inclusion to the same extent that it has in the United States Thus, it is not surprising that attitudes favoring empowerment were less prevalent in Israel than in the United States, and attitudes favoring exclusion were more prevalent in Israel.
The notion that the obtained differences in empowerment attitudes reflect differences in agency philosophy is given greater credence by the finding that Empowerment was the only subscale whose scores were significantly clustered by agency. Whereas variation in the other attitude subscales can be accounted for by individual differences, understanding variation in empowerment attitudes requires knowledge of the agencies at which staff members work.
Israeli staff also had higher scores on the Sheltering subscale than did United States staff. This difference, also, may reflect greater diffusion of inclusion attitudes in the United States There is an inherent tension between the values of self-determination and protection. This tension presents staff members with dilemmas whenever affirming self-determination poses a potential risk to safety for an individual with an intellectual disability. Because of this tension, greater endorsement of empowerment may be related to decreased endorsement of sheltering. In fact, Empowerment and Sheltering subscale scores were negatively correlated. Unlike Empowerment scores, however, Sheltering scores were not clustered by agency, indicating, perhaps, that the tension between these two attitude dimensions exists more at the individual level than at the agency level. That is, perhaps training in inclusion philosophy in agencies has stressed empowerment and self-determination without addressing the implications for the duty of protection.
Israeli staff members in this study were found to have higher mean scores on the Similarity subscale than did United States staff members. One possible explanation for this difference is that community-living philosophy may be grounded in different values in the United States than in Israel. The concept of empowerment, which is so central to inclusion efforts in the United States, is anchored in an individualistic ethos and may be less relevant in fostering inclusion in cultures that stress familialism rather than individualism (Roland, 1988). Perhaps in Israel, the philosophy underlying community inclusion efforts is centered on the idea that persons with and without disabilities share a common humanity. Just as the empowerment of people with intellectual disabilities is grounded in individualism in the United States, fostering community inclusion in Israel may require applying the value of similarity to people with intellectual disabilities. In Israel, arguing for inclusion based on the individual's rights may be less effective than arguing for inclusion of people with intellectual disabilities based on their common humanity. Future researchers who compare inclusion attitudes between the United States and other nations will help identify potentially important attitudinal elements that may foster inclusion in other cultures as well as in United States subcultures.
This investigation is preliminary in nature and limited by the nature and sizes of the data sets. The opportunity to examine inclusion attitudes in both nations was presented by the presence of parallel data sets. However, no systematic effort was made to match demographic categories, which had to be adjusted after both data sets were collected in order to facilitate analyses. Another limitation of this investigation was the size of the samples. Fortunately, the combined sample was large enough to provide reliable estimates using the hierarchical linear model methods because the power for detecting effects in a study such as this rests on the number of agencies (Raudenbush & Liu, 2000). However, the size of the sample was not sufficient to reliably examine the interaction between nation and job type. Attitude variation by job roles within nations may be a fruitful focus for future researchers.
This investigation makes new contributions to existing knowledge and suggests areas for further study. It is, to our knowledge, the first investigation in which United States and Israeli inclusion attitudes regarding persons with intellectual disabilities are compared. It provides additional support for the notion that inclusion attitudes comprise multiple dimensions that may exist in tension with each other.
Future cross-national research should also explore inclusion attitudes among Israeli Arabs and Palestinian Arabs because in Arab society agencies are beginning to replace more traditional familial arrangements for persons with intellectual disabilities. Data on staff inclusion attitudes is currently being collected in these programs. The present study underscores the value of cross-national research in attitudes toward persons with intellectual disabilities. Only about 6% of the world's population lives in the United States, and it cannot be assumed that knowledge gained with United States samples can be generalized to other nations. Comparing findings from other nations to existing findings with United States samples is critical to developing a more complete understanding of how staff and others think about persons with intellectual disabilities.
This article is based, in part, upon presentations given at the International Conference on Psychology, Haifa, Israel, June 2000, and the International Association for the Scientific Study of Intellectual Disabilities in Seattle, WA, in August 2000. The United States study was funded, in part, by a contract from the Illinois Department of Mental Health and Developmental Disabilities to the Institute on Disability and Human Development at the University of Illinois at Chicago. The authors thank Willi Horner-Johnson, Georgina Tegart, and Kiyoshi Yamaki for feedback on earlier versions of this paper. The authors also acknowledge the contribution of Marie McKay, David Njus, Joseph Kaplan, and Christine Smith for assistance with data-collection and entry. Daniel Giffort provided valuable support and assistance on administrative and logistical issues. Finally, the authors acknowledge the support and assistance of the Departments of Disability and Human Development, Psychology, and Psychiatry at the University of Illinois at Chicago, and the Faculty of Social Work at the University of Haifa, Israel. Requests for reprints should be sent to the first author.
Authors: David B. Henry, PhD ( email@example.com), Associate Professor, Institute for Juvenile Research, Department of Psychiatry; Christopher B. Keys, PhD, Professor, Department of Psychology; and Fabricio E. Balcazar, PhD, Associate Professor, Department of Disability and Human Development, University of Illinois at Chicago, 840 S. Wood St., Chicago, IL 60612. Ilana Duvdevany, PhD, Associate Professor, Faculty of Social Work, University of Haifa, Mount Carmel, Haifa 31905, Israel