There is little question that the strategies used to improve supported employment outcomes, namely higher wages and higher levels of integration, have changed since the mid-1980s. Innovations of natural supports and employer leadership have helped increase the capacity of provider agencies and the business community to include people with disabilities in the workforce. This report is the sixth in a series that focuses on features of natural supports and its relationship to outcomes. Our purpose in this paper is to describe an analysis designed to investigate the features of employment, wage, and integration outcomes of jobs acquired by people with disabilities early in the development of supported employment compared to more recent years.
Supported employment was defined for the first time in the Developmental Disabilities Act of 1984 (P.L. 98–527) as:
(i) paid employment for persons with developmental disabilities for whom competitive employment at or above minimum wage is unlikely and who need ongoing support to perform in a work setting, (ii) is conducted in a variety of settings in which persons without disabilities are employed, and (iii) is supported by any activity needed to sustain paid work including supervision, training, and transportation. (p. 2665)
For more than a decade, supported employment has been associated with innovations such as natural supports, business leadership, new methods of job development, modern marketing techniques, transition from school to work, expansion of choice, self-determination, person-centered planning, assistive technology, and coworker training. These areas have surely contributed to the large increases in the number of individuals with significant disabilities involved in integrated employment (Revell, Inge, Mank, & Wehman, 1999; Wehman, 2001). Work crews and enclaves (Mank, Rhodes, & Bellamy, 1986) were a part of the initial alternatives to activity centers and segregated workshops for people with severe disabilities as supported employment emerged nationwide. Although these approaches were improvements, it soon became clear that wages and integration were not reaching the desired levels and that individual placements were more likely to produce greater outcomes.
New approaches evolved. Employers have become more involved in the working lives of people with disabilities in their employ. Employers now may include their employee assistance programs, supervisors, and coworkers in the supports to people with disabilities (Callahan & Gold, 1993; Hagner & DiLeo, 1993; Hagner & Faris, 1994; Nisbet & Hagner, 1988; Rhodes, Sandow, Mank, Buckley, & Albin, 1991; Shafer, 1986). When the Rehabilitation Act Amendments of 1992 included natural supports as an “extended service option,” it became firmly rooted in what would be described as best practices or quality services (O'Brien, 1992). Demonstrations of choice in jobs and in selection of providers of service have improved job matching, greater individualization, and the responsibility of people with disabilities to determine how resources are used (Rogan, Hagner, & Murphy, 1993). Numerous articles about the importance of work place supports have been published (e.g., Chadsey, Linneman, Rusch, & Cimera, 1997; Cimera, 2001; Hagner, Rogan, & Murphy, 1992; Rogan, Banks, & Howard, 2000; Unger, 1999).
The Americans With Disabilities Act of 1990 (P.L. 101–336) protects individuals on the basis of disability from job discrimination if they can perform the essential functions of the job with or without reasonable accommodation. Shifting demographics, the economy, and the labor market have forced employers to look toward alternative sources of labor. The desires of people with disabilities and the interests of business more clearly overlap and present an opportunity for a relationship filling both requirements (Luecking, 2000). Supported employment appears to be implemented with more sophistication now than when first defined in 1985 and when the Rehabilitation Services Administration (RSA) of the U.S. Department of Education first funded systems change grants in 1986 (e.g., Mank & Revell, 2001; Wehman, 2001).
Despite overall problems common to programs across the country (staff turnover, funding issues, waiting lists, low staff pay), the numbers of people in supported employment continues to grow. Does it naturally follow that if supported employment is growing, it is also improving now that we have more than 15 years of experience? If so, how would we know? In order to gauge whether new ideas and innovations have contributed to improvements in supported employment outcomes, we can compare the nature of jobs and the outcomes of jobs acquired in the early years with the nature and outcomes of jobs acquired in the latter years of implementation.
Mank, Cioffi, and Yovanoff have created a study design that allows for a longitudinal look at outcomes of the implementation of supported employment with a database established in 1995. We are not aware of studies intended to compare the quality of outcomes from the early implementation of supported employment to the later implementation of supported employment. This research by Mank et al. (1997a, 1997b, 1998, 1999, 2000; Jenaro, Mank, Bottomley, Doose, & Tuckerman, 2002) has yielded results regarding natural supports, typical features of employment, coworker involvement, wage, and integration outcomes. The first report in 1997 showed that better outcomes in wages and integration appear to be a function of how the features of employment are developed and negotiated. The typicalness of the job acquisition process, compensation, similarity in work roles, and initial training and orientation is positively and strongly related to wage and integration outcomes. The second report in that same year showed that people with disabilities in work settings where coworkers received training by supported employment personnel had higher wages and more typical participation in social activities with coworkers without disabilities at, and away, from work. Also, people with disabilities working in companies where training is provided about diversity or disability awareness have more typical social participation with coworkers who do not have disabilities. The third report in 1998 documented that those persons with more severe mental retardation who had better outcomes than did their peers with similar disability labels were more likely to have had positive coworker relationships and been in work settings where their coworkers had received some form of training. They were also more likely to have had fewer hours of direct and indirect support on the job from supported employment personnel. In 1999, in the fourth paper in our series, we reported practical implications for practitioners of supported employment. Results of that report suggest that providing information and support to coworkers and supervisors in the immediate area is strongly associated with better wage and integration outcomes. In addition, better outcomes were also associated with providing specific information about the support needed by an employee and providing information to coworker and supervisors just as the employee with disabilities starts the job, rather than later (Mank et al., 1999). The most recent report (Mank et al., 2000) shows that although greater hours of direct support are negatively related to typicalness, job change, length of employment, and wages, the individuals who receive greater amounts of direct supports and have had their coworkers trained have better outcomes than if there were no coworker training. These and other reports about supported employment have been informative, and the innovations that have emerged are intriguing. However, have these changes resulted in improvements in the outcomes of supported employment?
Our purpose in this paper was to look for evidence that the outcomes of supported employment may be better in more recent years of performance as compared to supported employment outcomes during the early years of implementation. More specifically, our aim was to compare the outcomes of supported employment in the early 1990s with the outcomes in the later 1990s.
Method and Survey Design
Data for this study were provided by supported employment programs considered to be developing or using “natural supports” to improve the quality, stability, and integration of community employment for people with disabilities. These programs were nominated by colleagues from across the country and regarded as dedicated to producing measurable employment outcomes. Thirteen vocational programs in 8 states (Oregon, California, Colorado, Missouri, New York, Pennsylvania, Washington, and Virginia) contributed data that included demographic, disability, and employment information. This sample was specifically chosen because the investigators wished to study outcomes for individuals and organizations believed to be implementing best practices and producing high quality outcomes.
An initial draft of the survey outlining outcomes and features of supported employment jobs included about 40 items developed over a 2-year period. Literature reviews of workforce development, diversity issues, and natural support tools (Caudron, 1992; Hess, 1993; Murphy & Rogan, 1994; Reichers, 1987) were followed by repeated revisions of the data form. Two field tests resulted in additional revisions. The final version included 62 items covering three areas: general demographic information regarding gender, age, ethnicity, and residence; specific disability information, including primary disability, disability level, and behavior; and employment outcomes and features. The employment features section, which comprised the majority of the survey, included such items as job title, hiring date, hours worked weekly, wages and benefits, accommodations, role of support staff, and worksite characteristics (size of company, number of people with and without disabilities in immediate work area, nature of training provided, and information about company's history of employing people with disabilities). Types of jobs included assembly, warehouse stocking, health or human services, child care, food service, clerical or office work, janitorial, groundskeeping, recycling or sorting, retail, and other. Several questions centered on training provided by employers to their own employees. These questions inquired about whether the company in question provided training about diversity or disability in the work place to their own employees and, if so, who provided the training and under what conditions?
We designed the final section of the employment section to allow us to gather information about the “typicalness” of the employment circumstances of employees with disabilities compared to employees without disabilities in the same business. Instructions on the data form noted that this scale was not designed as a value judgment but, rather, as a means of comparison between the typical experiences of an employee without disabilities and an employee with disabilities. For example, respondents were asked to rate “how typical” the orientation process was for employees with disabilities compared to what occurred for other new employees who did not have disabilities. Items were rated on a 7-point scale, with 1 being not typical and 7 being very typical when compared to other employees. If the employee with disabilities was oriented to a new job by a job coach and to the exclusion of the processes by which employees without disabilities are oriented, then the orientation of the worker with disabilities would be atypical. If the employee with disabilities was oriented in the same way as other new workers, then orientation would be considered quite typical. Categories on this scale included Job Acquisition and Hiring (e.g., recruitment, application, interviewing), Compensation (e.g., work schedule, hours of work, hourly pay, company benefits), Initial Orientation and Training (e.g., orientation, initial job training), Work Roles (e.g., similarity in work tasks, opportunities for variety), and Social Aspects (e.g., participation in work and non-work social activities). Typical does not necessarily mean better. For example, accommodation for interviewing or training may be important and necessary and at the same time be somewhat less typical. One item asked about integration or worksite interactions in four categories (person rarely if ever interacts, person typically exchanges greetings or has very brief social interactions, person interacts in work-related situations with people who do not have disabilities, and person engages in frequent and ongoing interactions with persons who do not have disabilities at the job site). Finally, respondents completed a scale of 1 to 7 about the employees' overall job adjustment on factors such as comparable work rate, work quality, employment becoming more typical over time, and satisfaction. The word typical was used so that the reference point for the employee with disabilities was like that of other employees in the same work place. The intent in using this particular term was to gather information about features of employment across all aspects of the human resource process, from recruitment through ongoing supervision to social aspects compared to others in the same work place.
Initially, the directors of 14 programs were contacted by telephone to gauge their willingness to participate in the research project. All 14 administrators expressed an interest and were mailed the draft data form for their review. An explanation of the confidentiality protocol was explained, which maintained a list of people for whom data were collected along with a list of coded numbers. These numbers ensured confidentiality and allowed us to correlate data about individuals over time. Follow-up phone calls were made to all 14 respondents to gather their input on how to improve the data forms. At this point, one program dropped out of the research project.
The first round of data collection occurred in July 1995 and included information on 462 employees with developmental disabilities. The second round of data collection took place in November 1996 and included information on 538 employees with disabilities. For this second instance of data collection, the survey form was expanded to include questions about job separation and company personnel in an effort to capture information about why jobs were lost and the nature and content of company supports to their employees with disabilities. Representatives of vocational programs were asked to complete surveys on all individuals for whom they had previously collected data and any additional individuals they were currently supporting in individual jobs. The third and last round of data collection occurred in March 1999 and included information on 680 employees with disabilities. The employment specialist or program staff member who best knew the person and his or her employment circumstances was asked to complete each survey. Test–retest reliability was conducted on 12% of participants randomly selected from each vocational program for all three rounds of data collection. An average of 73% test–retest agreement was obtained with a standard deviation (SD) of 8% for the first instance of data collection. Agreement for Round 2 was 87% (SD = 5%); and Round 3, 91% (SD = 6%).
Financial remuneration of $15.00 per survey completion was offered to the agencies providing the data. Program respondents were asked to submit data about individuals with disabilities employed in individual community jobs or in very small group placements. Work crews or agency-owned businesses were excluded intentionally because we were interested in situations in which supported employees were most likely to be employed directly by the business and those in which the features of employment were more likely to be typical when compared to employees without disabilities.
This analysis defines the independent variable as the year the employee was hired, which is based on the data supplied for the starting date of the current job held by the individual. All three instances of data collection (1995, 1996, 1999) are included in this analysis. The year employees were hired was collapsed into two groups, those hired in the years 1990 to 1994 and those hired in 1995 or later. Table 1 contains descriptive statistics for the two comparison groups. Using chi-square tests of independence, we found that the year employees were hired is independent from the other demographic variables and is unrelated to employee demographics. Specifically, there was no difference by year of hire for gender, minority status, disability level, behavior issues, and severity of behaviors.
Types of jobs are also reported in Table 1 and are listed in order of highest frequency to lowest frequency overall. Most employees were employed in food service or custodial/janitorial jobs (22% and 18%, respectively). The least frequent job types were recycling and human services (3% and 2%). With the exception of custodial/janitorial jobs, χ2(1) = 11.17, p < .01, there were no differences between employees hired before 1995 and those hired in 1995 or later. A significantly higher percentage of employees hired in 1995 or later were hired into custodial jobs (22% compared to 11%). Although these data illustrate that people with disabilities are acquiring entry level jobs, it may also reflect major areas of expansion of employment opportunities in the last decade.
Table 2 summarizes the comparison group percentages for coworker training, hours of direct support, and hours of indirect support. These data show that, compared to employees hired before 1995, employees hired in 1995 or later were more likely to have received weekly direct support. χ2(1) = 10.37, p < .01, and weekly indirect support, χ2(1), p < .05 (vs. receiving less than weekly support). There were no differences between the groups with respect to coworker training.
Tables 2 also contains a summary of results for disability awareness training provided in the workplace by the employer. These survey questions inquired about whether the company provided such training to their own employees and if so, to whom? Overall, about 14% of the total sampled reported that such training was provided by the employer. Persons hired in 1995 or later were more likely to be employed by businesses who offer disability awareness training, χ2(1) = 5.97, p < .01.
Tables 3 and 4 show summaries of the four overall job adjustment survey items. The average adjustment ratings overall are approximately 3 for each of the job adjustment scales (scaled 1 to 7). Persons hired in 1995 or later reported a higher rating in work rate, work quality, and positive coworker relationships.
Tables 3 and 4 also contain summaries for the four typicalness scales that range in value from 1 (not typical) to 7 (very typical). The average typicalness ratings overall are approximately 5 for each of the typicalness scales. With the exception of Compensation Package, where there were no group differences, persons hired in 1995 or later reported statistically higher typicalness ratings for job acquisition, job role, and job orientation.
Tables 3 and 4 also summarize three work-related outcome variables. People hired in 1995 or later had higher levels of worksite integration. There were no differences in hours worked or weekly and hourly wage. On average, persons were working approximately half time (22 hours per week), and their hourly wage was approximately $6.13.
Overall, these data show little or no increase in wages or hours outcomes for people in jobs acquired later in the 1990s compared to earlier in the 1990s. There are, however, improvements in the typical features of job acquisition, job roles, and job orientation. Those with jobs acquired in the late 1990s were also considered to have more positive work rate and work quality. In addition, they were considered to have more positive relationships with coworkers. There were small improvements indicated in worksite integration. These data also indicate that those hired in later years received more direct and indirect support.
Before discussing possible implications of these data, we should note possible limitations in this research, similar to the limitations noted previously by Mank et al. (1997a, 1997b, 1998, 1999, 2000). These data are based on a sample of individuals in 13 programs across 8 states. They were not randomly chosen. These individuals may not be representative of all people in supported employment. There is evidence, however, that they are comparable in levels of mental retardation to data from the Research and Training Center on Supported Employment at Virginia Commonwealth University (Revell et al., 1999) of nearly 140,000 people in supported employment.
The supported employment personnel who best knew each individual whose data are included here provided the information about persons with disabilities served by their agencies. Some bias in reporting is possible. Some of the items required a judgment on the part of the staff person providing the data (e.g., how typical is someone's employment compared to employees without disabilities in the same work setting?). It is possible that the information provided is not wholly accurate or stable. This concern has been addressed in part by collecting test–retest information on 12% of the total sample and by the fact that the data have now been collected three times with similar results. In all three iterations, the major findings have been confirmed.
Finally, because of the nature of this database, we do not suggest direct cause and effect relationships about supported employment. These data provide snapshots over time and the opportunity to look for relationships in the data. Although we might suspect that coworker involvement causes better outcomes, these data only show us that the relationships exist. There may be alternative explanations.
Is supported employment being implemented in better ways than when it initially emerged in the 1980s? Such is the question we began with for this analysis. We know that supported employment's history has been accompanied by ideas and process improvements, such as natural supports, employer leadership, better methods of work interests and abilities, improvements in job matching, increases in choice and self-determination, assistive technology, and person-centered planning. Have these kinds of ideas, however, improved the implementation of supported employment? Do people in supported employment truly benefit from these process improvements? Is there evidence of improvement in the features of employment and wage and integration outcomes of jobs acquired early in the 1990s compared to people with significant disabilities who acquired jobs later in the 1990s?
The first look at these results may be cause for disappointment. There are no real increases in wages or hours worked, two important areas for outcomes. In addition, there is no real change in the types of jobs people are acquiring. People are still largely receiving entry-level jobs in service industries. The greater typicalness worksite integration and job adjustment in the later 1990s compared to the early 1990s is encouraging, however.
One explanation for the lack of significant change in wages and integration and types of jobs may be that supported employment is simply delivering the outcomes it is capable of delivering and that implementers of supported employment have been able to do that for more than a decade. There is, however, a more complicated explanation that we argue may account for the lack of change in wages and types of jobs. The former explanation suggests that improvements or no improvements are in the realm of and controlled by the implementers. Our explanation is that it may well be the larger system issues that define the extent to which supported employment can deliver improvements over time. Wage, and to some extent, integration outcomes are largely driven by the amount of hours worked. The average amount of 22 to 25 hours worked means it may be almost impossible to show a significant difference in these outcomes in such a comparison. Further, hours worked is driven by system issues. The structure and funding of supported employment mean that it is difficult to invest heavily in work assessments, job development, job exploration, and job matching; and yet, creating excellent job matches, finding better jobs, and finding jobs with near full time hours and benefits takes longer and is more difficult. Part time jobs in the service sector of the economy are much more readily available and have expanded rapidly in the last decade. We know the best predictor of how much money someone will make in 5 to 10 years is the wage scale when they started their current job. In addition to the difficulty (and expense) in acquiring better jobs, long-term support structures and funding may make it more difficult to support more people in full time or near full time jobs. There are no financial incentives for helping people choose, get, and keep full time jobs rather than part time jobs. Further, depending on where a person lives, the structure of his or her residential circumstances can make it more difficult to work near full time. Transportation problems exacerbate the issue as well.
In addition to the structure and funding issue, systemic personal disincentives still reign in perception even if they are becoming less a barrier in regulation. Individuals and families still make decisions. The system issues and the concern about how much to work based on concerns about the loss of Social Security benefits mean that the systemic issues may define more about the outcomes of supported employment than the practitioners who work inside that system.
The improvement over the years in typicalness, in integration, and on our composite notion of overall work adjustment is encouraging. We believe these differences are the result of the quality of the work of practitioners who are making use of the innovations of the last 15 years. Greater typicalness, integration, and overall work adjustment clearly suggest that practitioners are investing in ways to ensure that people with significant disabilities are “fitting in” better in their work places. Notions of coworker involvement and natural supports are clearly associated with such improvements. This may well mean that practitioners are improving the overall implementation of supported employment everyday in the community to the extent that it can be improved given the systemic context in which it now operates.
Although these findings are focused on the community implementation of supported employment, there are also system implications. Systems and funding structures should be developed that encourage more full time work, more variety in types of jobs, investment in job development and job match, and reduction of personal disincentives. In this paper we suggest that supported employment implementers are now putting into operation innovations related to natural supports and typicalness in employment. If systemic barriers were addressed, it is clearly possible that supported employment implementers could produce improvements in job variety, work hours, and supported employment outcomes as well.
These data also show that persons hired in later years receive more direct and indirect support (support to coworkers or supervisors). These data seem difficult to interpret. It is possible that those hired more recently receive more direct support because they have been in their jobs less time or it could be the result of funding incentives to provide direct support in order to maximize billing income. Increases in indirect support may be additional evidence of investment in coworker and natural support strategies.
Supported employment now benefits tens of thousands of people world wide. Future improvements in the career outcomes for people with significant disabilities will not simply be a matter of what practitioners do in the next 10 years; it may be more of a matter of what we are able to do with the policy, funding, and systemic issues that can be improved to support the work of people in local communities.
Authors:David Mank, PhD, Director, Indiana Institute on Disability and Community, and Professor, School of Education, Indiana University, 2852 E. Tenth St., Bloomington, IN 47408 ( firstname.lastname@example.org). Andrea Cioffi, MA, Senior Research Assistant, School and Community Supports, College of Education, and Paul Yovanoff, PhD, Research Associate/Adjunct Assistant Professor, 1235 University of Oregon, Eugene, OR 97405-1235