Post-placement outcomes in competitive employment: how do urban young adults with developmental disabilities fare over time?

Despite successful placements into community-based supported employment programs, thousands of previously excluded individuals with developmental and psychiatric disabilities find retaining jobs to be a major challenge (Flynn et al., 1991; Kregel et al., 1994; Wehman & Kregel, 1995). The reasons for a job separation are complex and may involve interacting factors in various levels of an individual's entire social system. Job changes may result from problems related to the family, workplace, provider agency, or community, as well as to external social, political, and economic policies. Job retention can be influenced, for example, by family responsibilities (Lustig & Thomas, 1997), work disincentives such as loss of Medicaid or Medicare health insurance, transportation problems, general turnover rates or economic downturns, social isolation on the job, burnout due to staff shortages and a stressful workplace, low pay, or absence of benefits (Lagomarcino, 1990). Diagnosis-related problems can also have an impact on job maintenance (Gervey, Parrish & Bond, 1995). Individuals with psychiatric disabilities, for example, experience more job separations than others due to relapse, sensitivity to stress, or failure to continue medications (Lagomarcino, 1990). Socio-demographic factors, such as an individual's economic status, racial/ethnic group, age, gender, and whether he or she lives in an urban/rural/suburban setting, may also be important considerations.

Other external explanations for job separations may include program weaknesses (e.g. poor job match, inadequate training, insufficient work hours, inappropriate social skills, poor job coach supervision), as well as job market problems (e.g. employer or coworker attitudes) and social policy issues (e.g. federal/state legislation, administrative, political, or economic inconsistencies and disincentives). Supported competitive employment is clearly a dynamic rather than static process. In low skill jobs, most individuals without disabilities (as well as those with disabilities) experience regular movement in and out of the labor market, with checkered patterns of job loss and job changes (Lagomarcino, 1990; Shafer et al., 1991).

In the supported employment field detailed information about job maintenance over time remains sparse (Shafer et al., 1991). Longitudinal data are difficult and expensive to collect and analysis methods are complex. Another problem with longitudinal research is the ever-changing, shifting external market forces that occur during the course of a study. The expanding market of the mid-1980s gave way in the mid-1990s to economic downsizing and rampant job instability affecting jobs in all sectors, including the service industries. These economic forces coincided with changes in the supported employment field. Vocational support models changed from providing formal training and supports to a more informal, natural supported employment model (Hagner et al., 1992). Regardless of the methodological difficulties, research on factors related to job retention in supported employment is important to help with design and implementation decisions about service eligibility, finances, and programming policies (Shafer et al., 1991).

The current study examined post-placement outcomes and tracked changes in wages, hours, and benefits in competitive employment over a 24-month period among a cohort of individuals with mental retardation, severe learning disabilities, and psychiatric disabilities. As service providers in a major urban center, an important goal of the study was identifying those risk and protective factors that might be associated with the ability to maintain a job. Lack of research on urban, nonwhite women and men with developmental disabilities from a variety of socioeconomic circumstances limits our understanding of resource allocation, access, and equity in the implementation of supported employment services.

This paper examines post-placement outcomes for individuals who worked between 20-40 hours per week for at least 30 days after they began their first placement. The outcome variables analyzed were: length of employment, job changes, leaving the employment program, and three employment outcomes over time (wages, hours worked per week, and number of benefits received).

The YAI Employment Program

The YAI employment training sites currently serve individuals with developmental and psychiatric disabilities in the New York metropolitan area. The length of the training program is three months, although a participant may obtain a job before then if she or he is ready for a job and opportunities are available in his or her chosen field. For the study reported here, participants learn vocational skills through simulated food service, maintenance and clerical work samples, and also participate in internships in banks, clinics and hospitals, offices, and retail establishments.

Placements are made in competitive integrated employment settings with pay at or above the minimum wage. Placements reflect the participant's interest and skill level. An employment training specialist (ETS) provides on-site support to the participant. This may be working along side the person and actually assisting him or her to perform the task, or it may consist of assisting the participant to process and problem-solve issues with coworkers and/or supervisors (on and off site). Participants receive additional placements if the initial or secondary placement ends in job termination or resignation. At this point an individual and his or her job are reevaluated to insure an appropriate job match.

The data reported here were obtained from a longitudinal project designed to monitor specific aspects of the job training process and to examine post-training outcomes over time for young adults with disabilities. Details on this cohort are discussed elsewhere (Levy, Botuck, & Gross, 1995; Rimmerman, Botuck, & Levy, 1995; Rimmerman, Levy, & Botuck, 1995).

Method

Procedures and Measures

Socio-demographic and background information was collected at intake. This included basic psychological and physical health information, as well as information regarding sources of family income. The employment training specialist completed monthly forms with information about the training process and actual employment outcomes (e.g., job title, industry, job site, weekly hours worked, hourly wages earned, number and type of benefits received) for as long as the individual received services.

For analyses summarized in this paper, the sample was limited to those who obtained a job placement defined as working at least 30 consecutive days for 20 to 40 hours per week. The definition of job attainment was based on guidelines from the New York City Department of Employment.

Measures

The outcome variables analyzed were: job retention and job separation; leaving the YAI program; and three employment outcomes over time (wages, hours worked per week, and number of benefits received). To clarify the myriad and complex causes for job separations, reasons for job separation were defined and collected in two ways. When the employee initiated the separation, his or her reason for separation was used (e.g., family move, financial disincentives, etc.), whereas when the employer initiated the separation, the employer's reason was used (e.g., layoff, termination due to lack of skills, inappropriate behavior, etc.).

The set of predictors tested in the analyses consisted of the following variables: diagnostic category represented as a three-category variable (mental retardation, severe learning disability or psychiatric disability); gender; age (represented either as a continuous or a four-level variable, 17-20, 21-29, 30-39, 40-60); race/ethnicity (represented as a four-category variable, African American, Latino, White, Asian American); a measure of family economic status represented as a three-category variable (individual's family received income assistance at intake, e.g. SSI, SSDI, AFDC, or Food Stamps; vs. individual's family does not receive income assistance vs. individual lives alone or in an out-of-family setting); whether the participant had experienced paid employment prior to entering the YAI employment program (yes or no); months before competitive employment (represented as a continuous variable); job category (represented as a six-category variable: food service; maintenance/housekeeping; mailroom/messenger; clerical, retail stock clerk or sales assistant; other category such as teacher's assistant, peer counselor, or veterinary assistant); and training site location (urban vs. suburban).

Analysis Methods

The statistical analyses used in this paper included Chi-square, t-tests, CHAID, and Hierarchical Linear Models (HLM). CHAID is an iterative Chi-square analysis technique (Magidson, 1993) that identifies both key predictor variables as well as interactions among them. The first step of the CHAID analysis consisted of cross-tabulating the binary dependent variable, leaving or not leaving the YAI program (yes or no), with each individual predictor. The variable with the highest Chi-square (with significance level [is less than] .05) was identified first. The analysis was repeated until significant predictors were no longer identified. The CHAID analysis adjusts the significance level in order to correct for the problem of performing multiple Chi-square tests.

To investigate longitudinal changes in job and work characteristics, a series of Hierarchical Linear Model (HLM) analyses were performed (Bryk & Raudenbush, 1992). The outcome variables were: hourly wage, number of hours worked per week, and number of benefits received. The goal of the HLM analyses was to study how each of the three dependent variables change over time for each participant, and to detect predictor variables which may explain individual differences. Data for the three time-varying dependent variables across 24 consecutive monthly time periods were available for all of the employed participants. Because of attrition at various times over the 24 months after program entry, participants had varying numbers of data points. The problem of "missing data" across subjects, however, produces no difficulty for HLM procedures, but results for subjects with more data points are estimated with greater precision.

Results

Participant Characteristics at Entry

Over a two-year period a total of 109 individuals with mental retardation, severe learning disabilities, and psychiatric disabilities who enrolled in the training program between January 1991 and December 1992 were placed in competitive employment positions lasting at least 30 consecutive days. Background information is presented in Table 1. Forty-four percent of the participants had a primary diagnosis of mental retardation with most functioning in the moderate to mild range. The diagnosis of severe learning disability was based on the federal definition as determined by school personnel and vocational rehabilitation counselors (Carney, 1990). Eighty percent of the participants were younger than 30 years old and 80% lived in New York City. None of the participants was married, but 8% had children (12% of the women and 6% of the men). Most lived with family members (83%) and 35% of their families were receiving some type of income assistance.

Table 1 Frequency Distribution of Participants'
Characteristics

Characteristics                    f        %

Age:
    17-20                          55     50.5
    21-29                          32     29.4
    30-39                          14     12.8
    40-60                           8      7.3

Primary Diagnosis:
    Mental Retardation             48     44.0
    Severe Learning Disability     41     37.6
    Psychiatric Disability         20     18.4

Secondary Diagnosis (n=36)
    Mental Retardation              1      0.9
    Learning Disability             3      2.8
    Psychiatric Disability         11     10.1
    Other                          21     19.2

Gender:
    Males                          67     61.5
    Females                        42     38.5

Race:
    African American               48     44.0
    Latino                         32     29.4
    White                          27     24.8
    Asian-American                  2      1.8

Residence:
    Live with Family               91     83.5
    Group Home                      4      3.7
    Supported Apartment             6      5.5
    Independently                   8      7.3

Family Receives Public
Assistance Benefits (n=91):
    Yes                             32    35.2
    No                              59    64.8

Children:
   Yes                               9      8.3
   No                              100     91.7

Timing of job placement, job characteristics, and job duration

Overall, the average length of time it took for participants to obtain a job lasting at least 30 days was 6 1/2 months after program entry (SD = 3.92, range = 2.5 to 17.5 months, median = 6 months). Men obtained jobs at a faster pace than did women (5.8 months vs. 7.5 months, t (107) = 2.21, p [is less than] .05). Participants' first jobs were primarily in maintenance (30%) and retail (24%) sectors. Messenger, clerical, and maintenance positions were found primarily in urban sites and food service positions were available primarily in the suburban site [x.sup.2] (5, N= 109)=13.75, p [is less than] .05. Women were more likely to find food service, clerical, or retail positions; men worked as messengers [x.sup.2] (5, N = 109) = 11.99, p [is less than] .05. The longest consecutive period of time that study participants were employed ranged from 6 weeks to 24 months when data collection ended. Twenty-eight percent of the participants worked between 12 and 24 months, 24% worked 6 to 11 months, 27% worked 2 to 5 months and 22% worked between 6 weeks and 2 months. When the 30 people who worked one year or more were compared to those who worked a shorter period of time, no statistically significant socio-demographic or other factors were identified.

Job Separations

Frequent job turnover among low-skilled jobs can be expected for those without disabilities as well as those with disabilities. A total of 84 study participants (77%) changed jobs at least once (14 people changed jobs twice and four people changed jobs three times). Five people who experienced layoffs and three people who left their job for a better job secured second jobs immediately or within one week, while others experienced a gap of three to 64 weeks before attaining their second job.

The reasons for a participant's first job change or termination can be categorized as primarily employer initiated, employee choice, or mixed (as in the case of layoffs). Layoffs can be interpreted as an employer decision reflecting job market factors, but employee shortcomings may also play a role. In our sample, layoffs affected 35% of those individuals experiencing their first job change (n=29). Thirty-seven percent of the first job changes (n=31) were due to employer terminations (8 terminations due to absenteeism, 12 due to a lack of skills, 11 as a result of behavior problems). Twenty-one individuals left their first job of their own volition (five individuals found a better job, seven quit because they didn't like the job, six reported financial disincentives, and three moved out of the area).

Those who were employed 12 or more consecutive months were significantly more likely to change their first job by their own choice, while those who were employed less than a year changed jobs because of employer-initiated reasons (layoffs, absenteeism, lack of skills, or inappropriate behavior). Compared to those working less than a year, 67% of individuals holding jobs for more than a year were more likely to change their first job on their own volition (e.g. quit, financial disincentives, better job, or moved) vs. 22% for those with shorter job duration (Fisher exact 2-tail p-value = .004). The results were similar when the layoff category was excluded from the analysis (N = 55, Fisher exact p = .015). For those who changed jobs a second time the reasons given were similar to the first job changes. None of the predictors was related to either experiencing a job separation or the number of job separations experienced (e.g., age, gender, diagnostic category, race/ethnicity, family economic status, months before competitive employment, previous paid employment experience, and training site location). Having a job separation, however, was highly predictive of leaving the YAI program.

Hours worked per week, hourly wage, and number of benefits

The results of the Hierarchical Linear Model analysis for hours worked per week identified gender as the only statistically significant variable which distinguished between the number of hours people worked over the 24-month period. On the average, women worked about four hours fewer per week than men (24 hours vs. 28 hours per week), and participants' hours per week did not increase or decrease over time. No other factors were statistically significant (age, race, diagnostic category, family economic status, previous paid employment experience, or training site location).

A preliminary analysis showed the average hourly wage over the 24-month period was $5.50. Further, the HLM found diagnostic category significantly predicted initial and over time differences in hourly wage. (It should be noted that whereas the difference between MR/LD and PSYCH/LD is large and statistically significant, this is not the case for the MR/PSYCH comparison whose probabilities were similar. Therefore, the MR/PSYCH group is combined in Figure 1.) Initially, individuals with learning disabilities differed from the MR/PSYCH group, and earned significantly higher hourly wages than did individuals with mental retardation or psychiatric disabilities ($6.05 vs. $4.99 per hour). Over time, however, while the wages of those with severe learning disabilities remained the same, the wages of the others increased approximately nine cents a month. After ten months, those with mental retardation and psychiatric disabilities surpassed workers with learning disabilities, and at 24 months earned significantly higher wages ($7.37 vs. $6.05 per hour). Other factors tested were not significant predictors of hourly wage (i.e., age, gender, race, family economic status, previous paid employment experience, or training site location).

[Figure 1 ILLUSTRATION OMITTED]

The number of benefits received by study participants is presented in Figure 2. None of the predictor variables tested was associated with the number of benefits an individual received at baseline. Overall, the number of benefits an individual received increased over time. Diagnostic category predicted the rate of benefits acquisition over time. Individuals with psychiatric disabilities increased the number of benefits they received to an average of all five (paid sick days, paid vacation days, medical insurance, dental insurance, retirement/pension benefits), whereas those with mental retardation and severe learning disabilities increased two or fewer. No other factors predicted initial or over time benefits (e.g., age, gender, race, family economic status, previous paid employment experience, or training site location or number of hours worked).

[Figure 2 ILLUSTRATION OMITTED]

Attrition/leaving the employment program

Fifty percent of those who worked 30 consecutive days left the YAI program (some even after working 18 or more consecutive months). Those who experienced job changes were more likely to leave the program compared to those who did not change jobs. The CHAID analyses indicated that 64% of those who experienced at least one job change left the program compared to none of the people who did not change jobs. In the second stage of CHAID analysis, among those who changed jobs and left the program, individuals living in families who received income assistance were more likely to leave the program (90%) compared to others who had job changes but whose families did not receive income assistance (58%). Other factors did not predict leaving the program (age group, gender, race, diagnostic category, prior paid employment experience, training site location, or months before competitive employment.)

Discussion

This study examined post-placement outcomes in competitive employment over time among a group of individuals with mental retardation, severe learning disabilities, and psychiatric disabilities living in a major metropolitan area. Its major contributions are the longitudinal microanalysis of a cohort of primarily African American and Latino individuals and the tracking of changes in their job status over a 24-month period. The findings reflect the difficulties of maintaining competitive employment for individuals with developmental, severe learning, and psychiatric disabilities (Flynn et al., 1991; Kregel et al., 1994; Lagomarcino, 1990). This difficulty is illustrated by the finding that 28% of the individuals retained their jobs for 12 to 24 months and 72% changed their jobs during the 24-month period. This percentage is similar to the 30% retention rate reported by Shafer et al. (1991) for 302 individuals with mild and moderate mental retardation from the Virginia Supported Employment Information System.

The surprising finding is that no significant factors were associated with job retention, suggesting that job retention and separation are situational and reflect a combination of circumstances related to the person, the employer, and the job market (Lagomarcino, 1990). Additional evidence that the problem is multifaceted rather than due to specific determinants is indicated by the wide range of time it took for individuals to rejoin the workforce -- from three to 64 weeks.

While it is unclear which factors lead to job separation, the findings reveal that in comparison to individuals who worked less than one year, those who worked from one to two years were more likely to change jobs of their own volition. One possible explanation for this finding is that those who remained employed longer gained more confidence and self-esteem and were ultimately able to leave a job for a more interesting, higher paying, more challenging, or easier job, whereas those who were terminated worked for a shorter period of time. In a study of job separation in Illinois, Lagomarcino (1990) found that 18% of job separations were explained by "lack of job responsibility" (does not want to work, poor attendance/tardy, poor work attitude, criminal behavior) and 17% were due to employer financial problems (seasonal layoffs, economic downsizing). To a lesser extent, other reasons for 10 to 13% of the Illinois separations were: task production problems (low quality, slow rate, prompting required); social-vocational behavior problems (poor social skills or appearance, insubordination, difficulties following directions or responding to criticism); positive job status changes (took a better job, support services not needed); and miscellaneous reasons such as moving away, program termination, transportation, or parental choice (Lagomarcino, 1990). Earlier studies of job terminations among food service employees found no significant differences between those with and without mental retardation; social and production problems were the major contributors to job terminations for both groups (Martin et al., 1986). Future research will need to more carefully examine the personal, social, and situational predictors of job retention, identifying those supports needed to keep people working and succeeding at their jobs. The 24-month analyses in our study indicate that wages, hours worked, and number of benefits surpass expectations for sheltered work settings. However as Mank (1994) has observed, this compensation does not approach the self-sufficiency employment outcomes of the rest of society. Although people in our study worked many hours, and 35% worked 40 hours per week or more, almost everyone including full-time employees remained eligible for some form of public assistance by earning less than $810 a month. This finding raises serious questions regarding the economic survival of working individuals with mental retardation and developmental disabilities. Certainly one solution is to raise the minimum wage to a living wage. Some advocates have suggested a guaranteed income supplement to ensure flexible employment which maximizes each individual's potential (Walter, 1993; Withorn, 1993).

The data regarding lack of benefits is particularly troublesome considering the higher incidence of medical and mental health issues for this group and the concomitant increases in cost (NASDDS, 1995). Careful planning is needed to guarantee that the full range of medical, dental, and mental health insurance benefits be provided for all individuals, including the coverage which will insure the particular needs of individuals with disabilities. Proposed revisions in Social Security programs (Batavia & Parker, 1995) must be reviewed carefully and the preservation of Medicaid entitlements is imperative.

The finding that women in our study worked fewer hours than men may reflect women's competing parenting, child care, or other family responsibilities in the home. Data were not available from this study to explore these issues. More detailed investigations of the extent of an individual's family responsibilities may be an important factor in understanding patterns of job placement and job retention (Botuck, Levy, & Rimmerman, 1996).

We found that diagnostic category at job entry and over time was related to both hourly wage and number of benefits. Initially, individuals with learning disabilities received higher hourly wages than did those with mental retardation and psychiatric disabilities. Over time, however, individuals with mental retardation and psychiatric disabilities received increases in their hourly wages, whereas those with learning disabilities received no such increases. It is possible that the initially higher hourly pay rate for individuals with learning disabilities reflects a more diverse range of job responsibilities open to them. The finding that individuals with psychiatric disabilities were more likely to receive both increased wages and benefits over time may be an indication of their superior and unexploited skill level at job entry and that they were rewarded for their performance over time.

Unfortunately many of those who experienced a job change left the program. Individuals who were members of families receiving public assistance were more likely to leave the program after a job change as compared to individuals who were from families not receiving public assistance. While our current data do not give us definitive reasons, we can speculate as to why some individuals left the program after a job separation. As proposed by Coulter (1996), an ecological understanding of how risk factors from multiple dimensions interact might suggest that nonvoluntary job separations could have an adverse impact on the employment aspirations of an entry level worker with limited skills, especially a person with developmental or psychiatric disabilities. Expectations about failure, self-esteem, self-perceptions of capabilities, and feelings of control over outside factors, as well as actual support persons and realistic alternative employment opportunities, all interact in determining reentry into the workforce after a job separation. Future programs serving people with disabilities from poor, urban areas will require linkages to extended supports and counseling at the point of job separation - an active crisis intervention approach which can be implemented at the point of job separation. This may help to avert the withdrawal of the individual from the employment market after a job loss. This type of crisis intervention can include natural, community-based supports as well as job club and clinical counseling services. In a review of supported employment over the past 10 years, Wehman & Kregel (1995) have expressed their concern with momentum of "support" for supported employment at the state and national levels. The findings from this study affirm the need to advocate for programs that respond to complex individual needs.

Acknowledgements

The authors would like to express our appreciation to Michael Kramer and the staff of the Department of Employment Initiatives of YAI for their assistance and for generously sharing their knowledge and expertise. We also thank Alan L. Gross for his recommendations and consistent assistance with the analytical procedures and Jacqueline M. Royce for her assistance on an earlier version of the manuscript.

References

Batavia, A.I. & Parker, S.B. (1995). From disability roles to payrolls: A proposal for Social Security program reform. Journal of Disability Policy Studies, 6(1), 73-86.

Botuck, S., Levy, J.M., & Rimmerman, A. (1996). Gender related differences in placement rates of young adults with mental retardation and severe learning disabilities. International Journal of Rehabilitation Research (19), 259-263.

Bryk, A. S. & Raudenbush, S.W. (1992). Hierarchical linear models: Applications and data analysis methods. Newbury Park, CA: Sage Publications.

Coulter, D.L. (1996). Prevention as a form of support: Implications for the new definition. Mental Retardation, 34(2), 108-116.

Flynn, T., Wacker, D., Berg, W., Green, K., & Hurd, R. (1991). Long-term job retention of workers placed in supported employment. Vocational Rehabilitation, (January), 25-34.

Gervey, R., Parrish, A., & Bond, G. (1995). Survey of exemplary supported employment for persons with psychiatric disabilities. Journal of Vocational Rehabilitation, 5(2), 115-126.

Kregel, J., Parent, W., & West, M. (1994). The impact of behavioral deficits on employment retention: An illustration from supported employment. NeuroRehabilitation, 4(1), 1-14.

Lagomarcino, T.R. (1990). Job separation issues in supported employment. In F.R. Rusch (ed.). Supported employment: Models, methods, and issues. Sycamore, IL: Sycamore Publishing.

Levy, J.M., Botuck, S., & Gross, A.L. (1995, June). Predictors of employment outcomes for urban youth with disabilities: Factors leading to success over time. Paper presented at the 119th Annual Meeting of the American Association on Mental Retardation. San Francisco, CA.

Lustig, D. & Thomas, K. (1997). Adaptation of families to the entry of young adults with mental retardation into supported employment. Education and Training in Mental Retardation and Developmental Disabilities, 32(1), 21-31.

Magidson, J. (1993). SPSS for Windows: CHAID, Release 6.0. Chicago, IL: SPSS Inc.

Mank, D. (1994). The underachievement of supported employment: A call for reinvestment. Journal of Disability Policy Studies, 5(2), 1-24.

Martin, J., Rusch, F.R., Lagomarcino, T.R., & Chadsey-Rusch, J. (1986). Comparison between nonhandicapped and mentally retarded workers: Why they lose their jobs. Applied Research in Mental Retardation, 7(4), 467-474.

National Association of State Directors of Developmental Disabilities Services. (1995, July). NASDDS policy statement: Managed care and long-term supports for people with developmental disabilities. (Available from NASDDS, 113 Oronoco Street, Alexandria, VA 22314).

Rimmerman, A., Botuck, S., & Levy, J.M. (1995). Predicting the likelihood of job placement: A short-term perspective. Journal of Rehabilitation, 61(1), 50-54.

Rimmerman, A., Botuck, S., & Levy, J.M. (1995). Job placement for individuals with psychiatric disabilities in supported employment. Psychosocial Rehabilitation Journal, 19(2), 3744.

Shafer, M.S., Banks, P.D., & Kregel, J. (1991). Employment retention and career movement among individuals with mental retardation working in supported employment. Mental Retardation, 29(2), 103-110.

Walter, T. (1993). Basic income in Europe (with special reference to Britain): The story so far. Journal of Progressive Human Services, 4(1), 19-28.

Wehman, P. & Kregel, J. (1995). At the crossroads: Supported employment a decade later. Journal of the Association for Persons with Severe Handicaps, 20(4), 286-299.

Withorn, A. (1993). Women and basic income in the US: Is one man's ceiling another woman's floor? Journal of Progressive Human Services, 4(1), 29-42.

Shelly Botuck, Ph.D., Coordinator of Research, Research Department, YAI/National Institute for People with Disabilities, 460 West 34th Street, New York, NY 10001