Recent welfare reform initiatives have focused on moving adults with disabilities from income maintenance programs such as Supplemental Security Income (SSI) into paid work. Young adults with disabilities are a key group for targeted reforms because of their potential for obtaining early and long-term employment. Although there is substantial research focusing on the employment characteristics of working-age adults with disabilities, little recent data describe or explain the relationship between work and SSI participation for young adults with disabilities. This study examined data from the 1994 and 1995 National Health Interview Survey on Disability Supplement (NHIS-D) to describe the employment characteristics of young adults with disabilities, ages 18 to 29 years. In addition, this study examined the extent to which sociodemographic, education, health, disability, and program participation factors were associated with employment for this population. Findings are discussed in light of current SSI and employment policies as well as recommendations for reform.

For eligible persons with disabilities, the Supplemental Security Income (SSI) program provides cash benefits, work incentives, and linkages to Medicaid and vocational rehabilitation. Eligibility for SSI is based on an individual's "inability to engage in any substantial gainful activity by reason of any medically determinable physical or mental impairment(s) which can be expected to result in death or which has lasted or can be expected to last for a continuous period of not less than 12 months" (SSA, 1998b, p. 2). For adults 18 years and older, substantial gainful activity refers to one's ability to earn $500 or more per month through paid employment. For children under the age of 18 years, the standard is that the impairment must result in marked and severe functional limitations.

Because adult eligibility criteria include work earnings, the issue of employment among persons with disabilities has received much attention. A recent Louis Harris and Associates (1998) poll showed that the employment rate for working-age persons with disabilities in the United States has not substantially improved over the last decade despite an historically low unemployment rate for the general population. In the survey year, 79% of persons without disabilities worked fill or part time compared to 29% of persons with disabilities. Further, this survey revealed that persons with disabilities experience large gaps in terms of education, community participation, health care, income, and overall satisfaction with life. These findings are discouraging given the longstanding goals of legislation such as the Individuals with Disabilities Education Act of 1990 (IDEA) and the Americans with Disabilities Act of 1990 (ADA). Also, these data highlight employment as a key issue that must be addressed if persons with disabilities are to gain improved independent living and community integration outcomes.

For many children and adults with disabilities, SS1 is not a supplement but a primary source of income (Thornton & Sears, 199 1 ). SSI may thus only mitigate the effects of poverty for these individuals and their families rather than lift household income beyond the poverty threshold. Indeed, unless participants receive income for other sources, they may remain on SSI and live in poverty for extended periods of time (Rupp & Scott, 1995). On the other hand, SSI serves as a mechanism for ensuring a base level of income that would be threatened if SSI were not available. Social policy focusing on employment for SSI participants should therefore consider the influence of income maintenance on the employment decisions and outcomes of working age persons with disabilities. In particular, there needs to be an emphasis on youth and young adults with disabilities who demonstrate a high potential for early and continued employment (National Academy of Social Insurance, 1995; Scott, 1992).

The Personal Responsibility and Work Opportunity Reconciliation Act of 1996 mandated that all youth SSI participants undergo an age-18 eligibility redetermination. That is, youth receiving SSI benefits through the Childhood SSI program had to be reevaluated to determine whether they met specific adult SSI criteria. As a result, the Social Security Administration recommended benefit cessation for 56% of age-18 SSI participants nationally (SSA, 1998c). Initial continuance rates, or recommended eligibility rates reported at the first level of consideration, varied widely across states. For example, states such as Louisiana, Mississippi, and Arkansas had continuance rates that fell below 30%, whereas the rates of Maine, California, and Hawaii exceeded 60%. Persons recommended for benefit cessation may appeal the age- 18 redetermination decision or reapply for SSI at a later time, but the employment and quality-of-life outcomes of these individuals arc not known at present. Nevertheless, the increased risk of losing SSI eligibility underscored the importance of employment as a critical outcome for young adults with disabilities and their families.

The purpose of this study was to describe the employment characteristics of young adult SSI participants and nonparticipants with disabilities and to examine factors associated with employment for this population. By considering the contextual factors that influence employment, recommendations for improving social policies for this population may be critically assessed.

Method

Population and Sampling

The research design of this study was based on Phases I and 11 of the National Health Interview Survey on Disability (NHIS-D). Phase I of the NHIS-D was used in 1994 and 1995 to collect basic information regarding disability and to determine eligibility for Phase 11. That is, if respondents met specified disability criteria for in Phase 1, then follow-back questions were asked for in Phase 11 in 1994 and 1995. To focus on transition working-age persons, this study focused on respondents whose ages ranged from I8 to 29 years (n = 1,640; NCHS, 199 8, 1998b, 1998c, 1998d, 1998e). Given that the NHIS-D used a two-phased approach with conditional questions, the sample size for specific questions varied accordingly~ Relative sample sizes arc therefore specified for each variable. Variable descriptions and sample sizes are shown in Figure 1.

To obtain reliable population estimates, a weighting system was applied to both NHIS-D phases. Weights were used to adjust for the multistage sample design, nonresponse, and age, race, and gender characteristics (National Center for Health Statistics, 1998c). Thus, the 202,560 persons interviewed for 1994 and 1995 were scaled to estimate the 260.8 million civilians living within the United States. From the entire NHIS-D sample, the 1,640 cases that were selected represented 2.3 million civilians with disabilities between the ages of 18 and 29 years.

The overall response rate for both phases of the 1994 NHIS-D was 84.8% (National Health Interview Survey, 1998a). This figure was calculated by multiplying the response rate for the NHIS-D Phase I by the response rate for the NHIS-D Phase 11. By this same method, the response rate for the 1995 NHIS-D was 85,5% (National Health Interview Survey, 1998b). Therefore, the average response rate for the 1994 and 1995 NHIS-D was 85.1%.

Data Analysis

Data were analyzed with SUDAAN 7.5 software using specific variables available from both phases with the NHIS-D (Research Triangle Institute, 1997 ' ). This software allows standard error estimation calculations for multistage, stratified sample designs such as that which was employed for the NHIS-D. Variables addressing these research questions were combined using cases from both the 1994 and 1995 survey years.

SSI participation and non participation statuses were c1cfincd using a modification of the criteria developed by McNeil (1993) and Kruse (1997). SSI participants were identified for respondents reporting enrollment during the time of the survey month. A person was identified as an SSI nonparticipant with a disability if lie or she (a) had difficulty with one or more functional activities of daily living (ADLs); (b) used a wheelchair; (c) used a cane, crutches, or a walker; (d) had a disabling mental or emotional condition; (e ) reported a limitation in-kind or amount of work or housework he or she could do; or (f) received SSDI or was covered by Medicare. On the basis of these criteria, a dichotomous variable was created to identify responses of individuals with disabilities.

The SUDAAN 7.5 CROSSTAB and DESCRIPT procedures were used to produce national percentage, mean, and median estimates. In addition to obtaining these estimates, parametric and nonparametric tests of significance were performed for all reported comparisons. The extent to which specific sociodemographic, disability, health, education, and program participation factors were associated with employment for young adult SSI participants and nonparticipants with disabilities was addressed using the SUDAAN 7.5 LOGISTIC modeling procedure for the dichotomous dependent variable, employment during the past 2 weeks. Logistic regression modeling resembles linear regression but allows for the prediction of dichotomous c1cpendent variables rather than continuous variables. Logistic regression also allows for the interpretation of results via odds ratios.

Regression diagnostic procedures were also performed to assess the adequacy of the logistic model. Specifically, aspects regarding collinearity, tolerance, and residual analysis were performed and analyzed (Research Triangle Institute, 1997).

Results

Table I illustrates the employment characteristics of young adult SSI participants and nonparticipants with disabilities. First, the percentage of SSI participants who reported ever having worked was substantially lower than that of nonparticipants, 65.7% versus 93.1%, respectively, X2 (df = 1) = 62.16, p < .001. Second, non-SSI participants were more than twice as likely to report having worked in the past 2 weeks prior to the interview date than SSI participants, X 2 (df = 1) = 75.4, p < .001. A third finding was that, for persons who were reportedly working, the mean number of months employed for the past 12 months was not different for nonparticipants with disabilities at the .05 level of significance. Fourth, although months employed during the year were not markedly different, the mean hours worked per week were less for employed SSI participants than for nonparticipants, 27.8 versus 35 hours, t = 54.71, SE -1.78, p < .0001. In addition, job income was markedly different. The mean monthly income for working SSI participants was $607.11, whereas the mean earnings for transition-age nonparticipants was $1,116.57, t = 67.38, SE = 514.8 1, p <.000 1. Notably, the median income reported for employed SSI participants was $500 per month, the current SGA level for program eligibility.

FIGURE 1. National Health Interview Survey on Disability.

Table 2 shows findings from the employment logistic regression model. The focus of this analysis was the extent to which specific independent variables were associated with the dependent dichotomous variable, working in the past 2 weeks. The degree to which the model fit the sampled data was determined by testing whether the model and defined variables could significantly predict employment beyond the null hypothesis.

Regarding socio-demographic variables, the model showed that the employment odds for transition-age males with disabilities were 50% greater than those for females when other factors were controlled for, P = .4075, SE = .1402, p = .0039. Race analyses indicated that the odds ratio for African Americans was .63 when Caucasians were the model reference group, P = -.4616, SE = .2112, p = .0291. No statistically significant findings were evident for the region in which transition-age persons resided. Respondents living with families that had an annual total income equal to or exceeding $20,000 had an employment odds ratio of 1.63 relative to those with family incomes below this amount, = .4859, SE = .1372, p = .0004.

Education influences were substantial. Transition-age individuals with disabilities who had completed high school had odds for employment that were double of those who had not completed high school, P = .8248, SE = .4697, p = .000 1. For those with more than a 12th grade education, the odds ratio was still higher, 3.27, P = 1.1835, SE = .1992, p < .0000.

Disability influences were examined using categorization according to the presence of ADLs and/or instrumental activities of daily living (IADLs). Transition-age persons with reported disabilities but with neither ADLs or more IADLs were used as a reference group. Controlling for other factors, persons with one or more ADLs and IADLs showed an employment odds ratio of .28, P = -1.2594, SE = .2301, p < .0001. Similar patterns were observed with an odds ratio of. 56 for individuals with one or more ADLs only, P = -.5715, SE = .2576, p = .0298, and .44 for those with one or more IADLs only, P = -.8319, SE = .1920, p < .0000. Thus, transition-age persons with disabilities that resulted in greater activity limitations were far less likely to report recent employment.

Reported health status indicated significant associations with this given model. Respondents with excellent or very good health had employment odds that were almost 1.5 times greater than those with poor health, B = .9001, SE =.2959, p = 0026. T-tests showed that the B for persons reporting good or fair health were not evident at the .05 level of significance.

Hospitalization episodes within the past 12 months were negatively associated with employment. Transition-age persons with disabilities who had reported hospitalization had an odds ratio of .69 when other factors were controlled for, B = - .3716, SE = .1418, p = .0087.

Transition-age persons with disabilities who reportedly received vocational rehabilitation services were not more likely to be employed than those who did not receive these services, P = .2436, SE = .3120, p = .4353. Persons who participated in the SSI program, however, were less likely to report recent employment, P = -.6973, SE =.1778, p =.0241. Wald chi-square analyses did not reveal significant interaction between SSI participation and education, X2 (df = 2) = .79, p = .6745, or between SSI participation and vocational rehabilitation, X 2 (df = 1) = .55, p = .4592.

Discussion

The purpose of this study was to describe the employment characteristics of young adults with disabilities and to identify factors associated with employment outcomes, A number of important findings ,were produced. First, SSI participants and nonparticipants differed significantly in terms of work history, current employment, hours worked per week, and monthly, earnings. In terms of employment characteristics, SSI participants were clearly less successful than nonparticipants with disabilities. Several factors were significantly associated with employment outcomes, including the following: gender, race, family income, highest grade completed, activity limitation, health status, hospitalization, and SSI participation. Factors such as region and prior vocational rehabilitation services were not found to be significantly associated with employment outcomes for this population.

As was found by Wagner, Blackorby, Cameto, and Newman (1993) for transitioning special education students, this study also showed that persons who were African American were less likely to achieve employment outcomes than individuals who were White. These results arc consistent with Pickett's (1998) findings with regard to SSI participation and employment for all working-age persons. That is, this study showed that the odds of employment for transition-age individuals with disabilities who were White and male were substantially greater than the odds for their nonwhite and female counterparts.

In contrast to Scott (1992) and Pickett (1998), this study examined disability in the context of activity and functional limitations rather than by impairment or disability categories. The findings of this study indicated, not surprisingly, that transition-age persons who reported difficulties with ADLs or IADLs were less likely to be working than those who reported no such difficulties. These findings also parallel those of Wagner et al, (1993) in that transitioning youth with higher levels of independence across interpersonal, community, and residential domains were also more likely to achieve competitive employment outcomes.

Note. ADL = activity of daily living; IADL = instrumental activity of daily living; SSI = Supplemental Security Income, ns = nonsignificant (p < .05).

Extending the findings of Reis and Brown (1993), this study showed that self-reported health status was associated with employment for transition-age persons with disabilities. Self-reported health status of very good or excellent increased the odds of employment substantially. Also, as suggested by the results of Kruse (1997), hospitalization was negatively associated with employment for transition-age individuals with disabilities.

This study confirmed earlier findings by Wagner ct al. (1993) and Kruse (1997) that education played an important role in increasing the likelihood of employment outcomes among transition-age individuals with disabilities. This study showed that the odds of achieving employment were doubled for persons who had completed more than a 12thgrade education relative to those who had dropped out of school.

This study also built upon previous research findings regarding program participation. Similar to the findings of Burkhauser and Wittenburg (1996), this study showed that transition-age SSI participants were substantially less likely to work relative to nonparticipants with disabilities. These findings indicated that the odds of employment were reduced by more than one half for SSI participants. In addition, transition-age SSI participants were more likely than nonparticipants to participate in other income assistance programs. Burkhauser and Wittenburg suggested that the implications of participation in multiple income assistance programs may further discourage employment. Though not examined in this study's logistic analysis, transition-age SSI participants may experience increased employment disincentives when also participating in additional income transfer programs.

An earlier study showed that that transition-age SSI participants were more likely than nonparticipants with disabilities to report having received vocational rehabilitation services (Berry, Price-Ellingsrad, Halloran, & Finch, in press); the findings of the present study also indicated that vocational rehabilitation program participation was not significantly associated with employment outcomes for transition-age persons with disabilities, including SSI participants. Berry et al. found that SSI participants who exited the vocational rehabilitation were just as likely as SSI nonparticipants to achieve employment outcomes, yet the findings of the present study suggest that the overall benefit of vocational rehabilitation was negligible for transition-age persons with disabilities. A possible explanation for this may be that many persons did not achieve employment outcomes after receiving vocational rehabilitation services. For example, SSI participants achieved similar employment outcomes relative to transition-age non SSI participants, but Berry et al. still found that 45% of transition-age vocational rehabilitation participants were not employed after receiving services. Other possible explanations may be that the effectiveness of rehabilitation services does not endure over time and that periodic supports and interventions are needed to sustain employment. The overall effectiveness of the vocational rehabilitation program may therefore be mixed, given that the program is not structured to provide long-term supports.

Future research should address several specific areas relating to transition-age persons with disabilities. First, as mentioned, research should evaluate the effectiveness of systems change efforts designed to promote employment for transition-age persons with disabilities. Analyses should determine the effectiveness and quality of investments to assist persons with disabilities to enter and remain in the workforce. Second, the influences of race and gender on SSI and labor force participation should be further examined for transition-age young adults. Given that minority and women SSI participants were less likely to engage in paid employment, the relationships and policy implications of these groups should receive further attention. Third, the effectiveness of vocational rehabilitation services for transition-age SSI participants should be further examined. This study examined vocational rehabilitation in the context of logistic modeling; the interaction of SSI and differing types of rehabilitation services should also be investigated for transition-age SSI program participants. In addition, studies should examine factors such as motivation, self-esteem, and willingness to work for those transition-age participants receiving vocational rehabilitation services. Fourth, as discussed by Auxter, Halloran, Berry  and O'Mara (1999), the implementation of age-18 redeterminations as related to assessment validity and programmatic equity warrants further investigation. Finally, the roles of educators, rehabilitation professionals, and others involved in the transition planning process should be examined to determine effective technical assistance stratcgies and training that may assist individuals with disabilities and their families to effectively negotiate SSI program benefits and work incentives.

ABOUT THE AUTHOR

Hugh G. Berry, EdD, is a policy analyst with the U.S. Department of Education, Office of' Special Education and Rehabilitation Services, in Washington, DC. Address: Hugh G, Berry, 330 C St. SW, MES Rm. 3127, Washington, DC 20202.

AUTHOR'S NOTE

The views expressed by the author do not necessarily represent those of the U.S. Department of Education or the Social Security Administration.

REFERENCES

Americans with Disabilities Act of 1990, 42 U.S.C. § 12101.

Berry, H., Price-Ellingstad, D., Halloran, W., & Finch, T. (in press). Supplemental Security Income and Vocational Rehabilitation: Early work experiences among transition-age individuals, with disabilities. Journal of Disability Policy Studies.

Burkhauser, R. V, & Wittenburg, D. C. (1996). How current disability transfer policies discourage work: Analysis of the 1990 SIPP. Syracuse, NY: Syracuse University.

Fox, J. (1991). Regression diagnostics. Thousand Oaks, CA: 61.2 Sage.

Individuals with Disabilities Education Act of J990, 20 U.S.C. § 1400-1485,

Kruse, D, (1997). Employment and disability: Characteristics of employed and non-employed people with disabilities (Final re port to the U.S. Department of Labor). Piscataway, NJ: Rutgers University.

Louis Harris and Associates. (1998). National Organization on Disability/Harris survey of Americans with disabilities, New York: Author.

Menard, S. (1995). Applied logistic regression analysis. Thousand Oaks', CA: Sage.

McNeil, J. M. (199 3). Americans with disabilities: 1991-1992 (Bureau of Census, Current Population Reports, P70-33). Washington, DC: U.S. Government Printing Office.

National Academy of Social Insurance. (1995, May). Restructuring the SSI disability program for children and adolescents. Washington, DC: Author.

National Center for Health Statistics. (1989). Design and estimation for the National Health Interview Survey, 1985-1994 (Series 2, No. 110). Hyattsville, MD: Author.

National Center for Health Statistics, (1998a). National health interview survey-1994 national health interview survey on disability, Phase I and 1-1 (Series 10, No. 8A) [CDROM]. Hyattsville, MD: Author.

National Center for Health Statistics. (1998b). National Health interview survey-1995national health interview survey on disability, Phase I and Il (Series 10, No. I OA) [CDROM], Hyattsville, MD: Author.

National Center for Health Statistics. (1998c, July). National health interview survey on disability.- Workshop on data release, analysis methods, and preliminary findings. Paper presented at a workshop conducted by the National Center for Health Statistics, Hyattsville, MD.

National Center for Health Statistics. (1998d). National health interview survey 1994 national health interview survey on family), resources [Electronic data tape]. Hyattsville, MD: Author.

National Center for Health Statistics. (I 998e). National health interview survey-1995 national health interview survey (Series 10, No. 10C) [CD-ROM]. Hyattsville, MD: Author.

Personal Responsibility and Work Opportunity Reconciliation Act of 1996, 42 U.S.C. 1305.

Pickett, C. (1998, March). Quarterly report on SSI disabled workers and work incentive provisions. Baltimore: Social Security Administration, Office of Research, Evaluation, and Statistics.

Research Triangle Institute. (1997). SUDAAN User's Manual (Vol. 1). Research Triangle Park, NC: Author.

Ries, P., & Brown, S. (1993). Disability and health: Characteristics of persons by limitation of activity and assessed health status, United States, 1984-1988. Hyattsville, MD: National Center for Health Statistics.

Rupp, K., & Scott, C. G. (1995). Length of stay in SSI disability program. Social Security Bulletin, 58, 29-47.

Scott, C. G. (1992). Disabled SSI recipients who work. Social Security Bulletin, 51, 26-36.

Social Security Administration. (1998a). Demonstration program: Cooperative agreements for state projects which increase employment for individuals with disabilities who receive public support (Program Announcement No. SSA-OD-98-1). Baltimore: Author.

Social Security Administration. (1998b). Disability evaluation under Social Security (SSA Publication No. 64-039). Baltimore: Author.

Social Security Administration. (1998c). SSI welfare reform age-18 redetermination status report through December 31, 1998. Baltimore: Author.

Thornton, C., & Sears, J. (1991, November). Characteristics of adults with developmental disabilities: Evidence from the Survey of Income and Program Participation. Paper presented at the meeting of the American Public Health Association, Atlanta.

U.S. General Accounting Office. (1995a). Social security: Disability rolls keep growing, while explanations remain elusive (GAO/ HEHS Publication No. 95-34). Washington, DC: Author.

U.S. General Accounting Office (1995b). Social Security: New functional assessments for children raise eligibility questions (GAO/ HEHS Publication No. 95-66). Washington, DC: Author.

Wagner, M., Blackorby, J., Cameto, R., & Newman, L. (1993). What makes a difference? Influences on postschool outcomes of youth with disabilities. Menlo Park, CA: SRI International.