Impact of Participation in School-to-Work Programs on Postsecondary Outcomes for Youth With Disabilities From Low-Income Families

Abstract

Individuals with disabilities who also live in poverty face a double jeopardy. Disability and poverty are each separately associated with poorer education and employment outcomes. One approach to ameliorate these poorer outcomes is to improve the transition from high school to adulthood. Using data from the National Longitudinal Study of Adolescent to Adult Health, this article examines the role of school-to-work training programs on adult outcomes for individuals with disabilities who live in welfare receiving households. A linear probability model identifies the differences in outcomes for youth by disability and welfare status. Participation in school-to-work programs for youth with disabilities from welfare receiving homes was found to predict higher rates of employment, lower rates of conviction, and lower wages. Implications of these results and recommendations for future research design are included.

Keywords

disability poverty school-to-work training programs adult outcomes

Poverty snares opportunities; it reduces the likelihood of engagement in work and education, especially for youth with disabilities (Wagner, Newman, & Javitz, 2014). Youth living in poverty experience a precarious environment that includes factors such as living in neighborhoods with high crime rates (Ellen & Turner, 1997), attending schools that are typically underfunded and lack appropriate educational programming to meet student needs (Davies, Rupp, & Wittenburg, 2009), lacking parental engagement in school, and experiencing lower expectations of postsecondary achievements (Park, Turnbull, & Turnbull, 2002). Wagner et al. (2014) used the National Longitudinal Transition Study 2 data to demonstrate that socioeconomic status (SES) itself explained about 25% of the probability of dropping out of high school, 60% of the probability of not engaging in competitive employment, and about 50% of the probability of not pursuing postsecondary education. Furthermore, Karpur, Nazarov, Brewer, and Bruyère (2014) illustrated that poverty explained one third of the lower probability of attending postsecondary education among youth with parents receiving public welfare.

Loprest and Wittenburg (2007) examined the transition from adolescence to adulthood for youth receiving Supplemental Security Income (SSI) and found that two out of five SSI youth never completed high school, one out of five were incarcerated, nearly two thirds were never engaged in training leading to employment or work, and between 30% and 50% remained on public welfare well into adulthood. Leventhal and Brooks-Gunn (2000) argued that multiple factors likely impact transition outcomes for these vulnerable youth, including lower expectations for postsecondary independence and engagement in work. Furthermore, the literature clearly demonstrates the transmission of poverty from parent to child and the lack of intergenerational income mobility in the United States (Chetty, Hendren, Lin, Majerovitz, & Scuderi, 2016; Solon, 1999). A pattern of dependency on welfare across generations in families has also been observed; such trends potentially discourage labor market engagement (Mitnik, 2007).

Despite the high utilization of safety-net programs by individuals with disabilities (Houtenville & Brucker, 2014), there is mixed evidence of the impact of public welfare programs like SSI. Duggan and Kearney (2005) identified a positive impact of participation in SSI on material hardship experiences for eligible families while Ghosh and Parish (2015) utilized the same data source to illustrate that households that received SSI experienced similar levels of hardships to other low-income households. Furthermore, Stapleton, Burkhauser, She, Weathers, and Livermore (2009) described welfare policies as being a “poverty trap” that has built-in disincentives for people with disabilities to engage in work.

Recognizing that many of the poorer outcomes for youth living in poverty are a function of limited family empowerment, engagement in education, and employment opportunities (Hemmeter, Kauff, & Wittenburg, 2008), the Social Security Administration (SSA) has supported transition programs to engage youth in work. The recent Youth Transition Demonstration (YTD) project, based on Guideposts for Success (National Collaborative on Workforce and Disability for Youth [NCWD/Y], 2012), used a combination of transition-related services and SSA waivers to allow youth to retain benefits as they worked in paid jobs. While findings were mixed, the evaluation of YTD generally indicated positive program effects of higher youth employment rates in three sites, and these effects were correlated with the intensity of services across the three successful sites (Fraker, Mamun, Honeycutt, Thompkins, & Valentine, 2014).

A separate strand of research has examined the role of participation in work-related training on labor market outcomes for SSI youth. These studies found important differences depending on the features of the training program or individual characteristics of the youth. Wagner, Newman, and Javitz (2016) reported no overall benefit on adult full-time employment of career and technical education (CTE) in high school for students with learning disabilities but did observe a positive effect of occupationally specific CTE on adult outcomes. Berry and Caplan (2010) reported poorer outcomes for vocational rehabilitation (VR) participants overall in means-tested welfare programs but also noted significant variations across individual characteristics and VR services. Specifically, youth with intellectual disabilities were more likely to be employed 2 years post exit compared with those with other disabilities, but these youth had lower earnings. Other researchers observed a positive impact of vocational training and rehabilitation programs for youth with disabilities as they transitioned to early adulthood (Hemmeter, Donovan, Cobb, & Asbury, 2015; Hoffman, Hemmeter, & Bailey, 2016).

Based on these studies, it is evident that services that provide opportunities for youth to engage in career development and work have positive, albeit mixed, effects on the employment of youth with disabilities living in poverty with welfare assistance. These findings form the basis of the current analysis that uses a nationally representative secondary dataset to understand the mitigating impact of participation in career development activities on postsecondary employment and earning success for youth with disabilities living in families receiving welfare benefits.

Method

Data

This research utilized the National Longitudinal Study of Adolescent Health (Add Health). The Add Health is a nationally representative school-based longitudinal study of youth in Grades 7 through 12 within the United States during the 1994–1995 academic year. The initial goal of Add Health was to study the factors that impact risk-taking behaviors and health outcomes of adolescents in the United States. As the respondents have aged, the objectives of the survey waves have broadened to include developmental and health trajectories. The study began with an in-school questionnaire and then continued with a series of in-home interviews of the youth. Additional surveys collected information from parents/guardians, siblings, and partners, as well as school administrators and peers. The study used a clustered sampling design, sampling schools, and then students. The first survey (Wave I) collected information from 20,745 youth and 17,670 matched parents, with a 79% response rate. Following this, three additional waves of data have been collected to create a longitudinal panel of youth survey data file: Wave II was conducted in 1996; Wave III in 2001–2002; and Wave IV in 2008. The fourth wave recorded data from the youth when they were between 24 and 32 years old. Mann and Wittenburg (2015) found that wage gaps emerge at age 24 for individuals with disabilities, which makes the age range in this dataset appropriate to assess predictors in wage differences for individuals with disabilities. Given the clustered sample design and response rate of more than 80%, the final sample weights contained within the data files are designed to address bias due to nonresponse (Brownstein et al., 2010; Chantala, Kalsbeek, & Andraca, 2005).

The data sample for this analysis was drawn from multiple Add Health files. Wave I contributed youth and parent demographic information, youth and parent disability measures, and secondary school information for this analysis. Several education measures were used from Waves II, III, and IV, and information about school-to-work (STW) training programs was drawn from Wave III. Finally, adult outcomes, including highest level of schooling, employment history, employment status, earnings, and criminal record, were pulled from Wave IV. Our baseline analytical sample contains 8,584 individuals without missing data in any of the regression model variables, from the original 20,725 youth who participated in the first wave of the survey. Due to missing wage data, a smaller subset of 6,952 individuals is used to study earnings outcomes.

Independent Variables

School-to-work participation

The variable of interest in this study was a measure of the youth’s participation in a STW program. In Wave III, youth were asked whether they had received vocational education or job training in a program that “lasted or will last for at least 3 months.” Our measure captured youth who responded affirmatively to the above question and also stated that the program was offered by their high school.

Household beneficiary receipt

This indicator variable, which identified youth from households receiving welfare, was a function of parents’ responses in the Wave I survey. Parents who responded affirmatively with regard to the receipt of Assistance to Families with Dependent Children (currently Temporary Assistance to Needy Families), Food Stamps, or SSI were flagged as indicating household beneficiary receipt.

Youth disability status

In this study, youth were identified as having a disability if they responded affirmatively to Wave I survey questions that inquired about use of special education services; having a developmental, visual, hearing, or learning disability; or having a difficulty using hands, feet, arms, or legs.

Youth characteristics

The individual characteristics collected about the youth in the analytical dataset included age, gender, race/ethnicity, and self-reported health. The Peabody Picture Vocabulary Test (PPVT), an instrument of verbal ability and aptitude, was also included in the dataset to provide information about heterogeneous ability across the youth populations (Blau, 1999). Furthermore, information about youth incarceration was included.

Other controls

Our analyses also included maternal characteristics. These data were collected from the parent interview. To focus on the impact of being raised in a household that received public benefits, we selected maternal variables including marital status, education, biological relationship to youth, and maternal disability status. Following the earlier work of Karpur et al. (2014), we included school characteristics: overall school size, average class size, and a measure of vocational focus in the program.

Dependent Variables

The objective of this analysis was to describe how STW programs impact transition-to-adulthood outcomes of young adults, particularly for those with disabilities and who come from homes receiving welfare. Consequently, we focused on three major events for young adults: education, criminal record, and employment. Extending the work of Karpur et al. (2014), we used the highest educational attainment measure from Wave IV to assess college attendance. Individuals who attended a 2-year or 4-year postsecondary school were classified as attending college, while for the purposes of this study, vocational postsecondary school was not classified as college attendance. To investigate criminal records among this population, we focused on those convicted of or who pled guilty to any charges other than a minor traffic violation, which was recorded in the fourth wave of the survey. We also assessed the impact of STW programs on employment by considering whether the youth had any history of paid work, was currently employed, and by examining their hourly wages.

Empirical Framework

In this article, we attempt to understand how the impact of STW programs differs for youth by disability status and household beneficiary status. Thus, there are eight groups we wish to study in this article, based on three binary measures: disability status, household beneficiary status, and participation in STW programs. We focus in particular on the group that encounters a dual disadvantage—that is, youth with a disability from a welfare beneficiary home (Karpur et al., 2014).

Regression models form the basis of the empirical framework in this analysis. They allow us to control for observed characteristics at the youth, parent, and school levels that would otherwise confound our understanding of the differential impact of STW programs. For example, a person’s highest level of education is highly correlated with both employment and earnings. Moreover, the literature has well established that individuals with disabilities face lower likelihood of participation in college and employment (Erickson, Lee, & von Schrader, 2014), so the inclusion of measures that influence these young adult outcomes allows us to more accurately estimate the impact of all the independent variables.

Specifically, we use a linear probability model that estimates the likelihood of a particular event for the binary dependent variables in this study. The empirical specification used to assess these outcomes is

\begin{array}{l} Y_{i j} = β_{0} + β_{1} D_{i j} + β_{2} B_{i j} + β_{3} D_{i j} B_{i j} + β_{4} S_{i j} + \\ β_{5} S_{i j} D_{i j} + β_{6} S_{i j} B_{i j} + β_{7} S_{i j} D_{i j} B_{i j} + γ X_{i j} + c_{j} + ε_{i j}, \end{array}

(1)

where Y_ij is the outcome of interest, that is, college attendance, history of employment, or conviction record, for individual i from school j. The indicator variables, D_ij, B_ij, and S_ij, are equal to one when the youth has a disability, comes from a beneficiary household, or participated in a STW program, respectively. X_ij is a set of control variables including age, gender, grade retention, self-reported health, PPVT score, maternal characteristics, and school characteristics. Unobserved school fixed effects are captured by c_j, and the remaining idiosyncratic error is represented by ε_ij.

The coefficients β₁ through β₇ represent seven of the eight groups of interest with the underlying reference group being youth without a disability and not from a beneficiary home who did not participate in a STW program. For example, the predictive role of youth disability status is estimated through β₁. Likewise, the β₇ estimator captures the predictive role of disability, benefit receipt, and STW participation interacted with each other (i.e., the Disability × Benefit × STW interaction estimate).

To understand the role of STW programs on earnings, a separate model designed for a continuous measure of wages must be employed. The field of labor economics relies upon the framework established by Mincer (1974) to examine the components that predict earnings. Card (1999) thoroughly discussed the history and various modifications of the Mincer model. The Mincer equation regresses log wages on measures of education, experience, and other attributes. Using a linear regression specification, the Mincer model in this article is

\begin{array}{l} W_{i j o} = δ_{0} + δ_{1} D_{i j o} + δ_{2} B_{i j o} + δ_{3} D_{i j o} B_{i j o} + δ_{4} S_{i j o} + \\ δ_{5} S_{i j o} D_{i j o} + δ_{6} S_{i j o} B_{i j o} + δ_{7} S_{i j o} D_{i j o} B_{i j o} + λ Z_{i j o} + \\ c_{j} + h_{o} + υ_{i j o}, \end{array}

(2)

where W_ijo are log wages of individual i from school j in occupation o. Disability, welfare receipt, and participation in a STW program are all modeled as above. Another difference in this model from above is the inclusion of 23 occupational fixed effects based on the first two digits of the Standard Occupational Classification for the job associated with wage, W_ijo. The sample here also differs from the above model’s sample in that it is restricted to individuals with a history of employment.

Results

Descriptive Analysis

A raw gap analysis is presented in Table 1, which compares the means of youth demographic measures across the four groups defined by disability and welfare status: nondisability and nonbeneficiary, nondisability and beneficiary, disability and nonbeneficiary, and disability and beneficiary. Panel A of Table 1 shows that 44% of the youth without disability or benefit status were male while 53% of the youth with disabilities from beneficiary homes were male. Black youth without disabilities were more than twice as likely to have come from a beneficiary home (16% vs. 36%) and that disproportionately jumped to nearly a threefold gap for black youth with a disability (11% vs. 29%). Both disability status and welfare receipt are associated with higher rates of grade retention, with the occurrence of both being an example of double jeopardy (12% vs. 28%, 12% vs. 38%, and 12% vs. 60%). Participation in STW programs ranges between 1.4% and 3.2% of the study sample. As shown by Karpur et al. (2014), youth with a disability from beneficiary households are more likely to have mothers who are unmarried, have disabilities, do not participate in their school’s parent teacher organization (PTO), have lower educational attainment, and have poorer health than the parents of youth without a disability from nonbeneficiary homes. Panel B of Table 1 describes the youth sample by disability type and shows that more than 80% of the youth with a disability have a cognitive disability. The majority of these students receive special education services, with a higher frequency among youth from beneficiary homes (54% vs. 69%), possibly indicating higher academic support needs of these youth.

Table 1.

Study Youth Demographics.

Demographic	(1)Nondisability & nonbeneficiary	(2)Nondisability & beneficiary	(3)Disability & nonbeneficiary	(4)Disability & beneficiary
Panel A
Male	0.44 (0.50)	0.38 (0.49)	0.60 (0.49)	0.53 (0.50)
Black	0.16 (0.37)	0.36 (0.48)	0.11 (0.31)	0.29 (0.46)
Hispanic	0.15 (0.35)	0.21 (0.41)	0.12 (0.33)	0.19 (0.39)
Asian	0.05 (0.22)	0.04 (0.20)	0.02 (0.13)	0.01 (0.11)
Other	0.05 (0.21)	0.06 (0.23)	0.05 (0.23)	0.04 (0.21)
Grade-level	9.62 (1.62)	9.38 (1.62)	9.43 (1.59)	8.95 (1.52)
Ever retained	0.12 (0.33)	0.28 (0.45)	0.38 (0.49)	0.60 (0.49)
S.R. health	2.05 (0.87)	2.21 (0.95)	2.26 (0.93)	2.33 (1.01)
Mental health	0.20 (0.40)	0.20 (0.40)	0.26 (0.44)	0.27 (0.44)
STW training	0.01 (0.12)	0.03 (0.18)	0.02 (0.12)	0.02 (0.15)
PPVT score	103.96 (13.34)	95.03 (14.65)	98.30 (13.45)	89.39 (13.78)
Panel B
Cognitive disabilities	0.00 (0.00)	0.00 (0.00)	0.82 (0.39)	0.83 (0.38)
Physical disability	0.00 (0.00)	0.00 (0.00)	0.06 (0.24)	0.10 (0.30)
Special education	0.00 (0.00)	0.00 (0.00)	0.54 (0.50)	0.69 (0.47)
Blind	0.00 (0.00)	0.00 (0.00)	0.01 (0.09)	0.01 (0.09)
Deaf	0.00 (0.00)	0.00 (0.00)	0.01 (0.10)	0.00 (0.06)
Cognitive impairment	0.00 (0.00)	0.00 (0.00)	0.03 (0.17)	0.04 (0.20)
Learning disability	0.00 (0.00)	0.00 (0.00)	0.80 (0.40)	0.82 (0.38)
Observations	n = 6,353	n = 1,001	n = 958	n = 272

Note. Means with standard deviations in parentheses. S.R. Health = Self-Reported Health; STW training = school-to-work training; PPVT = Peabody Picture Vocabulary Test.

Table 2 describes the raw differences in young adult outcomes. Considering current employment status, the proportion employed is lower for youth with disabilities, and this difference is magnified when beneficiary status is included (83% vs. 75% and 78% vs. 57%). We find a 26-percentage-point difference in the share of youth employed between youth with both a disability and welfare receipt compared with youth without either (57% vs. 83%). Similarly youth with disabilities from beneficiary households are less likely to have ever worked compared with those without disabilities not belonging to beneficiary households (84% vs. 96%). Panel A illustrates another pattern of dual disadvantage, where youth with disabilities from beneficiary households have a higher likelihood of ever being convicted of a crime compared with those with no disability from nonbeneficiary households (19% vs. 11%). Panel B reports differences in earnings on the subsample with wage data. Consistent with the literature, we find that wages are negatively associated with disability status and welfare receipt.

Table 2.

Mean Differences in Adult Outcomes.

Outcome	(1)Nondisability & nonbeneficiary	(2)Nondisability & beneficiary	(3)Disability &nonbeneficiary	(4)Disability & beneficiary
Panel A
Employed	0.83 (0.38)	0.75 (0.44)	0.78 (0.41)	0.57 (0.50)
Out of labor force	0.11 (0.32)	0.15 (0.36)	0.15 (0.35)	0.28 (0.45)
Active military	0.02 (0.14)	0.02 (0.13)	0.01 (0.11)	0.00 (0.06)
Incarcerated	0.00 (0.04)	0.01 (0.08)	0.00 (0.05)	0.02 (0.15)
Ever worked	0.96 (0.21)	0.95 (0.22)	0.94 (0.24)	0.84 (0.37)
Ever convicted	0.11 (0.31)	0.13 (0.34)	0.16 (0.37)	0.19 (0.39)
Observations	6,353	1,001	958	272
Panel B
Wages	22.14 (12.44)	17.56 (10.51)	19.30 (11.63)	13.97 (7.99)
Hours	41.82 (10.49)	40.51 (10.64)	42.85 (12.09)	39.84 (12.90)
Observations	n = 5,322	n = 755	n = 727	n = 148

Note. Means with standard deviations in parentheses. Hours refers to usual hours worked per week.

In addition to asking about the provider of job training, the Add Health survey asked respondents to report the field in which they received vocational education or job training. In total, just under one quarter of the baseline sample participated in in some type of job training. Conditional on receiving job training, Table 3 compares the areas of training for the full sample and also the subsample of individuals trained in a high school, which represents about 7% of the sample with any job training. Columns 1 and 2 report the percentage of respondents who reported training in a particular area either across all training institutions or only for those with training from a high school, respectively. Health-related training was the most common type of training overall (21.5%) and the second most common type for those with high school training (16.8%). Business management training has the third highest frequency among those with STW training.

Table 3.

Prevalence of Participation in Education or Job Training by Field of Training.

Field	(1)Participants in any training program(%)	(2)Participants in H.S. training program(%)
Skilled crafts	7.04	9.09
Machine operator, semiskilled	5.50	8.39
Technician, excluding medical; recording engineer; electronics; nuclear technician	6.40	2.1
Construction and building trades	7.17	9.79
Business management, restaurant management, retail management, leadership	8.45	15.38
Sales/retailing, telemarketing, buyer, insurance underwriter, real estate, travel agent	3.88	5.59
Food service/restaurant worker, excluding management	4.74	6.99
Drafting, surveyor, mechanical drawing, cartographer	2.30	5.59
Secretarial	4.31	9.79
Other office/clerical	3.97	7.69
Computer programmer	3.29	1.4
Computer, N.E.C.	4.44	3.5
Cosmetologist, barber, hair stylist, manicurist	4.57	2.8
Health-related first aid, nurse’s aide, LPN, medical office assistant, pharmacist’s assistant, cardiopulmonary resuscitation, EMT	21.50	16.78
Law enforcement, jailer training, military police, firefighter, police	3.97	0.7
Advertising, photography	0.43	0.7
Engineering	2.43	1.4
Art, music, drama, dance	1.88	2.8
Foreign language	0.34	0
Religion	0.55	0
Other	17.92	29.37
Observations	n = 2,012	n = 143

Note. Participation rates in job training fields conditional upon receiving training (Column 1) and upon receiving training in a high school (Column 2). The other office/clerical category also includes bookkeeping, stock or parts clerk, computer operator, receptionist, bank teller, and keypuncher training programs. N.E.C. = not elsewhere classified; LPN = licensed practical nurse; EMT = emergency medical technician.

Regression Analysis

The results of our analysis of the predictive roles of youth characteristics on young adult outcomes (Equation 1) are reported in Table 4. The first column reports the estimates concerning college attendance and reveals three interesting findings. First, disability status and welfare receipt are both negatively associated with the probability of attaining postsecondary education, and the presence of both intensifies the negative relationship. Compared with the reference group (youth without a disability or beneficiary status), youth without a disability who come from homes receiving welfare are 6.8% less likely to attend college. The same estimate for youth with both disability and beneficiary statuses indicates that these youth are 22% less likely to attend college.

Table 4.

Regression-Adjusted Probabilities of Outcomes for Transition-Age Youth.

Variable	College	Ever worked	Employed	Ever convicted
ND & B	−0.07 (0.02)***	0.01 (0.01)	−0.02 (0.02)	0.02 (0.02)
D & NB	−0.16 (0.02)***	−0.01 (0.01)	−0.01 (0.02)	−0.01 (0.02)
D & B	−0.22 (0.04)***	−0.10 (0.03)***	−0.17 (0.04)***	0.02 (0.03)
STW	−0.13 (0.07)**	−0.04 (0.04)	0.02 (0.05)	−0.10 (0.02)***
STW × ND & B	−0.09 (0.12)	0.03 (0.06)	0.07 (0.09)	0.14 (0.09)
STW × D & NB	0.19 (0.15)	0.08 (0.04)**	0.20 (0.08)**	0.19 (0.12)
STW × D & B	0.13 (0.15)	0.20 (0.05)***	0.30 (0.12)**	−0.10 (0.06)*
Std. PPVT score	0.08 (0.01)***	0.01 (0.00)**	−0.01 (0.01)	0.00 (0.01)
Mental health	−0.02 (0.02)	−0.01 (0.01)	−0.08 (0.02)***	0.04 (0.01)
School FE	Yes	Yes	Yes	Yes
Observations	n = 8,584	n = 8,584	n = 8,584	n = 8,584

Note. Regression estimates with standard errors in parentheses. ND = nondisability; B = beneficiary; D = disability; NB = nonbeneficiary; STW = school-to-work; Std. PPVT score = standardized Peabody Picture Vocabulary Test score; FE = fixed effects.

p < .10. **p < .05. ***p < .01.

As indicated in Equation 1, these estimates control for a wide range of youth, maternal, and school characteristics. The control measures lead to the second interesting finding: that the inclusion of the PPVT score attenuates the predictive role of disability and welfare receipt. Karpur et al. (2014) found estimates roughly 4 to 6 percentage points larger than our point estimates. This observation indicates that disability and welfare receipt play a significant role in predicting college attendance, though not as large as previously estimated. Finally, Column 1 shows that STW programs have a negative relationship with attending postsecondary school; youth participating in STW programs are 13% less likely to go on to college. Of course, this finding is influenced by the selection of students choosing to take STW programs and should not be interpreted as the causal effect of taking a STW program.

We next examine the factors that are associated with a young adult’s working history. In column 2 of Table 4, we see that youth with disabilities from a welfare beneficiary home who do not participate in a STW program have a 10% lower probability of ever having worked than the reference group. On the contrary, youth with disabilities who participate in STW, both with and without beneficiary status, have a higher likelihood of ever working compared with the reference group, 20% and 8%, respectively. These patterns carry over when considering the role of STW programs on the probability of being currently employed. Participation in a STW program is most positively associated with current employment among youth with disabilities. This finding reveals that while students with disabilities from welfare beneficiary households are less likely to have ever worked or currently be working, those that participated in STW programs have a significantly higher probability of working.

A final compelling result from our first model is the role of STW on conviction outcomes. We find weak evidence of a lower likelihood of having a conviction among youth with both disability and beneficiary statuses and STW experience. This estimate suggests that youth who chose to go through a STW program had more positive adult outcomes in terms of both employment and criminal records. While these models do not control for selection issues, the estimates indicate that STW programs may be a useful mechanism to help improve the lives of youth with disabilities from welfare recipient homes.

The results of our determinants of earnings model are reported in Table 5. Relative to the reference group, welfare receipt and the combination of welfare receipt and disability status are associated with lower young adult wages, 6% and 23%, respectively (using our preferred model with school fixed effects, occupation controls, and only full-time workers). This model also reveals a large negative predictive role (−41%) of STW programs for youth with both disabilities and beneficiary statuses. Integrating our findings here with the ones from Table 4 suggests that while STW programs are associated with higher likelihood of employment, they are also linked to lower wages given a person’s occupation and other control factors.

Table 5.

Estimated Wages From Determinants of Earnings Model.

Variable	Log wages	Log wages	Log wages	Log wages
NB & B	−0.09 (0.03)***	−0.05 (0.03)*	−0.09 (0.03)***	−0.06 (0.03)*
D & NB	−0.02 (0.03)	−0.02 (0.03)	−0.04 (0.03)	−0.04 (0.03)
D & B	−0.18 (0.06)***	−0.15 (0.06)***	−0.27 (0.06)***	−0.23 (0.06)***
STW	0.01 (0.08)	0.02 (0.06)	0.03 (0.08)	0.06 (0.06)
STW × (ND & B)	−0.06 (0.14)	−0.10 (0.12)	−0.13 (0.15)	−0.13 (0.13)
STW × (D & NB)	−0.00 (0.13)	0.04 (0.11)	−0.01 (0.13)	0.03 (0.11)
STW × (D & B)	−0.40 (0.21)*	−0.47 (0.16)***	−0.34 (0.20)*	−0.41 (0.17)**
Std. PPVT score	−0.01 (0.01)	−0.01 (0.01)	−0.01 (0.01)	−0.01 (0.01)
Mental health	−0.07 (0.02)***	−0.06 (0.02)***	−0.07 (0.02)***	−0.06 (0.02)***
Schooling	0.06 (0.02)***	0.06 (0.02)***	0.05 (0.02)***	0.05 (0.02)***
Experience	0.03 (0.02)	0.03 (0.02)	0.01 (0.02)	0.01 (0.02)
Experience square	−0.00 (0.00)**	−0.00 (0.00)*	−0.00 (0.00)	−0.00 (0.00)
School FE	No	Yes	No	Yes
Occupational FE	No	Yes	No	Yes
Part-time	Yes	Yes	No	No
Observations	n = 6,952	n = 6,952	n = 6,069	n = 6,069

Note. Regression estimates with standard errors in parentheses. NB = nonbeneficiary; B = beneficiary; D = disability; STW = school-to-work; ND = nondisability; Std. PPVT Score = standardized Peabody Picture Vocabulary Test score.

p < .10. **p < .05. ***p < .01.

Discussion

We conducted this research to understand whether participation in STW programs helps mitigate the adverse effect of growing up in a household receiving welfare benefits on the likelihood of positive transition-to-adulthood outcomes for youth with disabilities. It is apparent from the data that participation in STW predicts both an improved chance of employment and also a reduced likelihood of criminal justice involvement among these youth. This analysis adds to the existing literature by using a nationally representative sample of youth both with and without disabilities and by incorporating independent variables that might otherwise confound the estimated relationship between STW programs and adult outcomes, that is, the inclusion of a measure of aptitude and occupational fixed effects. Our findings support efforts of many welfare agencies, particularly the SSA, in promoting transition-to-adulthood programs with an emphasis on vocational and career development for these young people.

Prior research, including our own, documented the detrimental relationship of parental receipt of welfare benefits on adult productive engagement (i.e., postsecondary education participation) for youth with disabilities. Child developmental theories emphasize the interactive roles of social causation and social selection impacting young adult development. Schofield et al. (2011) empirically studied social causations and demonstrated that living in poverty or low SES impacts parental stress responses, resulting in a disruption of parental engagement and increased internalizing and externalizing problems for the child. Furthermore, they found that material deprivation reduces access to educational resources and that living in areas with poorer educational systems contributes to the poor outcomes of these children. Schofield et al. (2011) also demonstrated the role of social selection: High-SES parents were more emotionally invested in their children’s education, and these youth had better adaptive functioning and outcomes post school. It is likely that both social causation and social selection are playing crucial roles in influencing outcomes for youth with disabilities who live in households that receive public benefits. Specifically, the ability of PPVT to account for variation in the dependent variables of interest, as shown in this study, points to the presence of social selection process.

Prior research has indicated a predictive role of participation in work during high school on postsecondary employment outcomes. Karpur et al. (2014) illustrated that participating in paid work mediated about 25% of the likelihood of successful employment outcomes for youth participating in a model demonstration program in New York state. From a human capital theory perspective, paid work during high school leads to the development of skills as well as a way to communicate those skills to future employers. With high school diplomas becoming less differentiated over time (Kerckhoff, 2001; Levels, van der Velden, & Di Stasio, 2014), attaining a diploma itself has become increasingly less helpful in signaling attainment of specific skills valued by employers in the labor market. However, participation in STW training helps young people to build and articulate specific skills necessary to finding jobs.

Furthermore, STW programs contribute to higher school engagement and school completion rates among youth. Importantly, such programs have the ability to reduce criminal behaviors by reducing externalizing behaviors, especially among youth with serious emotional disturbances (Haber, Karpur, Deschênes, & Clark, 2008). Our analysis supports this positive aspect of STW program participation. Gottfredson, Gerstenblith, Soulé, Womer, and Lu (2004) argued that STW-type programs reduced delinquency among youth not by reducing unsupervised time or increasing time spent engaging in productive activities but through the reduction in drug use and improved peer-social relationships. That research emphasized the role of improved social capital as a mechanism for positive youth engagement in work.

It is important to note that although youth exposed to STW in this study were more likely to be currently employed, they had 41% lower earnings compared with their peers. Literature in VR continues to point to an increasing trend for people with disabilities to be employed in “low-rung” jobs and highlights precarious employment conditions (Kaye, 2010). Despite the outside evidence supporting the lower-wage jobs explanation, it is less likely to fit the results found here due to the inclusion of occupational fixed effects. Because the data only support an analysis using 23 broad occupational groupings, it remains possible that within a specific grouping, individuals with disabilities and STW experience are in “low-rung” jobs.

The explanation that these individuals are doing the same job for less pay seems to better fit these results. In many instances, means-tested welfare policies create disincentives for work participation among beneficiaries (Stapleton, O’Day, Livermore, & Imparato, 2006). It is important to note that for many of the SSA beneficiaries, losing access to benefits is not only accompanied by termination of monthly cash payments but also by termination of access to health care, as Medicaid eligibility is tied to SSA eligibility. Thus, it is possible that STW program providers could be encouraging these youth to take low-wage work to ensure they do not exceed the benefit threshold. Further research is needed to understand the reasons for lower wages among STW program participants. More specifically, more work is necessary on theoretical frameworks that explain the benefits of STW programs given the labor market and its ability to absorb talented youth with disabilities.

Initiatives such as Promoting the Readiness of Minors in Supplemental Security Income (PROMISE) will help expand the field’s understanding of factors that promote employment and overall economic well-being among youth with disabilities who are recipients of SSI. The PROMISE programs, currently being implemented across 11 sites nationally, utilize a randomized clinical trial design to test the impact of evidence-based innovative practices on transition to early adulthood.

Study Limitations

This is a secondary analysis of a nationally existing dataset and is certainly limited by the available information. Specifically, information on a youth’s direct receipt of public welfare would have highlighted the direct impact of participation in welfare programs on transition outcomes. This information would allow the analysis to focus on a group of more highly at-risk youth, those receiving SSI. Furthermore, this research is not in a position to test the issues of social causation and social selection as addressed by Schofield et al. (2011).

Another limitation is that the data were not collected specifically to study the impact of STW programs, so the sample size of individuals with disabilities that have STW experiences is moderate. This sample then decreases further when one additionally wants to explore the interactive role of benefit receipt. In addition, greater details regarding the STW programs would allow for an analysis of the heterogeneous role of STW programs on this population. Carter, Trainor, Cakiroglu, Swedeen, and Owens (2010) identified several school-level and practitioner-level factors that impact access to STW training for youth with disabilities and found a negative relationship between severity of disability and referrals for STW opportunities as well as evidence of insufficient professional development to promote STW programs among school staff. This could also impact the patterns of referrals to STW training for students across the schools. Their research also alluded to wide qualitative variations in STW activities. While their study was limited to 26 districts in one state, it is reasonable to extrapolate that such factors could also be influencing the outcomes in our analysis. Furthermore, several factors impact the choices of students with disabilities who participate in STW programs, and those factors, such as gradients in functioning, may impact adult outcomes. Studies that utilize randomized trial designs, such as PROMISE, could help eliminate such biases and study the causal impact of STW programs.

Conclusion

Youth with disabilities from welfare beneficiary households face a double jeopardy as they attempt to transition from adolescence to adulthood. A wide range of agencies and programs, like those supported by the SSA, have sought to reduce the barriers to successful employment and community engagement that this population faces. While evidence of effective policies and services is growing, these youth continue to have poorer labor market outcomes compared with their peers. This analysis presents evidence of a mitigating impact of STW programs on employment outcomes for youth with disabilities belonging to households that receive welfare benefits. This research provides further justification for the efforts made by programs, such as the PROMISE initiative, in supporting some aspects of the theory of change for adolescence-to-adulthood transition programs. While the analysis presents a limited picture, it reinforces the need for conducting systematic outreach for these youth to engage in STW programs and improve their employability while in high school settings.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: New York State PROMISE is funded by a grant from the United States Department of Education, H418P130011. The contents of this article were developed under a grant from the U.S. Department of Education. However, those contents do not necessarily represent the policy of the U.S. Department of Education, and you should not assume endorsement by the Federal Government.

ORCID iD

Hassan Enayati

References

Berry

H. G.

Caplan

L. J.

(2010). Employment and earnings growth among transition-age supplemental security income program participants. Journal of Disability Policy Studies, 21, 152–159.

Blau

D. M.

(1999). The effect of income on child development. The Review of Economic and Statistics, 81, 261–276.

Brownstein

Kalsbeek

Tabor

Entzel

Daza

Harris

K. M.

(2010). Non-response in wave IV of the National Longitudinal Study of Adolescent Health. In Add Health Users Conference. Retrieved from http://www.cpc.unc.edu/projects/addhealth/data/guides/W4_nonresponse.pdf

Card

(1999). The causal effect of education on earnings. In Ashenfelter

Card

(Eds.), Handbook of labor economics (Vol. 3, Part A, pp. 1801–1863). Amsterdam, The Netherlands: Elsevier.

Carter

E. W.

Trainor

A. A.

Cakiroglu

Swedeen

Owens

L. A.

(2010). Availability of and access to career development activities for transition-age youth with disabilities. Career Development for Exceptional Individuals, 33, 13–24.

Chantala

Kalsbeek

W. D.

Andraca

(2005). Non-response in Wave III of the Add Health Study. Health User Guides, Carolina Population Center, University of North Carolina at Chapel Hill. Retrieved from www.cpc.unc.edu/projects/addhealth/documentation/guides/W3nonres.pdf

Chetty

Hendren

Lin

Majerovitz

Scuderi

(2016). Childhood environment and gender gaps in adulthood. American Economic Review, 106, 282–288

Davies

P. S.

Rupp

Wittenburg

(2009). A life-cycle perspective on the transition to adulthood among children receiving Supplemental Security Income payments. Journal of Vocational Rehabilitation, 30, 133–151.

Duggan

Kearney

M. S.

(2005). The impact of child SSI enrollment on household outcomes: Evidence from the survey of income and program participation. (NBER working paper 11568). Retrieved from http://www.nber.org/papers/w11568

10.

Ellen

I. G.

Turner

M. A.

(1997). Does neighborhood matter? Assessing recent evidence. Housing Policy Debate, 8, 833–866.

11.

Erickson

Lee

von Schrader

(2014). 2013 disability status report: United States. Ithaca, NY: Cornell University.

12.

Fraker

Mamun

Honeycutt

Thompkins

Valentine

E. J.

(2014). Final report on the youth transition demonstration evaluation. Washington, DC. Retrieved from http://www.mdrc.org/sites/default/files/Final_Repot_on_the_Youth_Transition_Demonstration_Evaluation.pdf

13.

Ghosh

Parish

S. L.

(2015). Deprivation among U.S. children with disabilities who receive supplemental security income. Journal of Disability Policy Studies, 26, 173–183.

14.

Gottfredson

D. C.

Gerstenblith

S. A.

Soulé

D. A.

Womer

S. C.

(2004). Do after school programs reduce delinquency? Prevention Science, 5, 253–266.

15.

Haber

M. G.

Karpur

Deschênes

Clark

H. B.

(2008). Predicting improvement of transitioning young people in the partnerships for youth transition initiative: Findings from a multisite demonstration. The Journal of Behavioral Health Services & Research, 35, 488–513.

16.

Hemmeter

Donovan

Cobb

Asbury

(2015). Long term earnings and disability program participation outcomes of the Bridges transition program. Journal of Vocational Rehabilitation, 42(1), 1–15.

17.

Hemmeter

Kauff

Wittenburg

(2008). Changing circumstances: Experiences of child SSI recipients before and after their age-18 redetermination for adult benefits. Ithaca, NY: Employment and Disability Institute, Cornell University.

18.

Hoffman

Hemmeter

Bailey

M. S.

(2016, August). The relationship between youth services and adult outcomes among former child SSI recipients. Paper presented at the 4th annual meeting of the Disability Research Consortium, Washington, DC. Retrieved from http://www.nber.org/aging/drc/20160803drcmeeting/Panel%203.3%20Hoffman.pdf

19.

Houtenville

A. J.

Brucker

D. L.

(2014). Participation in safety-net programs and the utilization of employment services among working-age persons with disabilities. Journal of Disability Policy Studies, 25, 91–105.

20.

Karpur

Nazarov

Brewer

D. R.

Bruyère

S. M.

(2014). Impact of parental welfare participation: Transition to postsecondary education for youth with and without disabilities. Career Development and Transition for Exceptional Individuals, 37, 1–11.

21.

Kaye

H. S.

(2010). The impact of the 2007-09 recession on workers with disabilities. Monthly Labor Review, 133, 19–30.

22.

Kerckhoff

A. C.

(2001). Education and social stratification processes in comparative perspective. Sociology of Education, 74(extra issue), 3–18.

23.

Levels

van der Velden

R. K. W.

Di Stasio

(2014). From school to fitting work: How education-to-job matching of European school leavers is related to educational system characteristics. Acta Sociologica, 57, 341–361.

24.

Leventhal

Brooks-Gunn

(2000). The neighborhoods they live in: The effects of neighborhood residence on child and adolescent outcomes. Psychological Bulletin, 126, 309–337.

25.

Loprest

P. J.

Wittenburg

D. C.

(2007). Posttransition experiences of former child SSI recipients. Social Service Review, 81, 583–608.

26.

Mann

D. R.

Wittenburg

D. C.

(2015). Starting behind: Wage and employment differentials between young adults with and without disabilities. Journal of Disability Policy Studies, 26, 89–99.

27.

Mincer

(1974). Schooling, experience, and earnings. New York, NY: National Bureau of Economic Research.

28.

Mitnik

O. A.

(2007). Intergenerational transmission of welfare dependency: The effects of length of exposure (Working Paper No. 0715). Miami, FL: Department of Economics, University of Miami.

29.

National Collaborative on Workforce and Disability for Youth. (2012). The guideposts for success: A framework for families preparing youth for adulthood. Retrieved from http://www.ncwd-youth.info/sites/default/files/infobrief_36.pdf

30.

Park

Turnbull

A. P.

Turnbull

H. R.

III . (2002). Impacts of poverty on quality of life in families of children with disabilities. Council for Exceptional Children, 68, 151–170.

31.

Schofield

T. J.

Martin

M. J.

Conger

K. J.

Neppl

T. M.

Donnellan

M. B.

Conger

R. D.

(2011). Intergenerational transmission of adaptive functioning: A test of the interactionist model of SES and human development. Child Development, 82, 33–47.

32.

Solon

(1999). Intergenerational mobility in the labor market. In Handbook of labor economics (Vol. 3, Part A, pp. 1761–1800). Amsterdam, The Netherlands: Elsevier.

33.

Stapleton

D. C.

Burkhauser

R. V.

She

Weathers

R. R.

Livermore

G. A.

(2009). Income Security for workers: A stressed support system in need of innovation. Journal of Disability Policy Studies, 19, 204–220.

34.

Stapleton

D. C.

O’Day

B. L.

Livermore

G. A.

Imparato

A. J.

(2006). Dismantling the poverty trap: Disability policy for the twenty-first century. Milbank Quarterly, 84, 701–732.

35.

Wagner

M. M.

Newman

L. A.

Javitz

H. S.

(2014). The influence of family socioeconomic status on the post-high school outcomes of youth with disabilities. Career Development and Transition for Exceptional Individuals, 37, 5–17.

36.

Wagner

M. M.

Newman

L. A.

Javitz

H. S.

(2016). The benefits of high school career and technical education (CTE) for youth with learning disabilities. Journal of Learning Disabilities, 49, 658–670.