Personal and Contextual Factors Associated With Successful Vocational Rehabilitation and Employment Outcomes

Abstract

Some individuals with disabilities have relatively better labor outcomes than others. If vocational rehabilitation (VR) counselors can better understand what factors are correlated with positive outcomes, they could better identify what clients are at risk for poor outcomes and then intervene early to improve chances for success. Unfortunately, until recently VR survey and administrative data have had insufficient information to investigate these questions. We use a survey of 932 Ohio VR clients combined with administrative earnings data to examine the barriers and facilitators associated with employment and earnings outcomes. The survey data are from the 2014 Survey of Disability and Employment, a survey of VR applicants in three states. We compare VR case status and earnings through 2016, by demographics, functional limitation status, employment history and attitudes, and reasons stated for not working in the survey. We find that about 25% of survey respondents exited the program employed or had at least one quarter of average monthly earnings above the Social Security Administration’s substantial gainful activity amount. Pain, depression or anxiety, longer duration of job separation, personal and familial attitudes about work, and not working because of personal and health reasons were associated with poorer outcomes among survey respondents.

Keywords

vocational rehabilitation health disability employment

Introduction

The vocational rehabilitation (VR) program plays an important role in helping many individuals with disabilities achieve their employment goals. A federal-state partnership, VR is administered at the state level with the majority of its funding provided by the federal government. In 2013, state VR agencies (SVRAs) provided services for about one million individuals with disabilities who voluntarily applied for services. In the same year, approximately 183,000 individuals with disabilities were employed when they left the VR program (Rehabilitation Services Administration [RSA], 2016).

Previous studies have documented differences in outcomes across characteristics recorded in VR administrative data, such as demographic or impairment information. However, administrative data capture an incomplete picture of the characteristics that can help explain why some VR clients fare better than others. Though VR results in successful employment outcomes for many individuals, roughly two thirds of applicants exit the VR program before receiving services or without employment even after receiving services (RSA, 2014). In addition, though some VR clients who exit the program employed enjoy long-term labor market success, a substantial share do not remain employed for the long run (Mann, Honeycutt, Stegman Bailey, & O’Neill, 2017). If we can understand what applicant characteristics are most correlated with certain program outcomes, then there is an opportunity to tailor VR services to better address especially challenging barriers to employment.

Background

The social model of disability provides a useful framework for understanding employment and program participation among individuals with disabilities (Shakespeare, 2013). Although individuals may have functional impairments due to medical conditions, the social model posits that disability results from the interaction of those impairments with personal characteristics, societal attitudes, and the physical and social environment. The social model of disability groups the determinants of employment outcomes into three domains: (a) an individual’s health conditions, (2) his or her individual characteristics, and (c) features of his or her environment which can include but is not limited to family, local labor market, and policy environments. These domains are also embodied in the World Health Organization’s (2001) International Classification of Functioning, Disability and Health.

Existing research documents differences in outcomes across readily available measures of the first two domains but not the third. Individuals with sensory or communicative disabilities typically have better employment outcomes, followed by individuals with physical disabilities, and then by individuals with mental impairments (Rosenthal, Chan, Wong, Kundu, & Dutta, 2006). There are also substantial differences in VR outcomes by demographic groups (Martin, 2010). For example, Whites are more likely to be deemed eligible for VR services (Wilson, 2002; Wilson, Alston, Harley, & Mitchell, 2002) and African Americans are less likely to be placed in competitive employment (Capella, 2002; Feist-Price, 1995; Rosenthal et al., 2006; Rosenthal, Ferrin, & Wilson, 2005; Wilson, 1999). Higher levels of educational attainment are also associated with higher employment rates and earnings among persons with disabilities (Beveridge & Fabian, 2007; Gilmore & Bose, 2005; Loprest & Maag, 2007; Sevak, Houtenville, Brucker, & O’Neill, 2015). In addition, recent research has found that differences across demographic groups vary by type of impairment (O’Neill, Kaczetow, Pfaller, & Verkuilen, 2017).

Recent research has provided information about long-term earnings and benefit receipt outcomes of VR clients by linking VR data with Social Security Administration (SSA) data (Mann et al., 2017; Stapleton & Martin, 2017). Mann et al. (2017) found that though clients who closed with employment tended to have the best long-term employment and benefit receipt outcomes, those who did not receive services had better long-term employment and benefit receipt outcomes than did those who received services but closed without employment.

A richer set of personal and contextual factors may also be important for employment. Factors such as motivation, employment history, accommodations, and family support may play a role in whether an individual is employed or successfully engages with the VR system, but these factors have not been successfully examined quantitatively. One exception is Dutta et al. (2017), a study of 277 individuals with disabilities across eight states which provides some evidence how motivation or engagement is an important factor in the VR process (Dutta et al., 2017). Emerging evidence suggests that receipt of workplace accommodations is associated with better employment outcomes (Anand & Sevak, 2017; Chow, 2012; Cook, Burke-Miller, & Gray, 2015; Hill, Maestas, & Mullen, 2016; Kristman, Shaw, Boot, Delclos, & Sullivan, 2016; Sevak & Khan, 2017).

Research on the role these contextual and personal factors play is constrained by limitations in large data sets used in this field, including Rehabilitation Services Administration (RSA) administrative data, the American Community Survey (ACS), and SSA administrative data. Though VR counselors gather some of this information when they meet with new VR applicants, rather limited quantitative evidence is available.

This study contributes to the literature on VR and employment outcomes by making use of new data available from the Survey of Disability and Employment (SDE) linked with administrative data. We leverage this combined data to learn about how these contextual and personal factors are related to VR case status and earnings outcomes. Our analysis provides insight into what individual characteristics are correlated with poor VR outcomes, which should help VR counselors anticipate which clients may need extra assistance to achieve their employment goals.

Method

Data

Our analysis uses a sample of 932 working-age (25–64 years old) individuals with disabilities who applied for VR services in Ohio between August and December 2014.¹ These individuals were interviewed for the SDE between November 2014 and May 2015, most within several months after they applied for services. The goal of the survey was to capture information on the applicants’ backgrounds and needs at the time they applied for services. As a result, the SDE sample included VR applicants regardless of whether the applicants subsequently received VR services.

The SDE offers many advantages for this analysis compared with other data sets commonly used to examine employment outcomes for VR clients or the broader population of people with disabilities, such as the RSA-911, the ACS, or the National Beneficiary Survey (NBS). The SDE asks detailed questions about workplace characteristics and employment barriers that are not available in administrative data such as the RSA-911 or in most large national surveys like the ACS. Relative to the NBS—a survey of Social Security disability support program beneficiaries, who are typically unlikely to have recent employment experience—the SDE sample only includes people with disabilities who have signaled an interest in employment by applying for VR. In addition, most VR applicants are not Supplemental Security Income or Social Security Disability Insurance beneficiaries (Mann et al., 2017).

Eckstein, Sevak, and Wright (2017) provide a detailed description of the SDE design, data collection, measures, and response rates. The survey instrument was reviewed and approved by the institutional review board at the Kessler Foundation. Like most surveys, participation in the SDE was voluntary, with respondents giving verbal consent for the survey and for combining the survey with State administrative data during the telephone interview; as a result, estimates from the survey could be biased if responding sample members differ from the state’s VR population as a whole. The survey’s response rate for Ohio was 60%, and a nonresponse bias analysis illustrated that the weighted final sample did not differ from the VR applicant population along several key covariates. We adjusted all analyses using sampling weights that account for differential nonresponse by characteristics such as age, gender, race, and type of disability.

Several studies using the data from all three states have demonstrated the usefulness of the cross-sectional data in the SDE. Eckstein et al. (2017) describe the many barriers to employment experienced by respondents, and Sevak and Khan (2017) compare employment barriers and facilitators experienced by VR applicants with psychiatric versus physical disabilities, finding that individuals with psychiatric disabilities face more barriers, on average. Relative to those who are not employed when they apply for services, Brucker, Boticello, O’Neill, and Kutlik (2017) show that employed VR applicants have relatively more social capital and Anand and Sevak (2017) find that applicants who received accommodations at their most recent job are more likely to be employed at application than those who did not receive accommodations.

To better understand the extent to which VR and future employment outcomes differ across individuals who perceive different employment barriers and facilitators when they apply to VR, we merge data from the SDE with administrative data from Ohio’s VR agency. These data include VR case status for both open and closed cases, and earnings records from the Ohio Department of Labor. VR case status is reported as of December 2016, roughly 2 years after application, and earnings are reported quarterly for the six quarters from January 2015 to June 2016. The earnings records are available for all VR applicants in the SDE sample, regardless of VR closure status. This is particularly useful because when we conducted this study, VR administrative data did not contain case status information on applicants whose cases are still open. However, we do not have earnings data from jobs outside of Ohio; because some individuals will cross states to go to work, our estimates of earnings should be interpreted as a slight underestimate. The data were merged in a secure environment using the VR agency’s unique case ID. The case ID was removed from the merged data file before it was made available to researchers.

Data Analysis

We use these linked data to examine the relationships between VR applicant characteristics at the time of VR application and VR and employment outcomes within 18 to 24 months of application. Four sets of baseline characteristics were examined: demographic characteristics; functional limitations; employment search, history, and expectations; and reasons for not working (among those unemployed at application). The results section contains a discussion of the detailed values for these characteristics.

Using descriptive statistics, we examine the relationship between this information collected close to VR application to subsequent VR case status and earnings outcomes. VR closures were grouped into four categories: closed with employment, closed without employment but after receiving services or signing an Individualized Plan for Employment (IPE), closed before receiving services or signing an IPE, and not yet closed. For each baseline characteristic, we report the percentage of respondents who had a particular case status roughly 2 years after application.

We use descriptive statistics, rather than a multivariable regression framework for a number of reasons. First, we want to examine gross differences across characteristics, rather than net differences which control for other factors. Second, responses to many of the subjective questions such as reasons for not working are collinear and make estimation of a multivariable regression with all of the variables included concurrently problematic. Third, use a multivariate regression framework is often motivated by an attempt to get causal estimates; given the subjective nature of much of the baseline information of interest, we consider the findings to be descriptive.

Even in such univariate descriptive analysis, tests of significance can inform whether observed differences in means across groups are likely due to random variation or actual differences. For each baseline characteristic, we test whether there are significant differences in the closure status, using chi-square tests. We confirmed that the standard statistical assumptions for these tests were met.

We also examine two earnings outcomes using the earnings records from January 2015 to June 2016. The first earnings variable is average monthly earnings, which is calculated (in nominal dollars) across all 18 months, including months in which an individual did not have any earnings. The second earnings variable measures the percentage of individuals who had at least one quarter in which their average monthly earnings were above the SSA 2015 substantial gainful activity (SGA) threshold. The SGA amount is a monthly earnings threshold used by SSA for the Supplemental Security Income and Social Security Disability Insurance programs to assess whether an applicant or beneficiary is earning enough to be economically self-sufficient. The 2015 SGA amount was US$1,090 for nonblind beneficiaries. As with closure outcomes, we identify statistically significant differences across groups. For the SGA-level earnings outcome we use chi-square tests. For monthly earnings, we use the chi-square test statistic from the Kruskal–Wallis test. The Kruskal–Wallis test is a conservative, nonparametric test that can be used to test for differences in continuous outcomes in two or more samples.

Results

In this section, we describe our findings as illustrated in eight tables. In each table, VR or earnings outcomes are shown in columns and information about characteristics at application in rows. Percentages in all tables are calculated by row, not column. The mean percentages and dollars reported in the tables are survey weighted estimates.

Just over half of respondents exited the VR program before having a signed IPE. Roughly 25% of the 932 respondents in the analysis sample closed with employment (see Table 1). About 29% closed without employment but after IPE, and 42% of respondents exited VR before getting a signed IPE. Only about 5% of survey respondents still had an open case 2 years after applying for VR services. Across all respondents, including those not working, respondents earned an average of US$423 a month during the first 18 months after application and 26% had at least one quarter with earnings at or above the quarterly SGA amount ( see Table 2). When helpful in the text below, we compare these overall outcomes to outcomes by characteristic at application.

Table 1.

Differences in Future Case Status by Demographic Characteristics at Application.

	n	VR case status (%)				χ² test statistic
Characteristic	n	Closed employed	Closed after IPE but without employment	Closed before IPE	Not yet closed	χ² test statistic
Sample size	932	225	267	390	50
Total	932	24.6	28.6	41.5	5.2
Gender						6.02
Male	467	27.9	27.9	38.5	5.7
Female	465	21.3	29.3	44.6	4.8
Marital status						4.29
Married or partnered	226	28.3	25.7	41.2	4.7
Widowed/divorced/separated	266	20.9	29.9	42.6	6.6
Never married and other	440	24.8	29.3	41.1	4.8
Highest level of education						34.58*
Less than high school	144	14.7	32.4	48.5	4.4
High school graduate	400	26.3	28.5	41.3	3.9
Some college but no degree	164	16.4	27.9	48.8	6.8
Associate’s degree	133	28.7	31.8	32.1	7.4
Bachelor’s degree or higher	91	40.3	19.7	32.8	7.2
Race and ethnicity						33.39*
White alone	555	30.0	28.6	35.3	6.0
Black alone	287	13.6	28.9	53.4	4.0
Other	90	22.2	27.3	46.3	4.2
Age						4.35
25–34	214	26.3	31.6	39.0	3.1
35–44	271	23.2	30.1	39.9	6.8
45–54	314	25.8	25.1	44.2	4.9
55–64	133	20.8	27.1	44.3	7.7

Source. Weighted estimates from the Survey of Disability and Employment and Ohio administrative data.

Note. VR = vocational rehabilitation; IPE = Individualized Plan for Employment.

Statistically significant differences at the 5% level, based on chi-square test statistic.

Table 2.

Differences in 18-Month Earnings by Demographic Characteristics at Application.

	N	18-month earnings
Characteristic	N	Average monthly earnings (US$)	χ² test statistic	Average monthly earnings above SGA in one or more quarters (%)	χ² test statistic
Sample size	932	932		932
Total	932	423		26.1
Gender			1.37		1.74
Male	467	464		28.1
Female	465	381		24.1
Marital Status			1.17		8.61*
Married or partnered	226	569		30.5
Widowed/divorced/ separated	266	422		29.0
Never married and other	440	347		22.2
Highest level of education			31.75*		34.83*
Less than high school	144	164		12.5
High school graduate	400	388		25.6
Some college but no degree	164	335		21.8
Associate’s degree	133	544		36.7
Bachelor’s degree or higher	91	960		42.0
Race and ethnicity			13.53*		12.30
White alone	555	487		29.7
Black alone	287	302		18.5
Other	90	362		25.4
Age			4.364		2.53
25–34	214	370		28.4
35–44	271	395		23.3
45–54	314	493		28.5
55–64	133	427		21.6

Source. Weighted estimates from the Survey of Disability and Employment and Ohio administrative data.

Note. SGA = substantial gainful activity.

Statistically significant differences at the 5% level, based on chi-square test statistic.

Differences by Demographic Characteristics

Rates of some closure outcomes seem to differ by gender (see Table 1) though a chi-square test suggests that overall there were not significant differences in closure status by gender (χ² = 6.02). A higher share of men exited with employment than did women (28% vs. 21%) while women were more likely to exit without a signed IPE (45% vs. 39%). The percentage of applicants exiting without employment or still having an open case 2 years after application was similar across genders. While average earnings were higher for men than women (see Table 2), the difference was not statistically significant (US$464 vs. US$381; χ² = 1.37). Similarly, a slightly, but not statistically significant higher share of men had at least one quarter with earnings in excess of the quarterly SGA amount (28.1% vs. 24.1%;χ² = 1.74).

Marital status was also associated with differences in some but not all outcomes. While married and cohabiting applicants had the highest rate of exiting with employment (28.3%), the differences in closure outcomes by marital status groups were not statistically significant (χ² = 4.29). Married and cohabiting applicants have the highest earnings (US$569) but earnings did not differ significantly by marital status (χ² = 1.17). There were statistically significant differences in the share of individuals with SGA level earnings by marital status (χ² = 8.61) with the highest rate, 31%, among married or cohabiting individuals. Previously married individuals—that is, individuals who were widowed, divorced, or separated—experienced the lowest rates of successful closure (20.9%), but those who were never married had the lowest monthly earnings (US$347), and only 22% of them had at least one quarter of SGA earnings.

Unsurprisingly, closure status differed significantly across educational attainment groups (χ² = 34.58). Survey respondents with relatively higher levels of education typically had greater relative success exiting the VR program employed. Forty percent of those with a bachelor’s degree or higher exited the VR program with employment, whereas just 15% of those without a high school diploma had the same closure status. Interestingly, the closure status of those who had some college education but no degree looked most like those without a high school diploma, with just 16% exiting the program employed and 49% exiting without an IPE. Mean monthly earnings varied substantially by education (χ² = 31.75), ranging from US$164 for those without a high school diploma to US$960 for those with a bachelor’s degree or higher. Similarly, those with higher levels of education were more likely to have at least a quarter of earnings at or above SGA (42% of college graduates vs. 13% of those who had not completed high school; χ² = 34.83).

Race was also associated with meaningful and statistically significant variation in closure outcomes (χ² = 33.39). The employment at closure rate was 30% among Whites, 14% among Blacks, and 22% among individuals who identified with other or multiple races. Blacks had the highest rate, 53.4%, of closure without a signed IPE. Similarly, we found statistically significant differences by race in monthly earnings (χ² = 13.53). Average monthly earnings ranged from US$487 for Whites, US$362 for individuals identifying with other or multiple races and US$302 for Blacks. The SGA earnings rate was 30% among Whites, 25% among individuals who identified with other or multiple rates and 19% among Blacks though the differences in these rates were not statistically significant (χ² = 12.30).

While there were case status and earnings differences among the age groups, the differences were not consistent across age category and outcome and none were statistically significant. Those ages 55 to 64 had the lowest rates of employment at closure. While the pattern of average monthly earnings across age groups did not reveal statistically significant differences, it was consistent with the age-earnings profile, with earnings lowest among the youngest and rising with age until declining among those approaching retirement age.

Differences by Functional Limitations

SDE respondents were asked whether they had specific functional limitations, and they responded affirmatively or negatively to each separately. Across almost all functional assessment questions, having a particular functional limitation was associated with lower rates of employment at exit and/or higher rates of exiting before getting a signed IPE (see Table 3). Except for those with sensory impairments (hearing, vision, or speech), applicants who identified as having a particular functional limitation were less likely to exit the VR program with employment than those who did not report the limitation. Lowest rates of employment at closure were exhibited by those with difficultly climbing stairs (17.6%; χ² = 18.48), chronic pain (18.7%; χ² = 23.43), difficulty dressing or bathing (16.8%; χ² = 6.23), or very frequent depression or anxiety (17.8%; χ² = 17.71). Instead of exiting without employment, most of the applicants from these categories ended up exiting VR before getting a signed IPE.

Table 3.

Differences in Future Case Status by Presence of Functional Limitations at Application.

	n	VR case status (%)				χ² test statistic
Functional limitation	n	Closed employed	Closed after IPE but without employment	Closed before IPE	Not yet closed	χ² test statistic
All respondents	932	24.6	28.6	41.5	5.2
Deaf or difficulty hearing	144	25.5	28.6	38.5	7.4	0.63
Serious difficulty seeing even wearing glasses	231	22.8	26.6	44.9	5.6	1.09
Serious difficulty with speech or communicating with others	229	21.0	30.7	42.8	5.5	3.21
Serious difficulty walking or climbing stairs	417	17.6	27.6	48.9	5.9	18.48*
Currently experience chronic pain	527	18.7	27.5	48.4	5.5	23.43*
Difficulty dressing or bathing	167	16.8	25.7	51.7	5.8	6.23
Difficulty doing errands alone due to condition	350	21.7	27.3	45.1	5.8	3.14
Serious difficulty concentrating, remembering, or making decisions due to condition	558	22.1	28.7	44.6	4.6	9.97*
Frequently have problems with work or other regular activities because feel depressed or anxious						17.71*
Always	243	17.8	30.3	46.5	5.4
From time to time	486	24.2	30.6	40.9	4.3
Never	203	33.8	21.4	37.2	7.6

Source. Weighted estimates from the Survey of Disability and Employment and Ohio administrative data.

Note. VR = vocational rehabilitation; IPE = Individualized Plan for Employment.

Statistically significant differences at the 5% level, based on chi-square test statistic.

Turning to the earnings outcomes, for most functional limitation questions, those who indicated having a certain functional limitation had moderately poorer earnings outcomes than did respondents who did not have that functional limitation (see Table 4). However, those whose functional limitation was deafness or difficultly hearing were exceptions. Those who reported that functional limitation were significantly more likely to have had a quarter of SGA level earnings than respondents as a group (33.7% vs. 26.1%;χ² = 5.78).

Table 4.

Differences in 18-Month Earnings by Functional Limitations at Application.

	n	18-month earnings			χ² test statistic
Functional limitation	n	Average monthly earnings (US$)	χ² test statistic	Average monthly earnings above SGA in one or more quarters (%)	χ² test statistic
All respondents	932	423		26.1
Deaf or difficulty hearing	144	516	1.08	33.7	5.78*
Serious difficulty seeing even wearing glasses	231	358	7.06*	20.4	4.06*
Serious difficulty with speech or communicating with others	229	396	0.404	23.5	1.10
Serious difficulty walking or climbing stairs	417	355	29.68*	21.9	7.10*
Currently experience chronic pain	527	394	11.46*	24.9	0.69
Difficulty dressing or bathing	167	407	3.69	25.0	0.07
Difficulty doing errands alone due to condition	350	320	17.96*	20.2	9.41*
Serious difficulty concentrating, remembering, or making decisions due to condition	558	379	0.22	26.0	0.04
Frequently have problems with work or other regular activities because feel depressed or anxious			1.18		0.098
Always	243	388		25.5
From time to time	486	412		26.5
Never	203	491		25.9

Source. Weighted estimates from the Survey of Disability and Employment and Ohio administrative data.

Note. SGA = substantial gainful activity.

Statistically significant differences at the 5% level, based on chi-square test statistic.

Differences by Employment Search, History, and Expectations

More than any other of the characteristics we examined, closure outcomes differ the most by employment expectations, history, and search (χ² = 25.81; see Table 5). Over 26% of those who said work was very or extremely important exited with employment whereas 14% who said work was somewhat important exited with employment, and only 8% who said work was not at all important exited with employment. Those who placed the least importance on work had notably higher rates of exit before getting a signed IPE (74%). The pattern is similar based on whether a respondent sees himself or herself working in the next 6 months (χ² = 26.04) and whether he or she was actively searching for work (χ² = 12.66). Twenty-six percent of those who saw themselves working in the next 6 months exited the VR program employed compared with the 8% who responded to the question with “no.”

Table 5.

Differences in Case Status, by Employment Perceptions and History at Application.

	n	VR case status (%)				χ² test statistic
Employment perception and history	n	Closed employed	Closed after IPE but without employment	Closed before IPE	Not yet closed	χ² test statistic
How important is it to you that you work?						25.81*
Extremely important	410	26.3	29.8	37.9	6.0
Very important	342	28.6	26.3	39.5	5.7
Somewhat important	148	14.1	32.9	50.7	2.3
Not at all important	31	7.6	18.4	74.0	0.0
Employed or self-employed at application						26.34*
Yes	270	35.1	21.1	38.7	5.1
No	662	20.1	31.8	42.8	5.3
Actively looking for work in past month						12.66*
Yes	368	24.6	34.4	36.4	4.5
No	292	13.8	28.0	51.7	6.5
When last worked						51.32*
One year or less	255	25.8	32.2	37.4	4.5
More than 1, up to 5	206	19.5	32.8	39.9	7.8
More than 5	170	12.7	29.4	54.4	3.5
Do you see yourself working in the next 6 months?						26.04*
Yes	438	25.5	35.7	32.7	6.1
No	185	7.9	22.3	66.2	3.6

Source. Weighted estimates from the Survey of Disability and Employment and Ohio administrative data.

Note. VR = vocational rehabilitation; IPE = Individualized Plan for Employment.

Statistically significant differences at the 5% level, based on chi-square test statistic.

Length of separation from employment was also significantly associated with closure outcomes (χ² = 51.32). Those employed or self-employed at application had the highest rates of employment (35%), followed by those who last worked within a year prior to application (26%), those who last worked 1 to 5 years prior to application (20%), and those who had last worked 5 or more years ago (13%). Those who had last worked 5 or more years ago had the highest rate of closing without a signed IPE (54.4%).

Baseline employment characteristics were also strongly correlated with future monthly earnings (see Table 6) and SGA level earnings. Respondents who stated that working was extremely important to them had the highest mean monthly earnings (US$609, χ² = 69.7) and quarterly SGA percentage (36%, χ² = 47.35). Unsurprisingly, those working at application had better earnings outcomes than members of other groups. The earnings outcome disparity between those working and those not expecting to work in 6 months was most substantive. For example, there was a US$700 monthly earnings difference and 32 percentage point quarterly SGA difference in 18-month outcomes between those already employed at VR application and those who did not see themselves working in 6 months.

Table 6.

Differences in 18-Month Earnings, by Employment Perceptions and History at Application.

	n	18-month earnings			χ² test statistic
Employment perception and history	n	Average monthly earnings (US$)	χ² test statistic	Average monthly earnings above SGA in one or more quarters (%)	χ² test statistic
How important is it to you that you work?			69.70*		47.35*
Extremely important	410	609		36.0
Very important	342	354		22.5
Somewhat important	148	127		11.2
Not at all important	31	33		2.6
Employed or self-employed at application			112.53*		52.49*
Yes	270	817		42.1
No	662	254		19.3
Actively looking for work in past month			39.44*		17.52*
Yes	368	335		24.9
No	292	147		12.0
When last worked			162.64*		86.28*
One year or less	255	396		31.0
More than 1, up to 5	206	182		11.8
More than 5	170	142		11.1
Do you see yourself working in the next 6 months?			52.03*		15.37
Yes	438	315		23.8
No	185	117		9.9

Source. Weighted estimates from the Survey of Disability and Employment and Ohio administrative data.

Note. SGA = substantial gainful activity.

Statistically significant differences at the 5% level, based on chi-square test statistic.

Differences by Reasons for Not Working

The final set of characteristics we examined were for the subset of SDE respondents who were not working when they applied for VR services (see Tables 7 and 8). This includes 662 respondents, which is about 71% of the overall analysis sample. Overall, 20% of these respondents exited with employment, 32% exited without employment, 43% exited before getting a signed IPE, and 5% were still receiving services 2 years after application. Mean monthly earnings were US$254, and 19% had at least one quarter of earnings in excess of the quarterly SGA amount. For this subgroup, we explored differences in these outcomes by their self-reported reasons why they were not working at application. Respondents could respond affirmatively or negatively to as many reasons as they wished; Tables 7 and 8 list the cited reasons in order of frequency.

Table 7.

Differences in Case Status by Reasons Stated for Not Working at Baseline.

	n	VR case status (%)				χ² test statistic
Reason for not working	n	Closed employed	Closed after IPE but without employment	Closed before IPE	Not yet closed	χ² test statistic
All respondents not employed at application	662	20.1	31.7	42.8	5.3
Reasons stated for not working
Physical or mental condition	468	15.8	31.7	47.3	5.1	37.35*
Cannot find a job	389	24.6	31.2	38.0	6.2	2.17
Discouraged by previous attempts	348	17.6	29.6	47.9	4.9	15.48*
Lack of skills	307	18.4	27.5	50.2	3.9	14.28*
Employers will not give them a chance	303	24.3	31.8	38.2	5.6	1.54
Workplaces are not accessible	221	21.7	32.6	41.0	4.8	0.67
Lack of reliable transportation	210	17.3	32.1	46.6	4.0	7.75*
Concern about losing disability and health benefits	171	17.7	32.4	43.1	6.8	8.52*
Discouragement from family or friends	163	10.5	30.2	57.6	1.7	26.84*
Finishing school or training	162	17.9	34.6	39.0	8.4	6.11
Caring for someone else	97	21.7	37.1	40.1	1.0	5.16

Source Weighted estimates from the Survey of Disability and Employment and Ohio administrative data.

Note. VR = vocational rehabilitation; IPE = Individualized Plan for Employment.

Statistically significant differences at the 5% level, based on chi-square test statistic.

Table 8.

Differences in 18-Month Earnings by Reasons Stated for Not Working at Baseline.

	n	18-month earnings			χ² test statistic
Reason for not working	n	Average monthly earnings (US$)	χ² test statistic	Average monthly earnings above SGA in one or more quarters (%)	χ² test statistic
All respondents not employed at application	662	254		19.3
Reasons stated for not working
Physical or mental condition	468	203	102.27*	15.7	55.63*
Cannot find a job	389	288	11.33*	22.5	3.65
Discouraged by previous attempts	348	245	29.59*	16.5	25.02*
Lack of skills	307	219	36.36*	15.6	24.27*
Employers will not give them a chance	303	306	2.15	22.0	2.97
Workplaces are not accessible	221	268	16.29	17.7	10.12*
Lack of reliable transportation	210	181	30.68*	14.8	17.46*
Concern about losing disability and health benefits	171	190	23.41*	13.1	18.84*
Discouragement from family or friends	163	137	39.16*	10.2	25.12*
Finishing school or training	162	201	16.12*	17.3	8.92*
Caring for someone else	97	212	3.26	17.6	3.48

Source. Weighted estimates from the Survey of Disability and Employment and Ohio administrative data.

Note. SGA = substantial gainful activity.

Statistically significant differences at the 5% level, based on chi-square test statistic.

Discouragement from family and friends, although not among the more commonly cited reasons, had the strongest relationship with exiting VR without employment and prior to signed IPE (χ² = 26.84). Just 11% of individuals who said they were not working because their family or friends discouraged them were employed at exit, while 58% of these respondents exited before IPE. Reporting that a physical or mental condition prevented work also was associated with substantially lower employment rates. Several other reasons that are more labor market or employer focused—including inaccessible workspaces, not being able to find a job, or not having a chance to show that one can work—showed the opposite pattern, with those who responded in the affirmative having higher, though not statistically significant, rates of exit with employment than the group as a whole.

The correlation between the reason for not working and earnings was weaker. Similar to the case status distribution, the relationship between mean monthly earnings and reason for not working varied by reason but did not display a strong trend. For 8 of the 11 reasons, identifying with that reason was associated with lower earnings. For three reasons—not being able to find a job, employers will not give them a chance, and inaccessible workplaces—the opposite was true: respondents reporting those reasons had higher than average earnings among all respondents who were not employed at application.

Discussion

The findings reveal that contextual or personal factors are correlated with VR applicant outcomes in meaningful ways. A few themes are particularly noteworthy because of their implications for policy and practice.

Examining the demographic characteristics findings, we see that the respondents’ VR case status outcomes reflect broader trends in the employment literature. For example, the greater an individual’s educational attainment, the more likely the individual was to exit the VR program employed or earn more than the quarterly SGA amount. This theme suggests that in some ways, the barriers to employment experienced by VR applicants may be similar to the barriers experienced by people without disabilities, such as variation in demand for unskilled versus skilled labor. Consequently, components of evidence-based employment interventions such as Job Corps, which provides job training to transition age youth and has strong evidence of efficacy, may benefit people with disabilities who want to work.

The quarterly SGA outcome results throughout the study reveal substantive heterogeneity in SGA level earnings across respondent groups. For example, those without a high school degree, Blacks, those having difficulty doing errands, and those discouraged from working by family and friends had SGA earnings percentages well below the average. The most striking variation in quarterly SGA earnings is found across the employment history and attitude measures. Relatively few respondents with long separations from the labor market or with little expressed interested in working earned more than the quarterly SGA amount, whereas the opposite was true for recent workers and more motivated applicants. Because employment history and attitudes are strongly associated with outcomes, a future study may want to further explore this correlation to determine if there is a causal link. If there is such a link, then counselors and SVRAs could benefit from obtaining more detailed employment history and attitude information at application and then using that information to identify applicants with more substantive barriers to employment.

Among the stated reasons for not working, discouragement from family and friends was consistently correlated with the poorest outcomes among those not working when they applied for VR services. It could be that this discouragement was associated with an unwillingness among family and friends to help the VR applicant overcome his or her barriers to employment. But whatever the mechanism driving the result, encouragement (and possibly support) from family and friends seems to be a particularly important factor for VR counselors to learn about from the VR clients they are assisting. The services provided to applicants who face this type of discouragement may need to include additional resources that are typically offered by friends and family.

One other theme that emerged from the results is that except for those with difficulty hearing, the presence of specific symptoms or functional limitations was almost always associated with poorer outcomes (relative to those who did not have the condition). Some of the poorest outcomes were for those with chronic pain, depression and anxiety, and difficulty with activities of daily living (such as walking, climbing stairs, dressing, and bathing). Though important to understanding the nature of applicants’ work barriers, these symptoms and limitations are not currently recorded in VR administrative data. Particularly noteworthy, more than half the analytic sample reports chronic pain, well above the national estimate that 17% of American adults experience pain most days or every day (Nahin, 2012). The current opioid addiction crisis has increased awareness about the debilitating potential of chronic pain and its treatments. Counselors should be aware of the prevalence of chronic pain among VR applicants and the risks that it and other functional limitations raise for successful VR and employment outcomes.

The study is descriptive and does have some limitations. The descriptive findings should not be interpreted as causal estimates of VR because we cannot control for various applicant and counselor selection issues. Because the study was confined to Ohio VR agency applicants, the findings may not be applicable to other states whose VR agencies or economic landscapes—and VR applicant pools—are dissimilar to Ohio’s. Participation in the survey was voluntary, and a minority of participants in Ohio (about 10%) did not consent to have their survey responses linked to future administrative data. Consequently, the analysis sample is not fully representative of Ohio VR respondents. Despite these limitations, the descriptive findings of the study provide important information on factors associated with successful and unsuccessful outcomes that is helpful to practitioners, policymakers, advocates, and other stakeholder groups.

Footnotes

Acknowledgements

The authors thank Rachel Vogt for excellent programming support and Todd Honeycutt and two anonymous referees for helpful comments to strengthen the article.

Declaration of Conflicting Interests

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

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was funded by the National Institute on Disability, Independent Living, and Rehabilitation Research (NIDILRR, U.S. Department of Health and Human Services [HHS]) through the Rehabilitation and Research Training Center on Individual Characteristics at Kessler Foundation, under cooperative agreement 90RT5017-01-01. The findings and conclusions are those of the authors and do not represent the policy of HHS or NIDILRR.

Notes

References

Anand

Sevak

(2017). The role of workplace accommodations in the employment of people with disabilities. IZA Journal of Labor Policy, 6, 12. doi:10.1186/s40173-017-0090-4

Beveridge

Fabian

(2007). Vocational rehabilitation outcomes: Relationship between individualized plan for employment goals and employment outcomes. Rehabilitation Counseling Bulletin, 50, 238–246.

Brucker

Boticello

O’Neill

Kutlik

(2017). Variations in social capital among vocational rehabilitation applicants. Journal of Vocational Rehabilitation, 46, 187–194.

Capella

M. E.

(2002). Inequities in the VR system: Do they still exist? Rehabilitation Counseling Bulletin, 45, 143–153.

Chow

C. M.

(2012). Reasonable job accommodations: An econometric investigation into strategies addressing barriers to employment for individuals with psychiatric disabilities (Doctoral dissertation). Available from ProQuest (3539932)

Cook

J. A.

Burke-Miller

J. K.

Grey

D. D.

(2015). Reasons for job separation among people with psychiatric disabilities (DRC Brief 2015–04). Washington, DC: Mathematica Center for Studying Disability Policy.

Dutta

Chan

Kundu

M. M.

Kaya

Brooks

Sánchez

Tansey

T. N.

(2017). Assessing vocational rehabilitation engagement of people with disabilities: A factor-analytic approach. Rehabilitation Counseling Bulletin, 60, 145–154.

Eckstein

Sevak

Wright

(2017). Personal characteristics of vocational rehabilitation applicants: Findings from the Survey of Disability and Employment. Journal of Vocational Rehabilitation, 46, 177–185.

Feist-Price

(1995). African Americans with disabilities and equity in vocational rehabilitation services: One state’s review. Rehabilitation Counseling Bulletin, 39, 119–129.

10.

Gilmore

D. S.

Bose

(2005). Trends in postsecondary education: Participation within the vocational rehabilitation system. Journal of Vocational Rehabilitation, 22(1), 33–40.

11.

Hill

M. J.

Maestas

Mullen

K. J.

(2016). Employer accommodation and labor supply of disabled workers. Labour Economics, 41, 291–303.

12.

Kristman

V. L.

Shaw

W. S.

Boot

C. R. L.

Delclos

G. L.

Sullivan

M. J.

(2016). Researching complex and multi-level workplace factors affecting disability and prolonged sickness absence. Journal of Occupational Rehabilitation, 26, 399–416.

13.

Loprest

Maag

(2007). The relationship between early disability onset and education and employment. Journal of Vocational Rehabilitation, 26, 49–62.

14.

Mann

D. R.

Honeycutt

Stegman Bailey

O’Neill

(2017). Using administrative data to explore the employment and benefit receipt outcomes of vocational rehabilitation applicants years after program exit. Journal of Vocational Rehabilitation, 46, 159–176.

15.

Martin

F. H.

(2010). Racial variation in VR outcomes: A structural equation modeling approach. Rehabilitation Counseling Bulletin, 54(1), 26–35.

16.

Nahin

R. L.

(2012). Estimates of pain prevalence and severity in adults: United States. The Journal of Pain, 16, 769–780.

17.

O’Neill

Kaczetow

Pfaller

Verkuilen

(2017). Impairment, demographics and competitive employment in vocational rehabilitation. Journal of Vocational Rehabilitation, 46, 149–158.

18.

Rehabilitation Services Administration. (2014). Vocational rehabilitation program case service report (RSA-911). Washington, DC: U.S. Department of Education, Office of Special Education and Rehabilitative Services.

19.

Rehabilitation Services Administration. (2016). Annual report fiscal year 2013. Washington, DC: U.S. Department of Education, Office of Special Education and Rehabilitative Services.

20.

Rosenthal

D. A.

Chan

Wong

Kundu

Dutta

(2006). Predicting employment outcomes based on race, gender, disability, work disincentives, and vocational rehabilitation service patterns. Journal of Rehabilitation Administration, 29, 229–244.

21.

Rosenthal

D. A.

Ferrin

J. M.

Wilson

(2005). Acceptance rates of African American versus white consumers of vocational rehabilitation services: A meta-analysis. Journal of Rehabilitation, 71(3), 36–44.

22.

Sevak

Houtenville

A. J.

Brucker

D. L.

O’Neill

(2015). Individual characteristics and the disability employment gap. Journal of Disability Policy Studies, 26, 80–88.

23.

Sevak

Khan

(2017). Psychiatric vs. physical disabilities: A comparison of barriers and facilitators to employment. Psychiatric Rehabilitation Journal, 40, 163–171.

24.

Shakespeare

(2013). The social model of disability. In Davis

L. J.

(Ed.), The disability studies reader. New York, NY: Routledge, 266-273.

25.

Stapleton

D. C.

Martin

(2017). What have we learned using merged administrative data from the social security administration and the rehabilitation services administration? Journal of Vocational Rehabilitation, 46, 121–136.

26.

Wilson

K. B.

(1999). Vocational rehabilitation acceptance: A tale of two races in a large midwestern state. Journal of Applied Rehabilitation Counseling, 30(2), 25–31.

27.

Wilson

K. B.

(2002). Exploration of VR acceptance and ethnicity: A national investigation. Rehabilitation Counseling Bulletin, 45, 168–176.

28.

Wilson

K. B.

Alston

R. J.

Harley

D. A.

Mitchell

N. A.

(2002). Predicting VR acceptance based on race, gender, education, work status at application, and primary source of support at application. Rehabilitation Counseling Bulletin, 45, 132–142.

29.

World Health Organization. (2001). International classification of functioning, disability and health. Geneva, Switzerland: World Health Organization Press.