The Role of Educational Attainment in Supported Employment

Abstract

Rehabilitation counselors provide vocational services to consumers living with serious mental illness (SMI) who have an estimated rate of workforce participation from 10% to 30%. Services, such as supported employment (SE), have strived to overcome these figures. Yet, people living with SMI are often only qualified for employment within the secondary labor market. Human capital theory offers a useful theoretical framework for employment for persons living with SMI. The Brief Symptom Inventory Global Severity Index (GSI) and educational level were used to predict employment outcomes in a sample which consists of 105 individuals with SMI recruited from five SE programs in the mid-Atlantic region. Logistic regression with the predictors of time in SE, GSI, and educational level achieved was used to predict whether someone became employed within the next 6 months. The variable educational level was a significant predictor of successful employment outcome at the 6-month follow-up, Wald χ² = 7.6, p = .003. The other two variables were not statistically significant. The current study suggests that it is difficult to ignore an individual’s education when considering his or her future achievement in employment. Furthermore, the findings of this study support utilizing human capital theory.

Keywords

supported employment vocational rehabilitation psychiatric rehabilitation human capital theory

Best practice in rehabilitation counseling should consider larger economic trends in the provision of vocational rehabilitation services. In the current knowledge-based economy, education or specialized training is crucial to gain entry into employment in which one could earn a living wage (Sharf, 2012; Sullivan, 2010). Most people living with serious mental illnesses (SMI) have intermittent recent work histories, which put these individuals in an almost intractable situation where they lose all attachment to the workforce, and develop an identity related to their status as a person receiving mental health services, or an illness identity (Yanos, Roe, & Lysaker, 2010). Workforce participation rates for persons with SMI remain very low, ranging from 10% to 30% (Bertram & Howard, 2006; Pratt, Gill, Barrett, & Roberts, 2014; Salkever et al., 2007). In addition, when persons living with SMI attempt to return to work, they are often only qualified for low-paying jobs, with no benefits, within the secondary labor market (Murphy, Mullen, & Spagnolo, 2005; Pratt et al., 2014; Waynor & Gill, 2015; Waynor, Gill, & Gao, 2016). To qualify for better paying jobs with benefits and a career ladder, obtaining more education or training is often necessary to rebuild one’s career after the onset of SMI (Mueser & Cook, 2012; Murphy et al., 2005).

Wewiorski and Fabian (2004) found, in a review of literature, that educational level was an ambiguous predictor of employment for people living with SMI in vocational programs. This review, published a decade ago, was largely based on articles from the 1980s and 1990s. The macroeconomic impact of globalization, with the loss of traditional manufacturing jobs in the U.S. economy, had not progressed to its current level. Thus, many participants from vocational programs in the Wewiorski and Fabian review could more readily enter positions and be trained on the job. Consequently, educational level may not have been the barrier to employment that it is today.

Human Capital Theory

In the last quarter of the 20th century, a theory that connects educational accomplishment to employment and earning emerged. Human capital theory is a relevant theory for rehabilitation counselors working with individuals pursuing vocational goals (Gao, Gill, Schmidt, & Pratt, 2010; Sharf, 2012). Human capital refers to knowledge, skills, and abilities acquired by an individual, traditionally through education and work experience (Bridges, De’Armond, & Dean, 2013). The basic premise of human capital theory is that increases in years of schooling and years of labor market experience lead to increased earnings (Becker, 1962). Skills gained through education, work experience, and training will give a person an advantage in the labor market (Gao et al., 2010; Sharf, 2012; Sullivan, 2010). The role of human capital is becoming increasingly crucial in our technology-based economy (Gao et al., 2010; Sullivan, 2010); consequently, rehabilitation counselors will need to consider how to help consumers living with disabilities gain human capital to be competitive in the labor market.

Recent research on human capital theory has focused on its role in career outcomes among individuals of various ability, ethnic, racial, socioeconomic, geographic, and gender classifications (Bridges et al., 2013; Chen & van der Klaauw, 2008; Gilleskie & Hoffman, 2014; Turner, 2014). Understanding the discrepancies in access to human capital among traditionally disenfranchised groups of people will support efforts to overcoming barriers to the development of human capital, which is essential for upward mobility. A key truism of human capital theory is that persons with a higher educational level receive higher pay and have greater job security (Gao et al., 2010; Sharf, 2012; Sullivan, 2010). In the current economy, this is becoming even more pronounced, and persons who lack access to human capital will fall further behind economically (Chen & van der Klaauw, 2008; Lewine, 2005; Linder, 2016). People living with disabilities are likely to have limited human capital and connection to the workforce; therefore, it becomes imperative for rehabilitation providers to focus on human capital theory.

Barriers to Human Capital: Psychiatric Disability

Gao, Schmidt, Gill, and Pratt (2011) studied the association between human capital including years of education, hireability, and earning potential for individuals with and without self-identified mental illness who utilized the services of a One-Stop career center. The human capital variables, educational level and total months worked in the past 5 years, made significant individual contributions and were predictive of hourly wages among participants with self-identified mental illness (Gao et al., 2011). The presence of a psychiatric disorder or symptomatology was not predictive of the likelihood of returning to work. Among persons who did return to work, educational level was predictive of earnings.

An important issue to consider regarding the Gao et al. (2011) study was participants were receiving employment services from the mainstream public employment system, and not employment services from the public mental health system. Therefore, it is not clear whether these findings would be replicated with a sample of participants receiving a rehabilitation intervention from the public mental health system. Furthermore, another issue to consider is that there are other forms of “capital” including financial capital and social capital. For the purpose of this study, we contend that the human capital variable of educational level may serve as an exemplar of social capital, as individuals with higher educational levels would have access to potentially larger social networks.

The current study is focused on one critical human capital variable, the educational level of supported employment (SE) participants living with SMI. This study assessed whether the human capital variable, educational level, is predictive of the achievement of an employment goal, after controlling for psychiatric symptoms, and time receiving SE that did not include educational services. The hypothesis is that a higher educational level will be predictive of achieving an employment goal, while controlling for a longer length of stay in SE, and symptomatology.

Method

Participants

The sample consists of 105 individuals with SMI recruited from five community mental health SE programs in the mid-Atlantic region. Self-reported diagnoses were categorized as schizophrenia spectrum disorder (41%, n = 39), bipolar disorder (32.5%, n = 34), major depressive disorder (22.9%, n = 24), and other (5.7%, n = 6). The sample was 59% male with a mean age of 44 (SD = 10.8). In addition, participants were ethnically diverse, with 57.1% Caucasian, 36.2% African American, 2.9% Hispanic, 1% Asian, and 2.9% identified as Other. More than two thirds of participants were receiving income support from the Social Security Administration, with 37% on Social Security Disability Insurance (SSDI), 19% on Supplemental Security Income (SSI), and 13% on SSI and SSDI. Approximately 10.5% were recipients of general assistance, and about 19% on other forms of income at baseline.

Procedure

Research staff met with participants at baseline, and after giving informed consent, participants were asked to complete an intake questionnaire to gather data on demographic information including education, work history, benefit status, diagnoses, disability history, and time receiving SE services. In addition, data were collected on the level of psychiatric symptoms. Research staff then met with participants for a 6-month follow-up assessment. At the follow-up meeting, research staff collected data on employment outcomes. Participants were asked to report on their employment activity, including participation in job seeking activities such as filling out applications, participating in job interviews, number of days employed, title and type of job, type of industry in which the job falls, number of hours per week employed, salary and benefits, and date of job termination (if applicable). The study protocol was approved by the University Institutional Review Board (IRB), and the research staff team consisted of one graduate student and several University faculty members.

Measures

Brief Symptom Inventory (BSI)

The BSI (Derogatis & Melisaratos, 1983) is a 53-item self-report measure of psychiatric symptoms. This instrument uses a 5-point scale ranging from 0 = not at all to 4 = extremely. This instrument asks individuals whether they experienced any of the following problems for a period within 1 week; items include “feeling no interest in anything” and “numbing and tingling in parts of your body.” In addition, a Global Severity Index (GSI) is computed, and an average score of 1.39 among all items answered is considered to be clinically significant (Derogatis & Melisaratos, 1983). The alpha coefficient was .96 for the entire scale in the current study (Waynor & Gill, 2015; Waynor et al., 2016).

Time in SE

To enroll in the study, participants were required to be not employed for at least 1 month to be eligible for the study; participants’ length of participation in SE varied markedly. Therefore, as a potential confounding contextual support variable, the time enrolled in SE at baseline was measured in months.

Employment outcome

The criterion measure for the hypothesis is a discrete dichotomous variable of whether employment was achieved in either a part-time or full-time competitive job during the 6-month follow-up. The employed group was coded as 1, whereas the not employed group was coded as 0.

Educational level

Due to missing data from the participants with less than a high school diploma and some college, educational level was measured as categorical data. Educational level was coded with 1 = no high school diploma; 2 = high school diploma; 3 = some college, but no degree; 4 = associate’s degree or bachelor’s degree; and 5 = master’s degree or professional degree. The rationale for including the associate’s degree and bachelor’s degree in a single category is because the associate’s degree is primarily focused on workforce, or as a step toward a higher degree. Therefore, participants with an associate’s degree were considered as college graduates in this analysis.

Data Analysis

To test the study hypothesis, hierarchical logistic regression will be employed to address whether educational level predicts employment outcomes for persons seeking employment in SE, while controlling for psychiatric symptoms and time receiving SE services. This method has the advantage of allowing the researcher to input the variables in the order of their choosing, thus allowing the variables to be assessed based on their theoretical importance in the model (Tabachnick & Fidell, 2007; Waynor & Gill, 2015; Waynor et al., 2016). Predictor variables will be entered in two blocks in the following order: (1) baseline GSI and baseline time in SE and (2) educational level; all analyses were conducted with SPSS version 20. To achieve a critical z value at least 80% of the time (power) and assuming 40% of participants will attain employment above an expected rate of 10%, a minimum odds ratio of 3.86 will be required and will require a minimum of 35 participants. The literature suggests that 10% is a lower estimate of how many people living with SMI are employed. With 82 participants, there will be more than sufficient power. Power calculations were computed with G-power 3.1.

Results

A total of 82 participants met with study staff for the 6-month follow-up, indicating an attrition rate of 22%. A total of 31 out of the 82 participants who met with study staff at 6 months achieved an employment goal, signifying a success rate of 38% among those participants who met for the follow-up assessment. In addition, most of the jobs were part-time within the secondary labor market (Waynor & Gill, 2015; Waynor et al., 2016). The predictor variables were entered in two blocks in the following order: (1) baseline GSI and baseline time in SE and (2) educational level. The variable educational level was a significant predictor of successful employment outcome at the 6-month follow-up, Wald χ² = 7.6, p = .003, with an R² = .15. The other two variables were not significant. Table 1 includes descriptive statistics for the predictor variables. Table 2 includes all the predictors entered into the equation at Block 2.

Table 1.

Descriptive Statistics for Predictor Variables.

Characteristic	n	M	SD
Psychiatric symptoms
GSI baseline	105	0.93	0.67
Months receiving SE services
Baseline time is SE	105	8.3	13.5
Educational level
No HS diploma	10 (9.5%)
HS graduate or GED	42 (40%)
Some college	27 (25.7%)
Associate’s degree	6 (5.7%)
Bachelor’s degree	14 (13.3%)
Master’s degree	6 (6%)

Note. GSI = Global Severity Index; SE = supported employment; HS = high school; GED = General Educational Development.

Table 2.

Logistic Regression Results for Baseline Predictors and 6-Month Employment Outcome (n = 82).

Variables	Wald	df	p
GSI	0.38	1	.75
Time in SE	0.001	1	.68
Educational level	7.60	1	.003

Note. GSI = Global Severity Index; SE = supported employment.

Level of education was rank-ordered as described in the “Method” section and correlated with employment in the first 6 months (1) or not (0). Education was positively correlated with being employed at 6 months, Spearman’s ρ(80) = .32, p = .002, one-tailed. Degrees of higher education were particularly associated with employment at 6 months. Among individuals with an associate’s degree or higher (n = 13), 69% (n = 9) were employed. Among those with less than an associate’s degree (n = 69), only 32% (n = 22) were employed. This difference in frequencies was significant, χ²(1) = 6.49, p = .01.

In addition, to examine whether another human capital variable, past work history, was related to the achievement of an employment outcome, a point-biserial correlational analysis assessed this relationship. The findings indicated that the work history variable of total months employed in the previous 5 years was not a predictor of an employment outcome—for the current study sample, r(80) = .044, p = .70.

Furthermore, to examine whether there were differences in educational level between ethnic, gender, and diagnostic categories, post hoc chi-square analyses were performed. The educational level variable was recoded into two levels: associate’s degree and above and less than associate’s degree. Ethnicity was recoded into two groups: White and non-White. In addition, diagnoses were recoded into two groups: schizophrenia spectrum disorder and other diagnosis. The findings indicated a nonsignificant relationship between ethnicity and educational level, χ²(1, 82) = 3.6, p = .06. In addition, the association between diagnosis and educational level was not significant, χ²(1, 82) = 0.55, p = .46. However, the relationship between gender and educational level was significant, χ²(1, 82) = 11.4, p = .001, with a higher proportion of female SE participants obtaining higher education.

Discussion

Among SE participants, education level predicted achievement of an employment goal. Symptoms and length of program participation were not predictive. These findings replicate the trend in the macroeconomy that higher education is related to greater employment success (Gao et al., 2010; Sharf, 2012; Sullivan, 2010). Recent research includes efforts to understand the discrepancies in access to human capital among traditionally disenfranchised groups of people (Bridges et al., 2013; Chen & van der Klaauw, 2008; Gilleskie & Hoffman, 2014; Turner, 2014). These findings emphasize that persons who lack access to human capital will fall further behind economically (Chen & van der Klaauw, 2008; Lewine, 2005; Linder, 2016). Interestingly, even those with higher education in the current study tended to obtain employment in the secondary labor market. However, this may not be too surprising as SE is an intervention designed for people with significant barriers to employment. Furthermore, several study participants who obtained college degrees shared with study staff that they were seeking part-time work so they could also focus on managing their health and well-being. Consequently, rehabilitation counselors will need to consider how to help consumers living with disabilities gain human capital to be competitive in the labor market.

The current findings support the salience of educational level among people living with SMI in the current labor market. In addition, the current findings replicate the finding in the Gao et al. (2010) study of the critical role of educational level, despite important differences between the samples. In contrast to the findings in the Gao et al. study, past work history was not related to successfully returning to work for the current study sample. However, diagnostically, the sample in the Gao et al. study differed considerably, as only 3% of the participants had schizophrenia spectrum disorder and the vast majority, 60%, had major depression. Nonetheless, Gao et al found that nearly half, 48%, were experiencing clinically significant psychiatric symptoms in their sample. While in the current study the vast majority of the participants were clinically stable, but more than 40% had a diagnosis of schizophrenia spectrum disorder. In addition, the Gao et al. study assessed a sample that were collecting unemployment benefits and receiving services from the One-Stop career center, whereas the sample of the current study was primarily on disability benefits from the Social Security Administration, and receiving services from the public mental health system. Nevertheless, both studies support the salience of human capital theory for rehabilitation counselors working with people living with SMI and other disabilities.

Thus, whether the counselor is working in a vocational rehabilitation program or community employment program, human capital theory needs to be included in any discussion of work outcomes for persons with disabilities. Human capital theory posits that people make investments in their training and education, which can be seen in their participation in the labor market and earning potential (Gao et al., 2010; Sharf, 2012). Regarding people with SMI, this theory may have to be augmented to account for the significantly lower rates of employment that are encountered in the workforce. Yet, evidence suggests that there may be little difference between the desires of people with SMI compared with people without mental illness in regard to working (Gao et al., 2010).

Limitations

The major limitations of this study are in the area of recruitment and retention of participants including potential self-selection. It is not clear whether the characteristics of the individuals in the study differ from those who did not volunteer to enroll. Furthermore, in the design, “human capital” was considered, but not other types of capital: financial or social. It is likely that all the participants were fairly equivalent in terms of their financial capital; the majority were recipients receiving forms of public assistance and wary of losing both financial and medical benefits.

In terms of social capital, previous research from our group found that the number of nonpaid natural support people in an SE participant’s life was predictive of employment success, although the causal direction could not be established (Roberts et al., 2010). Furthermore, SE programs that promoted person-centered, natural supports also had better employment outcomes compared with those that did not (Roberts et al., 2010). This was not assessed in the present study. However, the finding that educational level was predictive of employment success in SE supports our contention that educational level increases access to greater social capital, whether from better networks in the job seeking process or in a superior ability to utilize the supports provided by SE staff.

Another limitation to consider was that the educational level variable was collected as a demographic variable. Data were not collected on when the participants last attended school, or obtained their highest degree. It is feasible that employers considered the attainment of an educational goal as an asset on the application, and hired individuals because higher educational level was perceived to be linked with stronger job skills. Furthermore, employers may consider applicants with a college degree to be more goal oriented and ambitious. Nonetheless, this issue was not explicitly explored in the current study.

A further issue that was not explored was the age of onset of psychiatric illness. It is likely that some participants had already completed their education before experiencing significant psychiatric disability, and had started careers, while others may have attempted to obtain postsecondary education while in the recovery process. In addition, some of these participants may have received various supports while obtaining educational goals. However, this was examined in the current study.

Implications

Implications from this study suggest that rehabilitation counselor’s need to pay attention to increasing the level of education of people living with disabilities and increasing access to postsecondary education should be considered both an important intervention and goal. There is sometimes an understandable preference for short-term training programs, over long-term academic programs. But in light of human capital theory, the findings of the present study highlight the importance of persons acquiring advanced training and education as early as possible in their lives. Both employment and education should come early and often. For example, there is evidence that employment services should be part of early intervention for psychosis (Kane et al., 2016).

One approach might be to integrate supported education (SEd) services with the evidence-based individual placement and support (IPS) SE (Manthey, Rapp, Carlson, Holter, & Davis, 2012; Murphy et al., 2005; Nuechterlein, Subotnik, Turner, Becker, & Drake, 2008). SEd is a current practice that is gaining a lot of attention in the rehabilitation field that helps students living with SMI pursue their educational goals (Mueser & Cook, 2012; Murphy et al., 2005). SEd includes methods such as modification in exams, note-takers, access to tutors, supportive counseling, and cognitive interventions (Mueser & Cook, 2012). By focusing on improving their skills and marketability through education, persons living with SMI may avoid the all too common trend for this population to slip into a lower socioeconomic status (Lewine, 2005). There is preliminary evidence that it can enhance both education and employment outcomes (e.g., Schindler & Sauerwald, 2013).

Nuechterlein et al. (2008) studied an integrated IPS SE and SEd program and found that 36% of the participants choose an educational goal, 31% selected employment, and 33% selected both an educational and employment goal. While Manthey et al. (2012) surveyed 67 IPS SE programs to determine whether they provide SEd services concurrently with SE. They found that 57% of these programs delivered some form of educational services and support. Recent studies examining the integration of IPS and SEd are a sign of adapting to changing macroeconomic trends, and it is likely the role of educational level in SE success may be more readily elucidated in current studies of this population.

In addition to encouraging a merger of SE and SEd, future research can benefit by looking to other disciplines to aid in the development of new methods to help prepare people with SMI for successful employment outcomes earlier in their lives (Dixon & Lieberman, 2015; Kane et al., 2016). Increasingly, SE and SEd are part of new early psychosis interventions which have been found to successfully improve outcomes (Browne & Waghorn, 2010; Kane et al., 2016). Providing individuals recently diagnosed with psychotic disorders with an integrated intervention, including medication management and family psychoeducation, individualized therapy with SE and SEd shows promise to prevent individuals with these disorders from spiraling into a life of disability (Kane et al., 2016). Importantly, encouraging the enhancement of human capital plays a prominent role in this process.

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: This research was supported by a grant from the UNDNJ-SHRP Department of Psychiatric Rehabilitation’s Center for the Promotion of Recovery.

References

Becker

G. S.

(1962). Investment in human capital: A theoretical analysis. Journal of Political Economy 70(5), 9–49.

Bertram

Howard

(2006). Employment status and occupational care planning for people using mental health services. Psychiatric Bulletin, 30, 48–51.

Bridges

De’Armond

Dean

L. R.

(2013). Importance of human capital in recovery from divorce for women. Journal of Financial Planning, 26(11), 40–47.

Browne

D. J.

Waghorn

(2010). Employment services as an early intervention for young people with mental illness. Early Intervention in Psychiatry, 4, 327–335. doi:10.1111/j.1751-7893.2010.00188.x

Chen

van der Klaauw

(2008). The work disincentive effects of the disability insurance program in the 1990’s. Journal of Econometrics, 142, 757–784. doi:10.1016/j.jeconom.2007.05.016

Derogatis

L. R.

Melisaratos

(1983). The Brief Symptom Inventory: An introductory report. Psychological Medicine, 13, 595–605.

Dixon

Lieberman

(2015). Early psychosis intervention services: A work in progress. Schizophrenia Bulletin, 41, 23–24. doi:10.1093/schbul/sbu143

Gao

Gill

K. J.

Schmidt

L. T.

Pratt

C. W.

(2010). The application of human capital theory in vocational rehabilitation for individuals with mental illness. Journal of Vocational Rehabilitation, 32, 25–33.

Gao

Schmidt

L. T.

Gill

K. J.

Pratt

C. W.

(2011). Building human capital to increase earning power among people living with mental illnesses. Psychiatric Rehabilitation Journal, 35, 117–124. doi:10.2975/35.2.2011.117.124

10.

Gilleskie

Hoffman

(2014). Health capital and human capital as explanations for health-related wage disparities. Journal of Human Capital, 8, 235–279. doi:10.1086/677855

11.

Kane

J. M.

Robinson

D. G.

Schooler

N. R.

Mueser

K. T.

Penn

D. L.

Rosenheck

R. A.

. . . Heinssen

R. K.

(2016). Comprehensive versus usual community care for first-episode psychosis: 2-year outcomes from the NIMH RAISE Early Treatment Program. American Journal of Psychiatry, 173, 362–372. doi:10.1176/appi.ajp.2015.15050632

12.

Lewine

R. J.

(2005). Social class of origin, lost potential, and hopelessness in schizophrenia. Schizophrenia Research, 76, 329–335. doi:10.1016/j.schres.2004.07.006

13.

Linder

(2016). How do unemployment insurance benefits affect the decision to apply for Social Security Disability Insurance? Journal of Human Resources, 51, 62–94.

14.

Manthey

T. J.

Rapp

C. A.

Carlson

Holter

M. C.

Davis

J. K.

(2012). The perceived importance of integrated supported education and supported employment services. Journal of Rehabilitation, 78, 16–24.

15.

Mueser

K. T.

Cook

J. A.

(2012). Supported employment, supported education, and career development. Psychiatric Rehabilitation Journal, 35, 417–420. doi:10.1037/h0094573

16.

Murphy

A. A.

Mullen

M. G.

Spagnolo

A. B.

(2005). Enhancing IPS: Promoting job tenure by integrating natural supports and supported education. American Journal of Psychiatric Rehabilitation, 8, 37–61. doi:10.1080/15487760590953948

17.

Nuechterlein

K. L.

Subotnik

K. L.

Turner

L. R.

Becker

D. R.

Drake

R. E.

(2008). Individual placement and support for individuals with recent-onset schizophrenia: Integrating supported education and supported employment. Psychiatric Rehabilitation Journal, 31, 280–290. doi:10.2975/31.4.2008.340.349

18.

Pratt

C. W.

Gill

K. J.

Barrett

N. M.

Roberts

M. M.

(2014). Psychiatric rehabilitation. San Diego, CA: Academic Press.

19.

Roberts

M. M.

Murphy

A. A.

Dolce

Spagnolo

A. B.

Gill

K. J.

Librera

(2010). A study of the impact of social support development on job acquisition and retention among people with psychiatric disabilities. Journal of Vocational Rehabilitation, 33, 203–207.

20.

Salkever

D. S.

Karakus

M. C.

Slade

E. P.

Harding

C. M.

Hough

R. L.

Rosenheck

R. A.

. . . Yamada

A. M.

(2007). Measures and predictors of community based employment and earnings for persons with schizophrenia in a multisite study. Psychiatric Services, 58, 315–324.

21.

Schindler

V. P.

Sauerwald

(2013). Outcomes of a 4-year program with higher education and employment goals for individuals diagnosed with mental illness. Work: Journal of Prevention, Assessment & Rehabilitation, 46, 325–336.

22.

Sharf

R. S.

(2012). Applying career development theory to counseling. Belmont, CA: Brooks/Cole.

23.

Sullivan

(2010). A dynamic analysis of educational attainment, occupational choices, and job searches. International Economic Review, 51, 289–317.

24.

Tabachnick

B. G.

Fidell

L. S.

(2007). Using multivariate statistics. Boston, MA: Pearson.

25.

Turner

R. N.

(2014). Occupational stratification of Hispanics, whites, and blacks, in southern rural destinations: A quantitative analysis. Population Research Policy Review, 33, 717–746. doi:10.1007/s11113-014-9324-y

26.

Waynor

W. R.

Gill

K. J.

(2015). The role of hope in predicting supported employment success. Journal of Rehabilitation, 81, 26–31.

27.

Waynor

W. R.

Gill

K. J.

Gao

(2016). The role of work related self-efficacy in supported employment for people living with serious mental illnesses. Psychiatric Rehabilitation Journal, 39, 62–67.

28.

Wewiorski

N. J.

Fabian

E. S.

(2004). Association between demographic and diagnostic factors and employment outcomes for people with psychiatric disabilities: A synthesis of recent research. Mental Health Services Research, 6, 9–21.

29.

Yanos

P. T.

Roe

Lysaker

P. H.

(2010). The impact of illness identity on recovery from severe mental illness. American Journal of Psychiatric Rehabilitation, 13, 73–93. doi:10.1080/15487761003756860