Utilizing propensity score matching analysis to examine the effects of postsecondary education on the employment outcomes of young adults with mental illnesses

Abstract

Background

Helping young adults with mental illnesses attend college will increase their chances of being employed with benefits, which can pave the way for a successful career and a middle-class life.

Objective

This study sought to examine whether receiving vocational rehabilitation (VR) services of four-year college or university training would lead to improved employment outcomes for young adults with mental illnesses.

Methods

Based on the Rehabilitation Services Administration (RSA)-911 data, a propensity score matching technique was employed to create a comparison group of young adults with mental illnesses who did not receive college or university training and a treatment group of those who did. The quasi-experimental design investigated differences in the employment outcomes between the two groups.

Results

Analyses revealed that the group who received college or university training was more likely to be engaged in competitive integrated employment upon exit from the VR program than the matched comparison group who did not receive the training. Participants with mental illnesses who received college or university training and had secured competitive integrated employment at the time of exit not only had significantly higher hourly wages and worked more hours per week than did non-college or university training recipients, they were also more likely to obtain employer-subsidized health insurance benefits.

Conclusions

College and university training is positively related to employment outcomes for young people with mental illnesses. Implications for state VR counsellors are discussed.

Keywords

mental health competitive employment vocational rehabilitation postsecondary education propensity score matching

Introduction

In the U.S., mental illnesses affect approximately 20.6% of the working-age population annually, with anxiety disorders and mood disorders, such as depression and bipolar disorder, being particularly prevalent. This means that nearly one in five working-age adults experiences a mental health condition each year, underscoring the significant public health burden posed by mental illness (National Institute on Mental Health [NIMH], n.d.). Despite increased awareness, mental illnesses—particularly conditions like schizophrenia, bipolar disorder, and major depressive disorder—remain highly stigmatized in both the United States and globally (National Institutes of Health & Biological Sciences Curriculum Study, 2007). Unfortunately, these negative perceptions extend into the workplace, where employers often hold more unfavorable views of workers with mental illnesses compared to those with physical disabilities. Such attitudes, rooted in stereotypes that individuals with mental illnesses are fragile and unreliable, create substantial barriers, including discrimination, limited employment opportunities, and heightened workplace stigma (Østerud, 2023). Rumrill et al. (2021) highlighted that individuals with mental illnesses face significant obstacles in the workforce, including high levels of unemployment, underemployment, and discrimination, despite existing legal protections. The challenges become even more pronounced as the severity of mental illness increases. For example, full-time employment rates drop significantly, from 61.7% among individuals without mental illnesses to just 38.1% among those with serious mental illnesses (Luciano & Meara, 2014).

Education plays a crucial role in enhancing employment prospects for both individuals with and without disabilities. Higher educational attainment equips individuals with the skills and knowledge necessary for specific careers and boosts their competitiveness in the labor market, making them more attractive candidates to potential employers. Data from the Federal Reserve Bank of St. Louis (2024), the U.S. Bureau of Labor Statistics (2016), and the National Center for Education Statistics (n.d.) consistently show that individuals with higher levels of education—such as an associate's or bachelor's degree—earn more and experience lower unemployment rates compared to those with only a high school diploma. This emphasizes the critical role of higher education in improving employment outcomes.

The impact of education on employment is particularly significant for individuals with disabilities, positioning postsecondary education as a key factor in improving their opportunities, income, and self-sufficiency (Collett, 2023; McDonough & Revell, 2010). According to the National Center for Education Statistics (2017), individuals with disabilities who have a bachelor's degree or higher have higher employment rates compared to those with lower levels of education, which helps reduce employment disparities between individuals with and without disabilities. The Administration for Community Living (n.d.) similarly highlights that access to higher education enhances employment prospects, income potential, and quality of life for individuals with disabilities by providing essential career skills. Recent research further supports the connection between educational attainment and economic stability, particularly for individuals with mental health conditions. Bond et al. (2023) found that higher levels of education are associated with increased lifetime earnings and improved employment outcomes for those with mental health issues. Similarly, Luciano and Meara (2014) demonstrated that individuals with mental illnesses who complete a college degree are four to six times more likely to be employed compared to those without a degree, illustrating the transformative impact of education for this population.

Despite these clear benefits, individuals with mental health issues face significant barriers to completing higher education. For example, a recent survey found that 47% of adults living with schizophrenia leave college before graduating, compared to a 27% dropout rate in the U.S. overall (Baldwin, 2018). Poor mental health is significantly linked to dropping out among students in both vocational training and higher education settings. Hjorth et al. (2016) found that males in higher education experiencing poor mental health are five times more likely to drop out, although no such association was observed for females. Additionally, a 2019 study by the American Council on Education reported that students with poor mental health tend to have lower GPAs, take longer to complete their degrees, and drop out entirely (Tandet, 2023).

The state-federal vocational rehabilitation (VR) program, the largest provider of employment support for individuals with disabilities in the United States, plays a crucial role in facilitating access to higher education for people with disabilities, including those with mental illnesses, across all 50 states and U.S. territories (Rubin et al., 2016). Serving over one million individuals annually with a budget of $3.9 billion for fiscal year 2023, the VR program has a long history of successfully supporting individuals in achieving their independent living and employment goals (U.S. Department of Education, Rehabilitation Services Administration [RSA], 2023). Under the Workforce Innovation and Opportunity Act (WIOA) of 2014, VR services are available to youth with disabilities from age 14, including those with mental health issues (U.S. Bureau of Labor, 2014). Among these services is access to four-year college or university training at no cost to VR consumers, highlighting the critical role of educational support in improving employment outcomes. Rubin et al. (2016) emphasized that postsecondary education, particularly college training, is among the most cost-effective VR services for individuals with disabilities. Higher education has a significant impact on employment and lifetime earnings, not only for those without disabilities but also for individuals with mental health conditions. It acts as a protective factor against long-term unemployment and contributes to greater socioeconomic stability.

Despite the well-established benefits of postsecondary education on employment and earnings, limited research has focused on the impact of state-federal VR agencies’ “college or university training” services on employment outcomes for young adults with mental illnesses. To address this gap, this study examined the effect of college or university training on employment outcomes for young adults with mental illnesses who received VR services. To strengthen causal inference, the study utilized propensity score matching (PSM) technique to create matched samples of state VR consumers who received college or university training (the treatment group) and those who did not (the comparison group; Dahabreh et al., 2012; Rosenbaum & Rubin, 1983). PSM is particularly effective in addressing selection bias that arises from non-random assignment in observational studies. By matching individuals with similar characteristics, PSM helps to simulate the conditions of a randomized controlled trial. In this study, the propensity score—the probability of receiving college or university training—was estimated based on various observed covariates, gender, race, Medicaid, and assessment at intake. This matching process minimizes confounding factors, making it more likely that differences in employment outcomes can be attributed to the training rather than pre-existing differences between groups. Ultimately, this approach provides a robust analysis of the effectiveness of “college or university training” in improving employment outcomes, offering valuable insights for enhancing VR services for young adults with mental illnesses. The specific research questions addressed in this study were:

Is college or university training an effective intervention for increasing the rate of competitive integrated employment among young adults with mental illnesses who receive services from the state-federal VR program?

Is college or university training an effective intervention for improving the quality of employment outcomes for young adults with mental illnesses who achieve competitive integrated employment at the time of VR case closure?

Method

Data source

Data collected through the U.S. Department of Education, Rehabilitation Services Administration Case Service Report (RSA-911) database in Program Year 2018 for the state VR services programs were used in this study. The database's elements include participants’ demographic information at application, eligibility, order of selection, disability, trial work experience, individual plan for employment, pre-employment transition services, VR services, training services, career services, other services, measurable skill gain, employment, exit, and post-exit outcomes (U.S. Department of Education, Office of Special Education and Rehabilitative Services Rehabilitation Services Administration, 2017). The Rehabilitation Services Administration uses these data to assess, support, and evaluate the effectiveness of state VR agencies’ performance.

Participants

Several filter criteria were applied to identify the study sample. These filter criteria resulted in participants who: (a) were eligible VR applicants with an assigned individualized plan for employment (IPE); (b) were aged between 18 and 35 years old; (c) had a primary diagnosis of depressive and other mood disorders, anxiety disorders, mental illness, personality disorders, or Schizophrenia and other psychotic disorders; (d) had a high school diploma or less education at intake; (e) had an hourly wage at exit of $300 or less as the evaluation of the data quality in this study indicated that values greater than $300 on hourly wage would be either extreme outliers or a case of miscoding; and (f) had provided complete data on all the variables used in this study. The selection procedure resulted in a sample of 34,995 young adults with mental illnesses out of the 833,150 records in the Program Year 2018 data file, 367 of whom had received four-year college or university training.

Demographic information regarding the study sample is presented in Table 1. About 53% of the sample were male (n = 18,481), whereas some 47% (n = 16,514) were female. In descending order, the three largest race/ethnicity groups in the sample were Caucasians (n = 19,379; 55.4%), African Americans (n = 8,304; 23.7%), and Hispanics (n = 5,468; 15.6%). Multi-racial groups (n = 919; 2.6%), Asians (n = 530; 1.5%), American Indians (n = 295; 0.8%), and Native Hawaiians and Pacific Islanders (n = 100; 0.3%) occupied smaller proportions of the study sample (5.2% total). A little more than half of the sample (n = 18,547; 53%) were receiving Medicaid at the time of enrollment in the VR program. About 41% of the sample (n = 14,235; 40.7%) reported being long-term unemployed, and slightly less than two-thirds of the sample (n = 21,975; 62.8%) were classified as having low income. The average age of the sample was 25.4 years old, with a standard deviation of 5.5 years.

Table 1.

Demographic statistics of the study sample (N = 34,995).

Variable	Category	N	%	Mean (SD)
Gender	Female	16,514	47.2
	Male	18,481	52.8
Race	Caucasian	19,379	55.4
	African American	8,304	23.7
	American Indian	295	0.8
	Asian	530	1.5
	Hawaiian	100	0.3
	Hispanic	5,468	15.6
	Multi-racial	919	2.6
Medicaid	No Medicaid	16,448	47.0
	Medicaid	18,547	53.0
Low income	No	13,020	37.2
	Yes	21,975	62.8
Long-term unemployment	No	20,760	59.3
	Yes	14,235	40.7
Age		34,995	–	25.41 (5.46)

Variables and measures

There were four dependent variables in this study, all having to do with participants’ employment status at the time they exited the VR program: competitive employment, hourly wage, weekly working hours, and employer-based health insurance. Competitive integrated employment (CIE) was a dichotomous variable indicating whether a participant exited the program in competitive integrated or supported employment or not (i.e., 0 = no; 1 = yes). Hourly wage and weekly working hours were both continuous variables.

Employer-based health insurance was a dichotomous variable indicating whether a participant exiting the program obtained a health insurance through employer or not (i.e., 0 = no; 1 = yes). The independent variable or grouping variable was a dichotomous variable representing whether participants received four-year college or university training or not (i.e., 0 = no; 1 = yes). Young adults receiving four-year college or university training consisted of the treatment group.

Five covariates were also used in this study. These included age at application, gender (i.e., 0 = female; 1 = male), race (i.e., 0 = Caucasian; 1 = non-Caucasian), Medicaid (i.e., 0 = without Medicaid, 1 = with Medicaid), and assessment at intake (i.e., 0 = not receiving assessments at intake; 1 = receiving assessments at intake). Age at application was a continuous variable and the others were dichotomous variables. Among the covariates, Medicaid served to represent participants’ economic status and assessment at intake gauged participants’ potential to pursue postsecondary education.

Data analysis

Because young adults with mental illnesses who received four-year college or university training (recipients) and those who did not (non-recipients) may differ with regard to both their employment outcomes and other relevant characteristics, the findings from a simple, direct comparison between the two groups using the observational data that are available in the RSA-911 database could be misleading. To examine in a more controlled way the relationship between receipt of four-year college or university training and employment outcomes, we utilized propensity score matching (PSM) to create a group of non-recipients that is equivalent and comparable on the important demographic and background covariates to the group of recipients. PSM has been commonly used as an alternative method or a quasi-experimental technique to estimate causal effects in observational studies where potential selection biases need to be corrected by balancing or equating the groups based on the matching covariates (Rajeev & Sadek, 2002; Rosenbaum & Rubin, 1983). In this study, the PSM technique identified a sub-set of non-recipients in the sample whose estimated propensities obtained from the matching covariates were statistically equivalent to the recipients’.

All the analyses were conducted using the statistical package STATA 17 (StataCorp, 2021). Before selecting the matching non-recipients (i.e., the comparison group), the balance of the two groups of young adults with mental illnesses in the database was assessed. In other words, the two groups (recipients or the treatment group and non-recipients or the comparison group) were compared on the identified matching covariates of age at application, gender, race, Medicaid, and assessment at intake. Next, the PSMATCH2 module was used to conduct 1-to-1 nearest neighbor matching without replacement conditioned on the common support (Leuven & Sianesi, 2003). A propensity score, the probability of being in the four-year college or university training group, was estimated for each individual in the sample through a probit regression model based on the matching covariates. The estimated propensity scores were used to match each recipient with one non-recipient. After the matching procedure, the PSTEST module was utilized to assess the quality of matching. The group means or expected frequencies of matching covariates between the recipient group (i.e., the treatment group) and the identified matching non-recipient group (i.e., the comparison group) were compared through t-tests or chi-squared tests. If the matching non-recipient group was equivalent to the recipient group on the matching covariates, the t-tests and the chi-squared tests showed non-significant results.

To examine the effects of receiving four-year college or university training on the employment outcome of the two groups of young adults with mental illnesses, a chi-squared test was first performed to investigate if receiving the college or university training was associated with the chance of participating in competitive integrated employment at the time of exit from the VR program. Next, only those young adults with mental illnesses in both the treatment and the comparison groups who obtained competitive integrated employment at the exit were retained in the analysis; independent sample t-tests were conducted to compare if there were significant group differences in their hourly wages and weekly working hours and a chi-squared test was performed to examine if having an employer-based health insurance at exit was associated with the group membership. For all the independent sample t-tests performed in this study, the assumption of equal variance was always tested first. Whenever a violated assumption was identified, the Welch's t-test was used instead to address the bias caused by the unequal variances. The alpha significance level for the hypothesis tests was set at .05.

Results

Preliminary analyses

Before the matching procedure, we examined the comparability between the four-year college or university training recipients and the non-recipients on the matching covariates. The results indicated that the two groups were significantly different in their composition (see Table 2). Relatively, there were significantly more female (χ²= 38.22, df = 1, p < .001) and Caucasian (χ²= 22.33, df = 1, p < .001) participants in the recipient group than in the non-recipient group; the recipient group had a significantly higher percentage of participants who did not have Medicaid insurance (χ²= 23.91, df = 1, p < .001) and completed assessments at the intake (χ²= 29.65, df = 1, p < .001). In addition, the independent sample t-test showed that the training recipients were significantly younger than their non-recipient peers at the time of application (t = 8.22, df = 34,993, p < .001). These preliminary test results called for a matching procedure that would make the two groups more equivalent. Leaving the pre-existing group differences unaddressed would have led to biased and misleading results.

Table 2.

Comparability between training recipients and non-recipients on matching covariates before the matching procedure (N = 34,995).

Variable	Category	Recipients (n = 367)	Non-recipients (n = 34,628)	χ² _(df)	p
Gender	Female	232 (63.22%)	16,282 (47.02%)	38.22 ₍₁₎	.000***
	Male	135 (36.78%)	18,346 (52.98%)
Race	Caucasian	248 (67.57%)	19,131 (55.25%)	22.33 ₍₁₎	.000***
	Non-Caucasian	119 (32.43%)	15,497 (44.75%)
Medicaid	No Medicaid	219 (59.67%)	16,229 (46.87%)	23.91₍₁₎	.000***
	With Medicaid	148 (40.33%)	18,399 (53.13%)
Assessment at intake	No	258 (70.30%)	28,200 (81.44%)	29.65₍₁₎	.000***
Assessment at intake	Yes	109 (29.70%)	6,428 (18.56%)
		Mean (SD)		t _(df)	p
Age		23.08 (5.56)	25.43 (5.46)	8.22_(34,993)	.000***

Note. *** p < .001.

Propensity score matching procedure

Based on the five matching covariates, a 1-to-1 PSM procedure was performed, whereby a subset of 367 young adults with mental illnesses from the original non-recipient group was matched with the 367 recipients of college or university training. The matching quality was assessed to ensure that the balance on covariates had truly been achieved through the matching procedure. Specifically, chi-squared tests and a t-test were conducted to examine if the recipient (treatment) group and the matching non-recipient (comparison) group exhibited differences in their status on the covariates. As is displayed in Table 3, all matching covariates that were significantly different between the two groups prior to matching were no longer significantly different between the group of 367 four-year college or university training recipients and the matching comparison group of 367 non-recipients.

Table 3.

Comparability between training recipients and non-recipients on matching covariates after the matching procedure (N = 734).

Variable	Category	Recipients (n = 367)	Non-recipients (n = 367)	χ² _(df)	p
Gender	Female	232 (63.20%)	232 (63.20%)	0.00₍₁₎	1.00
	Male	135 (36.80%)	135 (36.80%)
Race	Caucasian	248 (67.60%)	248 (67.60%)	0.00₍₁₎	1.00
	Non-Caucasian	119 (32.40%)	119 (32.40%)
Medicaid	No Medicaid	219 (59.70%)	219 (59.70%)	0.00₍₁₎	1.00
	With Medicaid	148 (40.30%)	148 (40.30%)
Vocational assessment	No	258 (70.30%)	258 (70.30%)	0.00₍₁₎	1.00
Vocational assessment	Yes	109 (29.70%)	109 (29.70%)
		Mean (SD)		t _(df)	p
Age		23.08 (5.56)	23.08 (5.56)	0.00₍₇₃₂₎	1.00

Matched group comparisons

To answer the research questions, a chi-squared test of independence based on the matched data was first performed to investigate the relationship between the outcome variable of competitive integrated employment at exit and the group membership independent variable. The chi-squared test result indicated that obtaining competitive integrated employment at exit was positively associated with receipt of four-year college or university training (χ²= 7.94, df = 1, p < .01, see Table 4). Young adults with mental illnesses who received four-year college or university training were more likely to secure competitive integrated employment (67.3%) than their peers in the matching comparison group who did not receive college or university training (57.2%). The effect size of the association (Cramer's V = .10) was small by Cohen's definition (1988; i.e., .1 = small, .3 = moderate, and .5 = large).

Table 4.

Comparison between training recipients and non-recipients in competitive integrated employment at exit after the matching procedure (N = 734).

Variable	Category	Recipients (n = 367)	Non-recipients (n = 367)	χ² _(df)	p	Cramer's V
Competitive employment	No	120 (32.70%)	157 (42.78%)	7.94 ₍₁₎	.005**	0.104
at exit	Yes	247 (67.30%)	210 (57.22%)

Note. ** p < .01.

Next, only participants who obtained competitive integrated employment were retained in the analysis to answer the second research question. Independent sample t-tests were conducted to compare group means on the two continuous outcomes, namely, hourly wages and weekly working hours. Table 5 displays the group means on these two outcome variables and the test results. The two groups differed significantly in their hourly wages (Welch's t = 7.70, df = 439.62, p < .001) and weekly working hours (t = 5.61, df = 455, p < .001). On average, hourly wages of those competitively employed participants who received college or university training (M = $15.50, SD = $6.60) were $4.06 higher than those of their peers in the comparison group (M = $11.44, SD = $4.63). Young adults with mental illnesses who received college or university training (M = 34.00 h, SD = 9.40 h) also worked about 5 h (M__diff= 4.98) more per week than their peers in the comparison group (M = 29.02 h, SD = 9.52 h). According to Cohen's (1988) conventions (i.e., 0.2 = small, 0.5 = medium, and 0.8 = large), the effect size of the difference in hourly wages (d = 0.70) and in weekly working hours (d = 0.53) were large and medium, respectively.

Table 5.

Comparison between training recipients and non-recipients in hourly wage and weekly working hours after the matching procedure (N = 457).

	Recipients (n = 247)		Non-recipients (n = 210)
Variable	Mean	SD	Mean	SD	t(df)	p	Cohen's d
Hourly wage	15.50	6.60	11.44	4.63	7.70^a _(439.62)	.000***	0.70
Weekly working hours	34.00	9.40	29.02	9.52	5.61 ₍₄₅₅₎	.000***	0.53

Note. *** p < .001.

^aWelch's t-test.

In addition, a chi-squared test was conducted for the dependent variable of employer-based health insurance at exit; the result indicated that having an employer-based health insurance at exit was positively associated with receipt of four-year college or university training (χ²= 25.19, df = 1, p < .001, see Table 6). Young adults with mental illnesses who received college or university training and had secured competitive integrated employment by the time they exited VR services were more likely to obtain health insurance coverage through their employers (35.4%) than their peers in the comparison group who did not receive college or university training (14.4%). The effect size of the association (Cramer's V = .24) was close to medium according to Cohen's definition (1988; i.e., .1 = small, .3 = medium, and .5 = large).

Table 6.

Comparison between training recipients and non-recipients in employer-based health insurance at exit after the matching procedure (N = 444).

Variable	Category	Recipients (n = 243)	Non-recipients (n = 201)	χ² _(df)	p	Cramer's V
Employer-based health insurance	No	157 (64.61%)	172 (85.57%)	25.19 ₍₁₎	.000***	0.238
at exit	Yes	86 (35.39%)	29 (14.43%)

Note. *** p < .001.

Discussion

This study provides compelling evidence that postsecondary education, specifically state-federal VR agencies’ “four-year college or university training,” significantly enhances employment outcomes and job quality for young adults with mental illnesses receiving VR services. Compared to their peers without such training, participants were more likely to achieve competitive integrated employment upon exiting the VR program (67.30% compared to 57.22%). This finding underscores the vital role of postsecondary education in equipping young adults with the skills, knowledge, and credentials that employers seek, thereby increasing their employability and enhancing their chances of securing sustainable employment in competitive environments. In addition to higher employment rates, individuals who completed four-year college or university training experienced significantly improved economic outcomes. Participants earned higher hourly wages ($15.50 versus $11.44) than those without such training. This wage differential highlights the economic value of higher education, especially for young adults with mental illnesses, as it enables them to access positions that offer greater financial compensation. Higher wages are a key factor in fostering economic independence, allowing individuals to reduce their reliance on social services and achieve a more stable financial future. By helping young adults with mental illnesses secure better-paying jobs, postsecondary education contributes directly to their economic self-sufficiency, which is essential for their long-term well-being.

Moreover, participants who received four-year college or university training were found to work more hours per week compared to those without such training (34 h vs 29 h). This suggests that higher education not only helps these individuals secure employment but also prepares them for roles that require greater engagement and work commitment. The ability to work more hours reflects a deeper integration into the workforce and suggests that individuals with postsecondary education are better equipped to handle the demands of competitive employment. These outcomes highlight how education can bridge the gap between employability and meaningful participation in the labor market (Strauser, 2021).

An equally important aspect of job quality is access to employer-sponsored benefits, particularly health insurance. The study revealed that individuals who received four-year college or university training were more likely to have employer-sponsored health insurance (35.39% versus 14.43%). Health insurance is a crucial measure of employment quality, as it directly impacts the health and financial stability of individuals with mental illnesses. Access to health insurance not only provides essential medical support but reduces the financial burden associated with healthcare costs, thereby promoting overall health stability and security. By providing opportunities for young adults to obtain jobs with such benefits, postsecondary education plays a pivotal role in (a) enhancing their quality of life and (b) supporting their ability to maintain employment and provide for themselves and their families.

These findings collectively demonstrate that “four-year college or university training” provided by VR agencies is a critical service that can significantly impact both employment outcomes and employment quality for young adults with mental illnesses. Postsecondary education equips these individuals with the skills and competencies needed to improve their employability, placing them in a stronger position within the labor market. This aligns with previous research highlighting the role of higher education in reducing employment disparities for individuals with disabilities and promoting their inclusion in integrated work environments (Collett, 2023). By empowering young adults with mental illnesses to acquire the skills and qualifications that enhance their employability, postsecondary education reduces barriers to employment and fosters greater social and economic inclusion.

To fully realize the benefits of postsecondary education, it is essential that participants in college or university training programs receive adequate mental health supports and accommodations. Young adults with mental illnesses often face unique challenges during their educational journeys, which can be effectively mitigated by providing access to academic accommodations, mental health services, and mentorship. These supports are critical in helping them complete their studies successfully, ensuring they can fully take advantage of the opportunities that postsecondary education offers. Additionally, incorporating career preparatory activities—such as internships, networking opportunities, and job search assistance—is vital to enhancing workforce readiness. These activities provide practical experience, industry connections, and the confidence needed for a smooth transition into employment.

For VR professionals, actively promoting postsecondary education as a viable and advantageous service is key. This means not only advocating for educational opportunities but also offering individualized support that aligns with each participant's mental health needs. Once participants secure employment, ongoing support becomes especially important for sustaining positive outcomes. Implementing supported employment models, like Individual Placement and Support (IPS), can further enhance employment success by providing tailored job support during and after college, helping to bridge the transition from education to employment (Strauser, 2021). These models offer sustained assistance that addresses the unique challenges young adults with mental illnesses may face in the workplace. Additionally, focusing on the psychosocial aspects of the work environment—such as fostering a supportive workplace culture and providing robust support systems—ensures that these individuals are placed in environments conducive to their mental well-being, ultimately supporting long-term employment success.

Investing in “four-year college or university training” not only boosts the likelihood of employment but also supports economic independence, access to employer-sponsored benefits, and a deeper connection to the workforce. This type of training can be transformative, improving the quality of life for young adults with mental illnesses by empowering them to achieve financial stability, gain necessary healthcare access, and integrate fully into society. These investments foster long-term economic growth and contribute significantly to the social inclusion of individuals with mental illnesses, making them a vital aspect of state-federal VR services.

Limitations

The authors acknowledge several limitations that may affect the interpretation of results. First and perhaps foremost, the potential selection bias caused by the observational nature of the RSA-911 data could have impacted study results even though the rigorous PSM technique was used to isolate the effects of college or university training on participants’ employment outcomes. PSM does not always eliminate all possible selection bias even though it mimics the randomization found in experimental design (Rosenbaum & Rubin, 1983). Although the two groups in this study were determined to be equivalent on important covariates through the PSM procedure, we must acknowledge that there was no true random assignment of participants into the two groups. Therefore, caution is warranted when using the findings in this study to draw causal inferences.

Also delimiting is the assumption in PSM that unobserved characteristics of the sample do not affect group membership. Unobserved variables do exist, of course, and it is not known how they may have affected whether a particular VR client in this study attended a college or university as part of their Individualized Plan for Employment (IPE). More covariates could be collected and explored in future studies to help increase the accuracy of matching.

Conclusion

Incontrovertible is the fact that Americans who attend and complete college have better employment outcomes than Americans who do not. This study revealed evidence of the effects of college or university training on employment outcomes in a sample of young adults with mental illness, perhaps indicating that postsecondary education may serve as an important equalizer to level employment opportunities and outcomes for young workers with psychiatric disabilities. State VR counselors should consider postsecondary education as an important service that leads to success in competitive integrated employment settings for the growing and industrious population of jobseekers with mental illness. To that end, particular emphasis should be placed on mental health supports during college and the psychosocial features of the work environment after graduation.

Footnotes

Acknowledgment

The authors declare no further acknowledgments.

Ethics statement

Kent State University approved the study and the IRB approval number is Kent State University IRB#20-499.

Informed consent

Not applicable.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The contents of this paper were developed under a grant from the National Institute on Disability, Independent Living, and Rehabilitation Research (NIDILRR grant #90RTEM0003). NIDILRR is a Center within the Administration for Community Living (ACL), Department of Health and Human Services (HHS). The contents of this paper do not necessarily represent the policy of NIDILRR, ACL, or HHS, and you should not assume endorsement by the Federal Government.

Declaration of conflicting interests

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

References

Administration for Community Living (n. d.). Advancing independence and inclusion of older adults and people with disabilities. https://acl.gov/

Baldwin

(2018). A diagnosis of mental illness need not end a college career. https://www.nami.org/education/a-diagnosis-of-mental-illness-need-not-end-a-college-career/#:∼:text=Many%20young%20people%20with%20schizophrenia,discrimination%20is%20against%20the%20law

Bond

G. R.

Drake

R. E.

Becker

D. R.

(2023). Employment and education services for young adults with mental health conditions [Issue brief]. Advancing State Policy Integration for Recovery and Employment. https://www.dol.gov/sites/dolgov/files/ODEP/topics/pdf/Employment_and_Education_Services_for_Young_Adults_with_Mental_Health_conditions.pdf

Cohen

(1988). Statistical Power Analysis for the Behavioral Sciences. Lawrence Erlbaum.

Collett

J. W.

(2023). An introduction to this special issue of the Journal of Vocational Rehabilitation: Transition and individual supports in education and employment. Journal of Vocational Rehabilitation, 58(2), 99–102. https://doi.org/10.3233/JVR-230000

Dahabreh

I. J.

Sheldrick

R. C.

Paulus

J. K.

Chung

Varvarigou

Jafri

Rassen

J. A.

Trikalinos

T. A.

Kitsios

G. D.

(2012). Do observational studies using propensity score methods agree with randomized trials? A systematic comparison of studies on acute coronary syndromes. European Heart Journal, 33(15), 1893–1901. https://doi.org/10.1093/eurheartj/ehs114

Federal Reserve Bank of St. Louis (2024). Unemployment rate by educational attainment and age, monthly, not seasonally adjusted: 25 to 34 years. https://fred.stlouisfed.org/release/tables?eid=1197546&rid=50

Hjorth

C. F.

Bilgrav

Frandsen

L. S.

Overgaard

Torp-Pedersen

Nielsen

Bøggild

(2016). Mental health and school dropout across educational levels and genders: A 4.8-year follow-up study. BMC public Health, 16, 976. https://doi.org/10.1186/s12889-016-3622-8

Leuven

Sianesi

(2003). PSMATCH2: Stata module to perform full Mahalanobis and propensity score matching, common support graphing, and covariate imbalance testing. Statistical Software Components S432001, Boston College Department of Economics. http://ideas.repec.org/c/boc/bocode/s432001.html

10.

Luciano

Meara

(2014). Employment status of people with mental illness: National survey data from 2009 and 2010. Psychiatric Services (Washington, D.C.), 65(10), 1201–1209. https://doi.org/10.1176/appi.ps.201300335

11.

McDonough

J. T.

Revell

(2010). Accessing employment supports in the adult system for transitioning youth with autism spectrum disorders. Journal of Vocational Rehabilitation, 32(2), 89–100. https://doi.org/10.3233/JVR-2010-0498

12.

National Center for Education Statistics (2017). Disability rates and employment status by educational attainment. The Condition of Education. https://nces.ed.gov/programs/coe/pdf/coe_tad.pdf

13.

National Center for Education Statistics (n.d.). Employment outcomes of bachelor's degree holders. Condition of Education. https://nces.ed.gov/programs/coe/indicator/cbc

14.

National Institutes of Health & Biological Sciences Curriculum Study. (2007). NIH curriculum supplement series: Information about mental illness and the brain. National Institutes of Health (US). https://www.ncbi.nlm.nih.gov/books/NBK20369/

15.

National Institute of Mental Health (n.d.). Mental illness. https://www.nimh.nih.gov/health/statistics/mental-illness

16.

Østerud

K. L.

(2023). Mental illness stigma and employer evaluation in hiring: Stereotypes, discrimination and the role of experience. Sociology of Health & Illness, 45(1), 90–108. https://doi.org/10.1111/1467-9566.13544

17.

Rajeev

H. D.

Sadek

(2002). Propensity score-matching methods for nonexperimental causal studies. Review of Economics and Statistics, 84, 151–161. https://doi.org/10.1162/003465302317331982

18.

Rosenbaum

P. R.

Rubin

D. B.

(1983). The central role of the propensity score in observational studies for causal effects. Biometrika, 70(1), 41–55. https://doi.org/10.1093/biomet/70.1.41

19.

Rubin

S. E.

Roessler

R. T.

Rumrill

P. D.

(2016). Foundations of the vocational rehabilitation process (2nd ed.). PRO-ED.

20.

Rumrill

P. D.

Koch

L. C.

Strauser

D. R.

(2021). Introduction to the special issue: Employment and vocational rehabilitation considerations for people with disabilities during and after the COVID-19 pandemic. Journal of Vocational Rehabilitation, 54(1), 1–3. https://doi.org/10.3233/JVR-201121

21.

StataCorp (2021). Stata Statistical Software: Release 17. StataCorp LLC.

22.

Strauser

D. R.

(2021). Career development, employment and disability in rehabilitation: From theory to practice (2^nd ed.). Springer Publishing Company.

23.

Tandet

(2023). Improving Student Mental Health within Higher Education. Modern Campus. https://moderncampus.com/blog/student-mental-health.html#:∼:text=Investing%20in%20students’%20mental%20well,to%20graduate%20confident%20lifelong%20learners

24.

U.S. Bureau of Labor (2014). Workforce Innovation and Opportunity Act. https://www.dol.gov/agencies/eta/wioa

25.

U.S. Bureau of Labor Statistics (2016). Education matters. https://www.bls.gov/careeroutlook/2016/data-on-display/education-matters.htm#:∼:text=According%20to%20data%20from%20the,than%20those%20with%20less%20education

26.

U.S. Department of Education, Office of Special Education and Rehabilitative Services Rehabilitation Services Administration (2017). Reporting manual for the case service report (RSA-911). https://rsa.ed.gov/sites/default/files/subregulatory/pd-16-04.pdf

27.

U.S. Department of Education, Rehabilitation Services Administration. (2023). Fiscal year 2023 budget request. https://www2.ed.gov/about/overview/budget/budget23/justifications/i-rehab.pdf