The Impact of Personal Factors on Self-Determination and Early Adulthood Outcome Constructs in Youth With Disabilities

Abstract

Data from the National Longitudinal Transition Study–2 (NLTS2) were used to examine the impact of three personal factors—race/ethnicity, gender, and family income—on self-determination (i.e., autonomy, psychological empowerment, self-realization) and early adulthood outcome constructs. Findings suggest for those with high-incidence disabilities, family income predicts postsecondary education outcomes. And males with high-incidence disabilities have greater access to services and accommodations as adults, but lower use of financial supports (i.e., use of government support programs). African American youth across disability categories reported lower levels of financial independence. Females with intellectual disability reported greater social relationships but lower levels of financial support and employment. Implications for future research, policy, and practice are discussed.

Keywords

self-determination National Longitudinal Transition Study–2

Researchers in field of secondary transition have developed recommendations for increasing the quality of research in the transition field (Mazzotti, Rowe, Cameto, Test, & Morningstar, 2013) including increasing the inclusion of students from diverse backgrounds in research and systematically analyzing differential impacts of predictors of postschool success for diverse student groups in correlational research. Trainor, Lindstrom, Simon-Burroughs, Martin, and Sorrells (2008) highlight the need to understand that diversity and cultural identities are defined by multiple factors, including gender, racial/ethnic identity, disability, and socio-economic status and that research and intervention development must consider these factors, as well as systems-level (e.g., school policy and organization) factors that affect outcomes.

The importance of research with diverse groups is highlighted by research findings, using data from the National Longitudinal Transition Study–2 (NLTS2), suggesting that racially/ethnically diverse youth with disabilities are less likely to be employed competitively after high-school graduation, as are youth from lower socio-economic backgrounds (Wagner, Newman, Cameto, Garza, & Levine, 2005). Disability label also influences postschool employment. For example, young adults with intellectual disability have much lower rates of paid employment than other disability groups, particularly those with high-incidence disabilities (e.g., specific learning disability, other health impairments; Newman et al., 2011). Outside of employment, systematic differences based on gender, disability, race/ethnicity, and socio-economic status have been found in postsecondary access and community inclusion (Newman et al., 2011).

Researchers have also systematically reviewed transition research to identify predictors of positive postschool outcomes for youth and young adults with disabilities (Test et al., 2009). One factor that has been identified as having a significant and positive impact on postschool outcomes is promoting and enhancing self-determination (Dattilo & Rusch, 2012; Lachapelle et al., 2005; Mazzotti et al., 2013; Wehmeyer & Palmer, 2003; Wehmeyer & Schwartz, 1997). Shogren, Wehmeyer, Palmer, Rifenbark, and Little (2015) followed youth with disabilities who participated in a randomized control trial study of the impact of interventions to promote self-determination (Wehmeyer, Palmer, Shogren, Williams-Diehm, & Soukup, 2013). Youth who exited high school with higher levels of self-determination experienced more positive postschool community access and employment outcomes, although still less positive outcomes than their peers without disabilities. However, research also suggests that access to meaningful opportunities for the development of self-determination aligned with youth’s cultural beliefs and values is limited for diverse youth in secondary school (Leake & Boone, 2007; Shogren, 2012; Trainor, 2005). Researchers have found that race/ethnicity and disability label affect autonomy, psychological empowerment, and self-realization, three of four essential characteristics of self-determination (Shogren, Kennedy, Dowsett, Garnier Villarreal, & Little, 2014). For example, Hispanic youth tend to report lower levels of each of the constructs, and African American youth tend to report higher levels of autonomy than Hispanic youth. This body of research suggests the need for further research, as suggested by Mazzotti et al. (2013), that systematically explores impacts of personal factors on self-determination and early adult outcomes to inform policy and practice in secondary schools.

Purpose of the Study

Our purpose in the present study was to build on previous NLTS2 work and explore the impact of personal factors, namely, disability, race/ethnicity, general and family income on autonomy, psychological empowerment, and self-realization and postschool outcomes for youth with diverse disability labels. Our primary research question was as follows:

Research Question 1: Do race/ethnicity, gender, and family income predict differences in self-determination (i.e., autonomy, psychological empowerment, or self-realization) and postschool outcomes for students with high-incidence disabilities (i.e., specific learning disability, emotional disturbances, speech or language impairments, and other health impairments), intellectual disability, and cognitive disabilities (i.e., autism, multiple disabilities, and deaf-blindness)?

Method

Data Sources

We used data from NLTS2, building on previous analyses conducted to examine the impact of disability (Shogren, Kennedy, Dowsett, & Little, 2014) and race/ethnicity (Shogren, Kennedy, Dowsett, Garnier Villarreal, et al., 2014) on self-determination and postschool outcomes (Shogren, Shaw, & Little, 2016) with the intent of examining additional personal factors that had not been examined in previous research. NLTS2 data were collected over a 10-year period (2000–2009), with a sample of approximately 1,250 students within each federally recognized disability category (i.e., autism, deaf-blindness, emotional disturbance, hearing impairment, learning disability, intellectual disability, multiple disabilities, orthopedic impairment, other health impairment, speech and language impairment, traumatic brain injury, and visual impairment). Students were sampled from districts stratified by geographic region, size, and community wealth, and were between the ages of 13 and 16 at the start of data collection. Multiple data sources were collected over five waves of data collection (each wave represents a 2-year data collection period), one of which was the direct student assessment, which occurred once during Wave 1 or 2 while students were still in school for a subset of the sample (approximately 83%). The direct student assessment included a subset of 26 items from three of the four subscales of The Arc’s Self-Determination Scale (SDS; Wehmeyer & Kelchner, 1995). Shogren, Kennedy, Dowsett, and Little (2014) found that three (of four) essential characteristics of self-determination (autonomy, psychological empowerment, and self-realization) identified in the functional theory of self-determination (Wehmeyer, 2003) and measured on the SDS were adequately assessed in the NLTS2 data collection and latent constructs could be created. Shogren, Kennedy, Dowsett, and Little (2014) also found that the 12 disability groups could be collapsed into six groups based on similarities in their patterns of latent self-determination means and variances and disability-related characteristics, including a high-incidence disabilities group (specific learning disability, emotional disturbances, speech language impairments, and other health impairments), a sensory disabilities group (visual and hearing impairments), and a cognitive disabilities group (autism, multiple disabilities, and deaf-blindness) as well as three disability categories that could not be collapsed with any other group: intellectual disability, traumatic brain injury, and orthopedic impairments. The high-incidence, cognitive, and intellectual disability groups defined by Shogren, Kennedy, Dowsett, and Little (2014) were used in the present analyses as these were the largest disability groups and had the greatest diversity in outcome constructs, discussed subsequently.

The adult outcome constructs used in the present analyses were initially developed by Shogren et al. (2016). Shogren et al. (2016) used NLTS2 data collected during Wave 5 after young adults with disabilities exited school (ages 23–26) to develop and test 10 early adult outcome constructs organized around indicators of quality of life (social relationships, independent living, emotional well-being, access to services, health status, postsecondary education, financial supports [i.e., use of government support programs], financial independence, employment, and advocacy). To develop the constructs, indicators from the NTLS2 parent/youth survey conducted after youth exited school were reviewed, and potential indicators of the outcome constructs identified. The potential indicators were then subjected to empirical analysis using structural equation modeling, and constructs that had theoretical and empirical support (e.g., strong factor loadings, model fit, invariance across disability groups) were analyzed for differences across disability groups (Shogren et al., 2016). Table 1 provides an overview of the 10 adult outcome constructs. Shogren et al. (2016) found that students with high-incidence disabilities generally experienced more positive outcomes than students with intellectual, cognitive, or sensory disabilities. Shogren and Shaw (2016) then examined the degree to which three essential characteristics of self-determination (autonomy, psychological empowerment, and self-realization) predicted early adulthood outcomes and found that autonomy tended to be the strongest predictor of adult outcomes. There were differences based on disability label; students with more severe disabilities tended to have weaker relationships between self-determination and outcomes. However, beyond disability label, no research has examined the impact of other personal factors on these adult outcomes, which is the purpose of the present analysis.

Table 1.

Adult Outcome Constructs.

Adult outcome constructs	Brief description of NLTS2 data
1. Social relationships	Reported participation in community, volunteer, and group activities; frequency of invitations to social activities, talking on phone, engaging in social activities; feeling supported and cared about by friends and family
2. Independent living	Type and inclusiveness of current residential arrangement (e.g., independent or supported living arrangements vs. congregate or segregated settings)
3. Emotional well-being	Ratings of the degree to which young adults enjoy life are happy, feel good about themselves, and feel useful and able to get things done
4. Access to services	Rating of need for services beyond what is current available
5. Health status	Rating of general health status
6. Postsecondary education	Enrollment in any form of postsecondary education; duration and continuity of attendance; graduation status
7. Financial supports	Receives financial support from government sources, including SSI, food stamps, or any government program
8. Financial independence	Reports having checking, savings, and charge account
9. Employment	Employment status, duration and consistency of employment, number of hours worked, access to benefits, if promoted at current job, perceptions of treatment, compensation, and opportunities for advancement at current job
10. Advocating for needs	Reports communicating needed accommodations to employer

Source. Adapted from Shogren and Garnier Villarreal (2015); Shogren, Shaw, and Little (2016).

Note. NLTS2 = National Longitudinal Transition Study–2; SSI = Supplemental Security Income.

Sample

It is important to note there were differences in the sample used in the present analysis, compared with Shogren, Kennedy, Dowsett, and Little (2014) and Shogren et al. (2016). In the present analysis, we narrowed the NLTS2 sample to represent the three largest racial/ethnic categories, African American, Hispanic/Latino, and White to ensure sufficient power to analyze differences based on racial/ethnic group. Relatedly, we only targeted three (of six) disability groups from Shogren, Kennedy, Dowsett, and Little (2014) and Shogren et al. (2016) with the largest sample size when crossed with the racial/ethnic groups: high incidence (HIN; specific learning disability, emotional disturbances, speech or language impairments, and other health impairments), cognitive disabilities (COG; autism, multiple disabilities, and deaf-blindness), and intellectual disability (INT) groups. Participants who indicated “White” as their racial/ethnic group on the survey were used as the reference group in comparisons with respondents who indicated African American or Hispanic/Latino (w2_Eth6). Females served as the reference group for gender (w2_Gend2). Family income, a variable with 16 categories in increments of US$5,000, was included as a continuous predictor (np1K15Detail). In preparation for examining the research question, dummy variables were created for race/ethnicity and gender.

Missing Data

Because several of the adult outcome constructs used in the present analyses only applied to subgroups (e.g., employment questions were asked only of respondents who were employed), full information maximum likelihood (FIML) was used as the method to handle missing data. FIML generates estimates based only on what questions were answered rather than dropping incomplete records (e.g., list-wise deletion or imputing values as in multiple imputation; Enders, 2010). Under an assumption of missing at random (MAR) data (i.e., wherein the propensity to respond is entirely predicted by the observed portions of the data), FIML will produce unbiased and optimally efficient parameter estimates (Enders, 2010; Schafer & Graham, 2002).

Analysis

As mentioned previously, the present study built on previous NLTS2 analyses, and the starting point for the present analyses was the final model from the Shogren et al. (2016) including constructs representing the three self-determination constructs (autonomy, self-realization, and psychological empowerment) and the 10 early adult outcome domains (see Table 1). As described in the “Sample” section of this article, the sample differed slightly in the present analysis, necessitating invariance testing of the three self-determination and the 10 quality of life outcome constructs. After invariance testing, the additional variables related to race/ethnicity, gender, and family income were added to the model to explore our primary research question. All models were estimated using Mplus 7.0. Survey weights (variable “wt_na”), stratum, and cluster information was included by specifying complex analysis for the use of the robust maximum likelihood estimator (MLR). Multiple fit indices were used to evaluate overall model fit, including the root mean square error of approximation (RMSEA), the comparative fit index (CFI), and non-normed fit index (NNFI/Tucker–Lewis index [TLI]). RMSEA evaluates misfit in the covariances between the observed data and the model estimates (Brown, 2015).

Invariance constraints were tested, and change in CFI was used to establish measurement invariance (Cheung & Rensvold, 2002). Nested model testing with χ² difference testing was used to evaluate differences in latent constructs and covariates. More specifically, each latent parameter was constrained to equality between two groups at a time, and model fit was then compared with the model without that constraint; if the χ² difference was significant for one degree of freedom, the null hypothesis that the models were the same was rejected. Nested model testing used log likelihood values with the scaling factor correction to obtain corrected χ² values for the MLR estimator (Satorra, 2000). A cutoff of p < .01 was set a priori rather than p < .05 due to the large number of constructs in the model. All model constructs were evaluated for group differences at both the measurement and latent levels, and then the impact of race/ethnicity, gender, and income for the constructs was examined.

Results

Measurement Invariance

Invariance testing results for the self-determination and adult outcome constructs across the three disability groups (high incidence, intellectual disability, and cognitive disability) who reported their race/ethnicity as African American, Hispanic/Latino, and White were similar to those found in Shogren et al. (2016) with the larger sample of students. Specifically, model fit for the three self-determination constructs and 10 adult outcome constructs was adequate (see Table 2). Weak invariance was established, in which factor loadings were equated across groups with a change in CFI = .005. When testing strong invariance (i.e., indicator intercepts equated across groups), similar to Shogren et al., two indicators were identified through nested model testing to have different intercept levels across the groups, one from the employment construct, and one from the social relationships construct. When these two indicators were constrained to equality for the intellectual and cognitive disability groups but freed to vary from high-incidence disability estimates, partial invariance for the intercept model (strong invariance) was established with as few changes as possible made to the intercept constraints. Because these indicators were freed from the high-incidence disability group, latent mean comparisons for employment and social relationships can only be made between the intellectual and cognitive disability groups.

Table 2.

Measurement Invariance Testing Results.

Invariance tests	χ²	df	RMSEA	90% CI	CFI	NNFI
Configural	2,947.17	1,461	0.030	[0.028, 0.031]	0.833	0.796
Loadings	3,036.25	1,505	0.030	[0.028, 0.031]	0.828	0.796
Intercepts	3,300.84	1,549	0.031	[0.030, 0.033]	0.803	0.773
Intercepts (partial)^a	3,178.09	1,547	0.030	[0.029, 0.032]	0.816	0.788

Note. RMSEA = root mean square error of approximation; CI = confidence interval; CFI = comparative fit index; NNFI = non-normed fit index.

Parceled indicator (emotw) in employment constrained to equality intellectual and cognitive disability groups, and parceled indicator (srirs) in social relationships constrained to equality for those same two groups.

Latent Differences

Differences in the latent factors (i.e., variances, correlations, and means) for the self-determination and adult outcome constructs across the three disability groups (high incidence, intellectual disability, and cognitive disability) who reported their race/ethnicity as African American, Hispanic/Latino, and White were then evaluated with nested model comparisons against the partial intercept model identified in the measurement invariance process. Most differences in latent factors were similar or identical to those reported in Shogren et al. (2016), although there were some differences that emerged based on the differing samples used in the two analyses, suggesting the influence of disability label and race/ethnicity on the pattern of results. As shown in Table 3, in the present analyses, approximately one third of the latent variance estimates were identified as differing across groups with most of the differences appearing between the high-incidence group and one of the other groups. Generally, those with high-incidence disabilities tended to show higher variability in autonomy, financial independence, and advocacy, and lower variability in independent living outcomes and postsecondary outcomes. Participants with intellectual disability showed wider variability in living independently and less variability in postsecondary education outcomes when compared with participants in the cognitive disability group. As shown in Table 4, there were limited differences in latent correlations across the groups with only nine of 105 (9%) comparisons identifying group differences. As with latent variances, the differences were concentrated in comparisons of the high-incidence disability group and one of the other groups. Furthermore, all but one group difference was between high-incidence and cognitive disability groups, with students with high-incidence disabilities generally having higher correlations between constructs, suggesting stronger relationships between constructs in this population.

Table 3.

Latent Variance Differences Between Groups.

Construct	Disability Group 1	Disability Group 2	Variance Group 1	Variance Group 2	Ratio	Δχ^2a
Autonomy	HIN	COG	1.000	1.713	1.713	20.51
Financial support	HIN	INT	1.000	2.837	2.837	10.50
	HIN	COG	1.000	3.070	3.070	18.00
Advocacy	HIN	INT	0.674	1.013	1.503	141.87
	HIN	COG	0.674	0.953	1.414	108.56
Living situation	HIN	COG	0.500	0.364	0.728	15.39
	INT	COG	0.457	0.364	0.796	8.41
Postsecondary education	HIN	INT	0.341	0.201	0.589	23.82
	INT	COG	0.201	0.330	1.642	21.15

Note. HIN = High Incidence Disability Group; COG = Cognitive Disability Group; INT = Intellectual Disability Group.

All Δχ² values for the nested model tests were evaluated with 1 degree of freedom and found to have p < .005.

Table 4.

Latent Correlation Differences Between Groups.

Correlation	Disability Group 1	Disability Group 2	Group 1 correlation	Group 2 correlation	Differences
Autonomy–Financial independence	HIN	INT	.353	−.034	.387
	HIN	COG	.353	−.130	.483
Autonomy–Employment	HIN	COG	.304	−.055	.359
Autonomy–Postsecondary education	HIN	COG	.234	.020	.214
Employment–Emotional well-being	HIN	COG	.539	.179	.360
Self-realization–Advocacy	HIN	COG	−.106	−.404	.298

Note. The Difference column is computed by Group 1 Correlation–Group 2 Correlation; HIN = High Incidence Disability Group; INT = Intellectual Disability Group; COG = Cognitive Disability Group.

Finally, when examining differences in latent means via nested model comparison for 26 latent mean pairs, the three disability groups showed significant differences from each other in specific outcome constructs, similar to Shogren et al. (2016), although differences continued to emerge that were unique to the present sample. More than half of the latent mean comparisons (60%) showed differences. Based on Cohen’s D, effects sizes for the differences were calculated based on this formula and are reported in Table 5:

d = \frac{(α_{1} - α_{2})}{\sqrt{φ p o o l e d}} .

Table 5.

Latent Mean Differences Between Groups.

	Disability Group 1	Disability Group 2	Mean Group 1	Mean Group 2	Cohen’s D
Autonomy	INT	COG	0.178	−0.317	0.242
Self-realization	HIN	COG	0.000	−0.309	0.252
Psychological empowerment	HIN	INT	0.000	−0.692	0.593
	HIN	COG	0.000	−0.974	0.757
Financial independence	HIN	INT	0.000	−0.916	0.847
	HIN	COG	0.000	−0.412	0.360
	INT	COG	−0.916	−0.412	−0.343
Financial support	HIN	INT	0.000	1.469	−1.150
	HIN	COG	0.000	1.932	−1.314
Emotional well-being	HIN	INT	0.000	−0.394	0.353
	HIN	COG	0.000	−0.432	0.368
Advocacy	HIN	INT	0.324	1.118	−0.837
	HIN	COG	0.324	1.337	−0.995
Independent living^a	INT	COG	0.303	0.150	0.147
Services and accommodations	HIN	COG	0.342	0.520	−0.219
Postsecondary education	HIN	INT	0.326	0.083	0.391
	HIN	COG	0.326	0.246	0.117
	INT	COG	0.083	0.246	−0.187

Note. INT = Intellectual Disability Group; COG = Cognitive Disability Group; HIN = High Incidence Disability Group.

Only the test comparing Independent Living between INT and COG was reported because all indicator intercepts were constrained to equality for these two groups.

The largest mean differences were found between those with high-incidence disability and the other groups. For example, young adults with high-incidence disabilities tended to report greater financial independence but lower financial supports (i.e., use of government programs). Students with high-incidence disabilities also tended to report greater psychological empowerment and emotional well-being, as well as greater independent living and postsecondary education outcomes. Students with high-incidence disabilities did, however, report engaging in lower levels of advocacy and having less access to accommodations. Students with intellectual and cognitive disabilities tended to score more similarly, with the exception of those with intellectual disability reporting higher autonomy and independent living outcomes, and lower financial independence and access to postsecondary education.

The differences in the financial support construct between the high-incidence and intellectual disability groups in the present analysis (−1.15) and between the high-incidence and cognitive disability groups (−1.31) were much larger than that found in the Shogren et al. (2016) sample, which included all disability groups and race/ethnicities, again suggesting sample-related differences influenced by disability and race/ethnicity. Furthermore, statistically significant mean differences disappeared for employment and when comparing the high-incidence disability group with the other groups on independent living in the present analyses.

Impact of Race/Ethnicity, Gender, and Family Income

Table 6 provides an overview of the key findings, and the significant findings are discussed in the following section for each personal factor.

Table 6.

Role of Race/Ethnicity, Gender, and Income on Latent Outcomes.

Constructs	African American		Hispanic/Latino		Male		Income
Constructs	Estimate	SE	Estimate	SE	Estimate	SE	Estimate	SE
High-incidence disability
Financial independence	−0.300*	0.095					0.284*	0.063
Financial support					−0.253*	0.079
Postsecondary education	0.157*	0.049					0.170*	0.057
Access to services					0.140	0.056
Intellectual disability
Financial independence	−0.304*	0.097					0.235	0.114
Financial support			0.589	0.334	1.361*	0.522
Employment			0.578	0.321	1.447*	0.518
Social relationships			−0.319	0.189	−0.608	0.338
Cognitive disability
Financial independence	−0.354*	0.085					0.338*	0.073
Employment					0.188*	0.072
Social relationships			−0.179	0.072
Advocating for needs	−0.222*	0.074			0.177	0.072

significant at p < .01.

Race/ethnicity

When compared with young adults across the disability groups (HIN, INT, COG) who reported their race/ethnicity as White, participants who reported their race/ethnicity as African American reported lower levels of financial independence and lower advocacy levels in the cognitive disability group. Access to postsecondary education was higher in African American respondents in the high-incidence disability group than in the intellectual and cognitive disability groups. No differences were identified within in the high-incidence disability group between Hispanic/Latino and White respondents. Social relationship outcomes were lower for respondents who were Hispanic/Latino within the intellectual and cognitive disability groups as compared with respondents who were White. Hispanic/Latino respondents with intellectual disability reported higher levels of financial support and employment than White respondents.

Gender

No consistent patterns were identified based on gender. Males in the high-incidence disability group reported lower levels of financial support than females but higher levels of services and accommodations. In the intellectual disability group, males reported higher levels of financial support and employment and lower levels of social relationships when compared with females. Higher levels of employment and advocacy were reported by males in the cognitive disability group when compared with females in that same group.

Income

Across all three disability groups, income was a positive predictor of financial independence suggesting its impact for students with high- and low-incidence disabilities. Income also predicted postsecondary education for participants in the high-incidence disability group.

Discussion

The purpose of the present analysis was to build on previous NLTS2 research and to broadly examine the impact of three personal factors—race/ethnicity, gender, and family income—that had not been examined systematically in previous research on self-determination and early adulthood outcomes. The intent was to inform policy and practice related to secondary transition services and supports, as well as provide direction for future research. By understanding the personal factors that shape one’s cultural identity (Trainor et al., 2008) and influence self-determination and adult outcomes, considerations that should be embedded into individualized transition services and supports can be identified, implemented, and further researched. Specifically, by better understanding these factors, interventions and supports can be individualized based on cultural identities, as a limited body of research has done (Valenzuela & Martin, 2005).

Implications for Research and Practice

Consistent with previous work, the findings suggest that for students with high-incidence, intellectual, and cognitive (i.e., autism, multiple disabilities, and deaf-blindness) disabilities, ongoing work is needed to enhance self-determination and to promote valued postschool outcomes. Specific attention should be directed to examining the implementation and outcomes of interventions to promote self-determination in secondary schools, particularly across racial/ethnic groups. The present analysis confirms differences in self-determination constructs and early adult outcomes based on membership in differing disability and racial/ethnic groups, and research on the factors that shape these differences to inform practice is needed.

The present findings suggest that, generally—across diverse youth—those with high-incidence disabilities tend to report more positive outcomes, with greater access to postschool employment and independent living opportunities. This group shows greater variability in their outcomes, suggesting that a greater range of postschool options are experienced by youth with high-incidence disabilities. The greater range of experiences available to youth with high-incidence disabilities suggests the criticality of policy and practice that promote access to and experiences in integrated employment and living options for secondary youth with more severe disabilities. The importance of these experiences is confirmed by other national data (Hart, Grigal, & Weir, 2010; Siperstein, Parker, & Drascher, 2013) that suggest those with intellectual and cognitive disability tend to have lower rates of employment, independent living, and postsecondary education, and there is less variability in outcomes. Evidence-based practices for promoting transition into employment and postsecondary education exist that can be implemented in secondary schools (Test et al., 2009), and the present findings suggest that such practices need to be available and accessible to all students to enhance outcomes.

Interestingly, those with intellectual and cognitive disability reported higher levels of financial support, but lower levels of financial independence. This confirms the frequently discussed limitations of government benefits related to employment and the creation of financial capital (Soffer, McDonald, & Blanck, 2010), and suggests the need for ongoing focus on creating opportunities for competitive employment for people with intellectual and cognitive disabilities and promoting school-based employment opportunities and experiences that are linked with positive postschool employment opportunities for youth with intellectual and cognitive disabilities (Test et al., 2009).

Beyond disability-related findings, a major contribution of the present analysis was extending knowledge of the impact of other personal characteristics on self-determination and postschool outcomes. Although a limitation of the present analysis is only analyzing three additional factors—race/ethnicity, gender, and family income—which do not capture all the diverse factors that may shape one’s personal culture (Trainor et al., 2008), the results do provide preliminary information to guide future research, policy, and practice. For example, as might be expected from other research (Sima, Wehman, Chan, West, & Leucking, 2015; Stoneman, 2007; Wang et al., 2004), family income played a role, irrespective of disability label, on financial independence. Furthermore, for those with high-incidence disabilities, but not low-incidence disabilities such as cognitive and intellectual disability, family income also predicted postsecondary education outcomes. The data did not allow for a systematic analysis of the factors that shaped this relationship (e.g., greater ability to provide financial support—outside of government programs—for college or greater ability to access supports for accommodations), but they suggest that above and beyond financial support through government programs, family resources are a stronger predictor of access to and success in high education. Further research is needed to more systematically examine these findings and alternative asset-building programs (Soffer et al., 2010), as well as to explore, given the increase in the prevalence of postsecondary programs for people with intellectual and cognitive disability (Hart et al., 2010), the impact of family and government resources on access to these programs. Considering ways that access to and knowledge of these programs can be developed in secondary education and transition services has the potential to enhance practice and the ability of youth to transition into adulthood with the financial supports in place that enhance valued postschool outcomes.

In addition to family income, there were also associations between a person’s reported race/ethnicity and postschool outcomes. Shogren, Kennedy, Dowsett, Garnier Villarreal, et al. (2014) identified mean-level differences in self-determination constructs based on race/ethnicity, and the present analysis adds to those findings by providing additional details on the impact of race/ethnicity on postschool outcomes. For example, the present analyses found African American young adults reported lower levels of financial independence, consistent with other research (Fujiura & Yamaki, 1997; Fujiura, Yamaki, & Czechowicz, 1998). Work is also needed to develop and evaluate interventions in secondary schools that target enhanced self-determination and financial resource development for diverse students and families. Hispanic/Latino youth in the intellectual and cognitive disability groups tended to report fewer social relationships, but high levels of financial support, highlighting the importance of building not only paid but also natural supports for this population in adulthood. Research is also needed examining how various peer support and other relationship-focused interventions (Carter, Cushing, Clark, & Kennedy, 2005) in secondary schools can use culturally responsive practices to build social capital in schools and communities to promote valued self-determination and postschool outcomes.

As mentioned previously, there were no consistent gender differences; but males did tend to report, in the high-incidence disability group, greater access to services and accommodations and lower use of financial supports, perhaps indicating any ability to access accommodations in postsecondary and work environments independent of paid services and supports. For those with intellectual disability, females reported greater numbers of social relationships, but lower levels of financial support and employment, suggesting stronger social supports but less access to employment opportunities (Olson, Cioffi, Yovanoff, & Mank, 2000). Employment was also more frequent for males with cognitive disabilities, again suggesting ongoing gender-based disparities. Further work is needed to explore the degree to which these outcomes are shaped by gender-related stereotypes or planning practices that may be reinforced in schools, and can be targeted with policy and practice that directly address gender. This also highlights the need for future research to jointly examine personal and environmental factors.

Overall, the findings suggest that there is a relationship between a variety of personal factors (i.e., disability, race/ethnicity, family income, gender) and outcomes, and emphasize that looking at any one of these factors in isolation will not provide a full picture of the complex contextual factors that affect outcomes (Shogren, Luckasson, & Schalock, 2014). Furthermore, this work looked at a restricted range of personal factors, certainly not all factors that can define one’s personal culture, and did not examine the interactive role of the environment in shaping outcomes. Future work is critically needed applying an ecological perspective (Bronfenbrenner, 1979, 2005) to research on the development of interventions and policy to support implementation that addresses these factors concurrently and comprehensively. Further research is also needed to more specifically analyze the patterns of relationships of personal factors with specific outcome constructs (e.g., employment, postsecondary education), to further understand the unique influences within each construct. However, the findings continue to confirm that there are disability-related disparities in outcomes for youth and young adults with disabilities, as well as disparities influenced by race/ethnicity, gender, and family income that must be addressed through culturally responsive practices. Again, however, these disparities are simultaneously influenced by the multiple factors that define one’s personal culture and differ across outcome domains. Given this, work is also needed to explore protective factors as well as risk factors. The findings highlight that, particularly for those with high-incidence disabilities compared with those with intellectual and cognitive disabilities, there is a large amount of variability in outcomes (although this variability is still skewed to more youth having less than optimal outcomes), and that in some circumstances, certain groups, such as African American youth with high-incidence disabilities, are actually reporting greater access to postsecondary environments, although again, this may be an artifact of the sample and the population that was retained in the NLTS2 sample over time. Perhaps, however, youth who are provided with access to these opportunities are more likely to have generalized benefits. Identifying protective factors that enable youth from diverse backgrounds to use their self-determination to shape the adult outcomes they desire is a critical area for future research to enable practitioners to build on these protective factors and address risk factors (Trainor, 2008; Trainor et al., 2008).

Limitations and Directions for Future Research, Policy, and Practice

The same limitations that characterize all secondary data analysis of NLTS2 and other large national data sets apply to the present analyses (Mazzotti et al., in press; Shogren & Shaw, 2016; Shogren et al., 2016). First, using survey data to define latent constructs on a post hoc basis, when the variables were not originally selected to be representative of latent constructs, is a challenge and limits the ability to include all constructs that might be of interest or included in a given theoretical perspective. As described in Shogren et al. (2016) with regard to the adult outcome constructs and Shogren, Kennedy, Dowsett, and Little (2014) with regard to the self-determination constructs included in the present analysis, the full range of quality of life domains and essential characteristics of self-determination could not be represented using the questions in NLTS2. Although efforts were made to select the most representative variables to define latent constructs, the included constructs do not fully meet the definitional criteria present in the literature for the constructs. Furthermore, the sample included in the present analysis was restricted to those who were able to participate in the direct student assessment (approximately 83% of NLTS2 sample) and those who reported race/ethnicity of White, African American, or Hispanic/Latino and were classified as having a high-incidence, intellectual, or cognitive disability based on previous work with the self-determination constructs. As described in the “Results” section, this more restricted sample led to slight differences in the outcomes particularly when testing latent differences compared with Shogren, Kennedy, Dowsett, and Little (2014) where all disability categories and race/ethnicities included in the NLTS2 sample were part of the analyses. Furthermore, the disability grouping and the terminology used to describe these groupings (e.g., cognitive disability) were defined in previous analysis. Alternate groupings or descriptions of the groupings might be possible and were not tested in this study. Thus, the sample and the grouping of the sample must be considered in interpreting the analyses. Overall, the findings confirm the impact of sampling for specific characteristics on findings and the need for research with other disability and racial/ethnic groups as very limited research has examined factors that influence outcomes across disability groups in other racial or ethnic groups (Leake & Boone, 2007). Specifically, sampling plans in large policy studies must be developed to allow for sub-analyses of small disability and racial/ethnic groups, as this was not possible in the present analysis because of the small nature of some of the additional racial/ethnic groups represented in the sample (e.g., Asian American, Native American/Pacific Islander). In addition, the fact that the start of NLTS2 data collection was more than 15 years ago (2000) must be considered in interpreting these findings, and as there have been multiple changes in society and in schools since the early 2000s when data collection began, updated data are needed to further explore current student experiences.

Each of these limitations, however, confirms the need for further research exploring the best ways to understand the diverse personal and environmental factors that affect outcomes. Although the focus in the present analyses is on looking, broadly, at the pattern of relationships between personal factors, self-determination, and adult outcome constructs, future work is needed that more specifically analyzes the pattern of relationships between specific factors and outcome areas. Only then can frameworks for meaningfully using this information in policy and practice be developed to guide transition planning that is individualized to the needs of each adolescent with a disability to meaningfully affect outcomes.

Footnotes

Authors’ Note

The opinions expressed are those of the authors and do not represent views of the Institute or the U.S. Department of Education.

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: The research reported here was supported by the Institute of Education Sciences, U.S. Department of Education, through Grant R324A110041 to the University of Kansas.

References

Bronfenbrenner

(1979). The ecology of human development: Experiments by nature and design. Cambridge, MA: Harvard University Press.

Bronfenbrenner

(Ed.). (2005). Making human beings human: Bioecological perspectives on human development. Thousand Oaks, CA: SAGE.

Brown

T. A.

(2015). Confirmatory factor analysis for applied research (2nd ed.). New York, NY: Guilford Press.

Carter

E. W.

Cushing

L. S.

Clark

N. M.

Kennedy

C. H.

(2005). Effects of peer support interventions on students’ access to the general curriculum and social interactions. Research and Practice for Persons With Severe Disabilities, 30, 15–25.

Cheung

G. W.

Rensvold

R. B.

(2002). Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling, 9, 233–255.

Dattilo

Rusch

(2012). Teaching problem solving to promote self-determined leisure engagement. Therapeutic Recreation Journal, 46, 91–105.

Enders

C. K.

(2010). Applied missing data analysis. New York, NY: Guilford Press.

Fujiura

G. T.

Yamaki

(1997). Analysis of ethnic variations in developmental disability prevalence and household economic status. Mental Retardation, 35, 286–294.

Fujiura

G. T.

Yamaki

Czechowicz

(1998). Disability among ethnic and racial minorities in the United States: A summary of economic status and family structure. Journal of Disability Policy Studies, 9, 111–130. doi:10.1177/104420739800900207

10.

Hart

Grigal

Weir

(2010). Expanding the paradigm: Postsecondary education options for individuals with autism spectrum disorders and intellectual disabilities. Focus on Autism and Other Developmental Disabilities, 25, 134–150.

11.

Lachapelle

Wehmeyer

M. L.

Haelewyck

M. C.

Courbois

Keith

K. D.

Schalock

. . .Walsh

P. N.

(2005). The relationship between quality of life and self-determination: An international study. Journal of Intellectual Disability Research, 49, 740–744.

12.

Leake

Boone

(2007). Multicultural perspectives on self-determination from youth, parent, and teacher focus groups. Career Development for Exceptional Individuals, 30, 104–115.

13.

Mazzotti

V. L.

Rowe

D. A.

Cameto

Test

D. W.

Morningstar

M. E.

(2013). Identifying and promoting transition evidence-based practices and predictors of success: A position paper of the Division on Career Development and Transition. Career Development and Transition for Exceptional Individuals, 36, 140–151. doi:10.1177/2165143413503365

14.

Mazzotti

V. L.

Rowe

D. A.

Sinclair

Poppen

Woods

W. E.

Shearer

M. L.

(in press). Predictors of post-school success: A systematic review of NLTS2 secondary analyses. Career Development and Transition for Exceptional Individuals.

15.

Newman

Wagner

Knokey

A.-M.

Marder

Nagle

Shaver

. . .Schwarting

(2011). The post-high school outcomes of young adults with disabilities up to 8 years after high school (A report from the National Longitudinal Transition Study-2 [NLTS2; NCSER 2011-3005]). Menlo Park, CA: SRI International.

16.

Olson

Cioffi

Yovanoff

Mank

(2000). Gender differences in supported employment. Mental Retardation, 38, 89–96.

17.

Satorra

(2000). Scaled and adjusted restricted tests in multi-sample analysis of moment structures. In Heijmans

R. D. H.

Pollock

D. S. G.

Satorra

(Eds.), Innovations in multivariate statistical analysis (pp. 233–247). London, England: Kluwer Academic.

18.

Schafer

J. L.

Graham

J. W.

(2002). Missing data: Our view of state of the art. Psychological Methods, 7, 147–177.

19.

Shogren

K. A.

(2012). Hispanic mothers’ perceptions of self-determination. Research and Practice for Persons With Severe Disabilities, 37, 170–184.

20.

Shogren

K. A.

Garnier Villarreal

(2015). Developing student, family, and school constructs from NTLS2 data. Journal of Special Education, 49, 89–103. doi:10.1177/0022466913513336

21.

Shogren

K. A.

Kennedy

Dowsett

Garnier Villarreal

Little

T. D.

(2014). Exploring essential characteristics of self-determination for diverse students using data from NLTS2. Career Development and Transition for Exceptional Individuals, 37, 168–176.

22.

Shogren

K. A.

Kennedy

Dowsett

Little

T. D.

(2014). Autonomy, psychological empowerment, and self-realization: Exploring data on self-determination from NLTS2. Exceptional Children, 80, 221–235.

23.

Shogren

K. A.

Luckasson

Schalock

R. L.

(2014). The definition of context and its application in the field of intellectual disability. Journal of Policy and Practice in Intellectual Disabilities, 11, 109–116.

24.

Shogren

K. A.

Shaw

L. A.

(2016). The role of autonomy, self-realization, and psychological empowerment in predicting early adulthood outcomes for youth with disabilities. Remedial and Special Education, 37, 55–62. doi:10.1177/0741932515585003

25.

Shogren

K. A.

Shaw

L. A.

Little

T. D.

(2016). Measuring the early adulthood outcomes of young adults with disabilities: Developing constructs using NLTS2 data. Exceptionality, 24, 45–61. doi:10.1080/09362835.2015.1064416

26.

Shogren

K. A.

Wehmeyer

M. L.

Palmer

S. B.

Rifenbark

G. G.

Little

T. D.

(2015). Relationships between self-determination and postschool outcomes for youth with disabilities. Journal of Special Education, 53, 30–41. doi:10.1177/0022466913489733

27.

Sima

A. P.

Wehman

P. H.

Chan

West

M. D.

Leucking

R. G.

(2015). An evaluation of risk factors related to employment outcomes for youth with disabilities. Career Development and Transition for Exceptional Individuals, 38, 89–100.

28.

Siperstein

G. N.

Parker

R. C.

Drascher

(2013). National snapshot of adults with intellectual disabilities in the labor force. Journal of Vocational Rehabilitation, 39, 157–165.

29.

Soffer

McDonald

K. E.

Blanck

(2010). Poverty among adults with disabilities: Barriers to promoting asset accumulation in individual development accounts. American Journal of Community Psychology, 46, 376–385. doi:10.1007/s10464-010-9355-4

30.

Stoneman

(2007). Examining the Down syndrome advantage: Mothers and fathers of young children with disabilities. Journal of Intellectual Disability Research, 51, 1006–1017.

31.

Test

D. W.

Mazzotti

V. L.

Mustian

A. L.

Fowler

C. H.

Kortering

Kohler

(2009). Evidence-based secondary transition predictors for improving postschool outcomes for students with disabilities. Career Development for Exceptional Individuals, 32, 160–181. doi:10.1177/0885728809346960

32.

Trainor

A. A.

(2005). Self-determination perceptions and behaviors of diverse students with LD during the transition planning process. Journal of Learning Disabilities, 38, 233–248.

33.

Trainor

A. A.

(2008). Using cultural and social capital to improve postsecondary outcomes and expand transition models for youth with disabilities. The Journal of Special Education, 42, 148–162.

34.

Trainor

A. A.

Lindstrom

Simon-Burroughs

Martin

J. E.

Sorrells

A. M.

(2008). From marginalized to maximized opportunities for diverse youths with disabilities: A position paper of the Division on Career Development and Transition. Career Development for Exceptional Individuals, 31, 56–64.

35.

Valenzuela

R. L.

Martin

J. E.

(2005). Self-directed IEP: Bridging values of diverse cultures and secondary education. Career Development for Exceptional Individuals, 28, 4–14.

36.

Wagner

Newman

Cameto

Garza

Levine

(2005). After high school: A first look at the postschool experiences of youth with disabilities (A report from the National Longitudinal Transition Study–2 [NLTS-2]). Menlo Park, CA: SRI International.

37.

Wang

Turnbull

A. P.

Summers

J. A.

Little

T. D.

Poston

D. J.

Mannan

Turnbull

(2004). Severity of disability and income as predictors of parents’ satisfaction with their family quality of life during early childhood years. Research and Practice for Persons With Severe Disabilities, 29, 82–94. doi:10.2511/rpsd.29.2.82

38.

Wehmeyer

M. L.

(2003). A functional theory of self-determination: Model overview. In Wehmeyer

M. L.

Abery

Mithaug

D. E.

Stancliffe

(Eds.), Theory in self-determination: Foundations for educational practice (pp. 182–201). Springfield, IL: Charles C. Thomas Publishing.

39.

Wehmeyer

M. L.

Kelchner

(1995). The Arc’s self-determination scale. Arlington, TX: The Arc National Headquarters.

40.

Wehmeyer

M. L.

Palmer

S. B.

(2003). Adult outcomes for students with cognitive disabilities three-years after high school: The impact of self-determination. Education and Training in Developmental Disabilities, 38, 131–144.

41.

Wehmeyer

M. L.

Palmer

S. B.

Shogren

K. A.

Williams-Diehm

Soukup

J. H.

(2013). Establishing a causal relationship between interventions to promote self-determination and enhanced student self-determination. Journal of Special Education, 46, 195–210. doi:10.1177/0022466910392377

42.

Wehmeyer

M. L.

Schwartz

(1997). Self-determination and positive adult outcomes: A follow-up study of youth with mental retardation or learning disabilities. Exceptional Children, 63, 245–255.