top of page

Hiring Agents' Beliefs: A Barrier to Employment of Autistics

A study published in the summer of 2019 by Angela Marie Mai at Walden University in Minneapolis found that:

Capable, qualified, and working-age (18-65) autistic adults experienced an 83% unemployment rate in the United States in 2017 resulting in extreme poverty and severely decreased quality of life.

Research dating from 1957 to 2016 inferred hiring agents’ beliefs were the cause.

In this multiple regression study, the nature of the relationship between hiring agents’ beliefs and their selection of qualified autistic candidates was explored through Ajzen’s theory of planned behavior to determine what hiring agent’s beliefs, if any, influence the selection of qualified autistic candidates to fill open positions.

I used the Hiring Agent Survey Regarding Selection of Qualified Autistic Candidates to anonymously gather data from hiring agents throughout the contiguous United States.

Known values of the independent variable, the beliefs influencing hiring agents, were summed and clustered against the TPB-predicted, percentage-based, continuous-level dependent variable, which was hiring agents’ selection of qualified autistic candidates.

This statistically significant regression analysis, F(45, 73) = 36.067, p < .001, adj. R2 = .930, predicted the degree to which each control, normative, and behavioral belief influenced hiring agents’ selection.

Hiring agents’ desire for mandated comprehensive organizational diversity, along with their negative stereotypical associations and fear of embarrassment, signify a need for substantive policy and strategic interventions.

Results of such aggressive diversity initiatives could considerably improve the nation’s socioeconomic health and substantively increase autistics’ quality of life.


There is an unemployment crisis among autistic adults in the United States. At the end of 2017, the calculated U.S. unemployment for qualified autistic candidates was 83% (Mai, 2018), while the unemployment rate for the general United States population was 4% (U.S. Department of Labor [DOL] Bureau of Labor and Statistics [BLS], 2018).

A wide variety of equal opportunity public policies, such as the Americans with Disabilities Act of 1990 (ADA), ADA Amendments Act of 2008 (ADAAA; Equal Employment Opportunity Commission [EEOC], 2008), EEOC (2009), Equality Act 2010 (EqA), Ontarians with Disabilities Act (OWDA), and Persons with Disabilities Act of 2008 (PWD), have yet to achieve equal employment opportunities for autistic individuals as demonstrated by DOL and other scholarly reports.

At the end of 2010, roughly 3 million working-age, qualified, skilled, and capable autistics in the United States were unemployed (Christensen et al., 2016; Mai, 2018; Zablotsky, Black, Maenner, Schieve, & Blumberg, 2015). Autistics live in poverty conditions (i.e., 27% below the 2007 U.S. poverty line; Cimera, 2010; U.S. Office of the Assistant Secretary for Planning and Evaluation, 2014) due to severe unemployment issues (i.e., 83% employment rate) and a lack of state support services (Carpenter, 2015; Cheak-Zamora, Teti, Peters, & Maurer-Batjer, 2017; Fournier, 2016; Gentleman, 2011).

As Cimera’s, solo and joint extensive study, from 1996 through 2016 (Burgess & Cimera, 2014; Cimera, 1996, 2000, 2002, 2006a, 2006b, 2007a, 2007b, 2008a, 2008b, 2009a, 2009b, 2010, 2011a, 2011b, 2012a, 2012b, 2014a, 2014b; Cimera & Burgess, 2011; Cimera, Burgess, & Bedesem, 2014; Cimera, Burgess, & Wiley, 2013; Cimera & Cowan, 2009; Cimera & Oswald, 2009; Cimera & Rumrill, 2008; Inge, Cimera, Revell, Wehman, & Seward, 2015; Inge, Cimera, Rumrill, & Revell, 2016; Kregal, Wehman, Revell, Hill, & Cimera, 2012) demonstrated, even though increased income results in decreased social service benefits, competitive employment increases income.

In 2007, for every $1.00 of support services provided, autistics earned $5.28 (Cimera, 2010) and the U. S. Social Security Administration (SSA; 2009) report a ratio of $7.00 for every dollar. Employed autistics would not be as reliant on social services to provide income and medical services (Burgess & Cimera, 2014; Council of State Administrators of Vocational Rehabilitation [CSAVR], 2011; Rusch & Cimera, 1996). Autistics desire to work (Agovino, Paroki, & Sánchez Barajas, 2014; Anderson, McDonald, Edsall, Smith, & Taylor, 2015; Wik & Tøssebro, 2014) and are capable, skilled, and qualified to work (Krieger, Kinébanian, Prodinger, & Franziska, 2012; Taylor & Seltzer, 2011). Working improved autistics’ health and quality of life (Beyer, 2016; Chan & Rumrill, 2016; Katz, Dejak, & Gal, 2015). Reducing social service–related costs and increasing disposable earnings, increases national health. In addition, reducing autistic unemployment would noticeably improve autistics’ socioeconomic well-being and quality of life through increased independence and societal contribution. Such positive social change increases respect for human rights in accordance with the United Nations Convention on the Rights of Persons with Disabilities (2009).

Autism is a unique disability first diagnosed in 1908 but not formally designated until the third edition of the Diagnostic and Statistical Manual of Mental Disorders in 1980.

Research on autism and employment began in earnest in 1957 and has continued to date.

Clinical research conducted from 1957 to 2010 did not consider employer perspectives; rather, most studies either ascertained the percentage of the population diagnosed with an autism spectrum disorder (ASD) or their level of employable functionality.

While knowledge of both is critical a component of the equation, it does not address reasons why employers do not hire autistics.

From 2010 to the end of 2017, only three studies pertained to autistic employment from a demand-side perspective:

Stuckey’s (2016) cross-tabulation

Wallis’ (2012) attitudinal review.

In their studies, Stankova and Trajkovski and Stuckey explored potential employer-related factors associated with the high unemployment rates of autistics specifically.

However, Stuckey limited participation to the Rotary Club association in 23 states, which significantly reduced generalizability due to the lack of randomization and a potentially biased sample.

Stankova and Trajkovski’s research lacked generalizability beyond Macedonia. In addition, Stankova and Trajkovski did not limit their sample to those responsible for hiring nor did they test relative to only qualified autistic candidates.

In a periodical article, Wallis recounted hostile treatment received by qualified autistic candidates in the workplace.

Wallis’ qualitative narratives gave a description of the problem but not employer-related reasons for it.

Therefore, throughout 60 years of research, no study has been conducted to inquire why those responsible for hiring do not hire autistics from a nonbiased, generalizable participant sample. However, most literature on the topic inferred one commonality: the beliefs of those doing the hiring.

Problem Statement

Exorbatent autistic unemployment rates (i.e., 83%) are resulting in extreme poverty conditions, declining health, deteriorating quality of life, and maladaptive socioeconomic well-being.

Skilled, capable, unemployed autistics adults (aged 18–65) desire competive employment and deserve the equal opportunity for such the same as all persons.

For policymakers to develop effective policies addressing this problem, they need to know why employers do not hire qualified autistic candidates; however, prior research has not identified the cause. Many scholars inferred that the beliefs of those responsible for hiring are the reason why disabled candidates were not employed; however, the reasons why employers do not hire qualified autistics candidates remain unknown (Copeland, Chan, Bezyak, & Fraser, 2010; Harris Interactive, 2010; von Schrader, Malzer, Erickson, & Bruyère, 2011).

Purpose of the Study

I conducted this study to identify the reasons that hiring agents are not selecting qualified autistic candidates so that society can begin rectifying this horribly depressing problem.

Thus, the purpose of this quantitatively weighted (±90%), concurrent, mixed-methods (QUAN > qual), multiple linear regression study was to begin filling the gap in the literature by identifying why employers, as represented by their hiring agents, do not hire qualified autistic candidates to fill open positions.

To that end, I explored the nature of the relationship between hiring agents’ beliefs and their selection of qualified autistic candidates in this study through Ajzen’s theory of planned behavior conceptually crystalized with five established supporting theories.

In this study, I individually summed and clustered known values of the independent variable (IV; i.e., the control, normative, and behavioral beliefs influencing hiring agents) against the theory of planned behavior-predicted, percentage-based, continuous-level dependent variable (DV; i.e., hiring agents’ selection of qualified autistic candidates).

This regression analysis predicted the degree to which each IV influenced the DV. I used qualitative responses to add insight, depth, and clarity to quantitative findings and to provide participants the ability to provide additional information they deemed pertinent.

Theoretical Framework

Peoples beliefs drive all behavior, even highly analytic professionals such as hiring agents, which is why the theoretical framework for this study needed to focus on predicting behavior driven by beliefs. Ajzen’s (1985) theory of planned behavior was the foundational theoretical framework for the analyses in this inquiry because it provided a robust footing for explaining behavior.

Several scholars have used theory of planned behavior in similar studies, including an exploration of hiring influences (Hernandez et al., 2012).

From 1989 to 2014, scholars debated the value of theory of planned behavior as a stand-alone theoretical model (Bargh, 1989; Sniehotta, Presseau, & Araújo-Soares, 2014);

however, numerous scholars employed theory of planned behavior with reliability nearing 70% (Ang, Ramayah, & Khan VUN, 2013; Araten-Bergman, 2016; Lu, Kao, & Hsieh, 2011). Welbourne (2007) posited that acceptable correlations are usually 60%.

According to Ajzen (2011), even well-designed attitudinal measures of behavior rarely present reliabilities over 80%.

Ajzen listed several study results by other scholars, wherein their correlations ranged from 40% to 67%.

While theory of planned behavior offered a robust theoretical framework for this research, it was conceptually crystalized with other sound theories adding increased reliability to the findings.

In this study, I conceptually crystalized theory of planned behavior with ambivalence amplification theory (Katz, Wackenhut, & Glass, 1986), elaboration likelihood model of persuasion (Petty & Cacioppo, 1979), expectancy value theory (Fishbein, 1963), Fay’s (1987) critical theory, and Pratkanis’ (2000) altercasting theory.

Expectancy value theory added conceptual support to the influence that perceived control beliefs have upon behaviors. Pratkanis’ altercasting theory and Fay’s critical theory conceptually built upon the inference that society influences behavior, thus normative beliefs.

Ambivalence amplification theory and elaboration likelihood model conceptually supported the implication that personal beliefs and experiences (i.e., behavioral beliefs) directly affect behavior.

Such theoretical triangulation provided further support and explanation for Ajzen’s (1985) theory of planned behavior inference that the combination and strength of control, normative, and behavioral beliefs (i.e., the IVs) indicate intent and behavior.

Using this conceptual theoretical framework, this study significantly and statistically (p < .001) accounted for 93% (Mai, 2018) of the variability.

Thus, hiring agents’ combined control, normative, and behavioral beliefs have a significant impact on their selection of qualified autistic candidates.


Legislation, policy, expectations, environment, peer pressure, personal opinions, and more compose a hiring agent’s beliefs, which is why the theory of planned behavior informed the classification of all variables in this study.

Control, normative, and behavioral beliefs are the three theory of planned behavior IVs. I calculated the DV using the tenets of the theory of planned behavior and its predictive nature.

Independent Variables

From routine organizational policy to blatant discrimination, throughout the literature, a plethora of potential hiring agent beliefs surfaced (N = 45; Mai, 2018).

Beliefs are intrinsically inseparable due to the complex, socially related development of an individual’s belief systems (Ajzen, 1985, 2004, 2015).

However, the dominant association of those beliefs allowed categorization between control, normative, and behavioral items while remaining

(a) compatible and

(b) stable over time.

Each taxonomy of belief included subsequent related beliefs (n = 15) that best fit within that distinction.

Control beliefs included those values, attitudes, beliefs, and expectations (VABEs; Clawson, 2012) that stemmed from actual or perceived controlling factors.

Normative beliefs included those VABEs that arose from societal influences and memes. Behavioral beliefs included those VABEs related directly to opinion.

Post hoc analysis also revealed dimension items that hiring agents believe are beyond their control such as public policy mandates and autistic candidate contributions.

Dependent Variables

Whether legally driven or personally passionate, hiring agents’ likelihood to hire is driven by their underlying beliefs.

Ajzen’s (1985, 2004, 2015) predictive theory of planned behavior indicates that the strength of the combined control, normative, and behavioral beliefs infer the undetermined action; in this study, the hiring agents’ intent to hire autistic candidates.

With an unknown DV like this one, it was necessary to gather data directed at calculating the DV based on the theory of planned behavior tenets and predictive nature.

Given that 88% of participants reported a familiarity with autism (Mai, 2018), it was highly likely those associations influenced participants’ beliefs and, therefore, was factored into the equation for determining the DV. Thus, the strength of the IVs, adjusted for autistic familiarity, provided the basis for determining the DV (see Figure 1).

Mai’s (2018) formula used to deduce the DV is shown in equation (1).

First, I summed each participant’s Likert-type scale responses (Σ IV weights) and divided that by the total number of IVs (# of IV).

Each of the 45 IV prompts were measured on a 7-point Likert-type scale; thus, each Likert-type unit was valued at 14 (100/7=14.28571) by which I converted the IV weight into a percentage ((ΣIVweights/45)*14.2857).

Next, I accounted for hiring agent familiarity with autistics.

Participants were queried regarding

(a) how many, if any, autistics they knew,

(b) what those autistics’ level of functionality was,

(c) if they had ever worked with an autistic, and

(d) if their company employed any autistics.

Each of those four items was assigned a max weight (see Figure 2) from which the total DV weight was adjusted according to individual participant responses to each item (maxvalueof(eachcategory−X).

Research Questions and Hypothesis

Why do hiring agents disregard qualified autistic candidates?

In this multiple regression study, I explored one central overarching research question; however, since the theory of planned behavior includes three taxonomies, each represented a separate IV.

Ajzen (1985, 2004) posited that the theory of planned behavior validity as a predictive model rests on the interrelated cornerstone of all three taxonomies: control, normative, and behavioral beliefs.

Thus, the central overarching question of this study and corresponding hypotheses were,What is the nature of the relationship among the potential control, normative, and behavioral beliefs of hiring agents (IVs) and hiring agents’ selection of qualified autistic candidates (DV)?

Null Hypothesis: There is no statistically significant correlation among the IVs and the percentage-based continuous-level DV as assessed through the HASSQAC and predicted through the concepts of theory of planned behavior and measured using multiple regression and all B coefficient values are not statistically significantly different from zero.

Hypothesis A: There is a statistically significant correlation among the IVs and the percentage-based continuous-level DV as assessed through the HASSQAC and predicted through the concepts of theory of planned behavior and measured using multiple regression and at least one B coefficient value is statistically significantly different from zero.

Post-Hoc Analysis

How apropos was this study?

To ascertain the validity and reliability of theory of planned behavior as the fundamental theoretical model, I isolated and additionally explored each IV.

To establish the degree of interrelation between the variables, Mai (2018) ran three additional post hoc regressions on each IV separately.

To this end, the question and hypothesis format remained identical with one exception: the inclusion of all three IVs (i.e., control, normative, and behavioral) altered to reflect each IV:

Subquestion 1: What is the nature of the relationship between the control beliefs influencing hiring agents and their selection of qualified autistic candidates?

Subquestion 2: What is the nature of the relationship between the normative beliefs influencing hiring agents and their selection of qualified autistic candidates?

Subquestion 3: What is the nature of the relationship between the behavioral beliefs influencing hiring agents and their selection of qualified autistic candidates?

Nature of the Study

Albeit statistics drive organizational change, vision enables people to see the meaning behind those numbers and both drive public policy; thus, both quantitative and qualitative data were gathered.

In this quantitatively weighted (±90%), concurrent, mixed-methods (QUAN > qual), multiple linear regression study, I explored the degree to which hiring agents’ control, normative, and behavioral beliefs related to their selection of qualified autistic candidates.

Multiple regression presented an optimal quantitative venue for predicting the influence of known values of multiple IVs on an unknown value of a single DV.

Quantitative methodology provided a clear understanding of the predictive value that the IVs have on the DV and best demonstrated the inferential, statistical findings.

Qualitative responses added insight, depth, and clarity to quantitative findings and allowed participants to add anything they deemed applicable to the topic.

Internet-based participant solicitation and concurrent data collection allowed efficient study administration, reduced expenditures, and met the theory of planned behavior’s time-stable assumption.


To reach across the nation, I used a nation-wide forum. Using social media–style participant solicitation, I targeted a representative, simple, random probability sample of hiring agents serving medium-sized organizations (i.e., 50-249 employees) throughout the contiguous United States over the timespan of 9 months.

Some respondents were outside of those parameters (n = 7 [3.4%]) but were included in the analysis due to the random probability nature of the sample.

Of the total number of participants (N = 212), 88% reported familiarity with autistics. Social media solicitation influenced participation through homophily and transitivity.


Five highly reliable tools supported data collection and analysis. Resources (i.e., tools and instruments) used in this study included Qualtrics (2014) survey hosting, International Business Machine’s Statistical Package for the Social Sciences (SPSS), version 23; Microsoft Excel 2016; NVivo 11; and the HASSQAC (Mai, 2015).

All data collection occurred via the HASSQAC, which was administered electronically via Qualtrics.

HASSQAC development and testing occurred in four steps:

(a) completion of a literature review to identify potential instrument design and potential variables;

(b) adaptation of Copeland et al.’s (2010) version of the Affective Reactions, (c) subscale of the Disability Questionnaire (α = .69 to α = .85) and Kaye, Jans, and Jones’s (2011) Employer Questionnaire, Parts I and II manner of third-person inquiry;

(d) administration of the HASSQAC, which included one 7-point Likert-type scale prompt for each variable (N = 45) and nine demographic questions each directed at hiring agents and relating to their beliefs relative to potential autistic employees; and

(e) completion of peer review from a panel of professionals in the fields of business, vocational rehabilitation, autism clinicians, advocates, and autistics. Of 212 responses, 130 were complete data sets; consequently, the confidence level varied between 80.28% and 74.54%.


An anonymous person has nothing to hide; thus, the anonymous individual will tell all unhindered by accusation.

To ensure anonymity, Qualtrics gathered no IP addresses or any other identifiers pertaining to participants’ identities, and data were transferred directly from Qualtrics into SPSS, and then erased from Qualtrics’ servers.

Quantitative data remained in SPSS for analysis and the DV was ascertained through concepts of theory of planned behavior (see Equation 1).

Qualitative data were transferred to NVivo for analysis.

I used a multiple regression model to analyze the quantitative data gathered via a 7-point Likert-type scale survey tool (HASSQAC, Mai, 2015) through the lens of theory of planned behavior.

After substantiating assumptions and model fit, quantitative analysis consisted of four regressions: the main theory of planned behavior model and one post hoc analysis for each taxonomy.

Qualitative data were coded and analyzed for similarity, theme, redundancy, and alignment with participant’s response.

Content and Construct Validity

Numerous checks and balances ensured significant design validity. Study delimitations included several components to increase content and construct validity.

The study was not experimental in nature, so no threat of unreliable regression toward an overall mean existed. The absence of a pretest or posttest prevented prior knowledge of the survey from impacting the data.

Participants received no compensation; therefore, there was no potential for skewed results based on rivalry.

The purposeful, random sample selection targeting hiring agents serving medium-sized businesses in the contiguous United States minimized the chance of participant predisposition toward a specific outcome and reduced the likelihood of data diffusion.

Participants consisted of one group which minimized the potential for time-related threats.

However, since this study was time-bound, there is restricted transferability to past- or future-inhibited, time-related applications without replication throughout time; therefore, the detailed study design documentation allows for future replication.

The survey duration averaged 15 minutes and could be completed only once by each participant, reducing the potential for maturation (i.e., change) or tainted data during collection.

External factors and existing literature informed my exhaustive exploration of potential framework, theories, and populations to find the best design to increase validity, reliability, and generalizability of this study.

Framework crystallization increased internal validity, while purposeful sampling increased external validity.

The inclusion of the population (N ≈ 443,000) and setting characteristics, provided by Dunn & Bradstreet (DeJesus, 2015), increased generalizability throughout the contiguous United States and minimized the potential for incorrect inference beyond that population.

Publication of the survey in only the English language reduced the potential generalizability beyond English-speaking participants. However, my participation solicitations targeted a population, wherein fluent English is an essential skill; thus, minimizing that limitation.

I solicited participants via social media and administered the survey via the Internet. These remote data-collection techniques provided additional construct validity throughout the contiguous United States and allowed convenient participation which increased the sample size and presented the ability to control the line of questioning.

Albeit, the indirect communication and anonymity of the collection method increased the likelihood of participants misunderstanding the questions and, thus, some data sets were incomplete. The data-collection tool, the HASSQAC, tested highly reliable.

HASSQAC development and testing was a critical component to increasing content validity. First, I conducted a thorough and detailed literature review to identify potential instrument designs and variables.

My survey design incorporated theory of planned behavior crystallization with two highly reliable existing tools informed via extensive literature review to establish variables and question prompts (Mai, 2018).

Theory of planned behavior’s incorporation of control, normative, and behavioral beliefs presented a threat to variable cross-categorization, so each HASSQAC question pertained to each identified potential variable with 15 items within each of the three belief taxonomies (see Figure 1).

This format minimized potential cross-categorization and the risk of multiple treatments of variables and allowed the covariant examination of predicting the influence of one variable on another.

In this way, the study design incorporated the potential convergence of theory of planned behavior belief categories.

Finally, professionals in the fields of business, vocational rehabilitation, autism clinicians, advocates, and autistics (n = 13) reviewed the tool, ensuring the validity of questions, relevance, format, and phrasing.


Study reliability surpassed expectations.

The regression was statistically significant, F(45, 73) = 36.067, p < .001, adjusted R2 = .930, revealing a balanced influence of control, normative, and behavioral influencing beliefs.

The strongest influencers included:

hiring agents’ desire for comprehensive organizational diversity (B = 0.266, p < 001),

the strength of their stereotypical associations (B = 0.195, p < 001),

and their fear of embarrassment (B = 0.187, p < 001; see Table 1).

My post hoc analysis investigating each taxonomy individually supported the strength of these items and added insight into associated influencers within each taxonomy.


Hiring agents want mandated compliance. Isolation of the control taxonomy revealed additional insight into this belief type, F(15, 107) = 20.688, p < .001, adj. R2 = .708; see Table 2.

Organizational diversity plans including autistics must evolve from control-related mandates extending to dedicated personnel. In addition, the organizational ontology, represented by its commitment to hire autistics, resource allocation, and leadership behaviors, would increase hiring agents’ selection of qualified autistic candidates.

Hiring agents’ socially driven VABEs and personal memes reduced their belief in costly organizational burdens associated with autistic employees. Similarly, hiring agents’ normative and behavioral beliefs also dampened their aversion to redesigning work processes to accommodate autistic employees.

Overwhelmingly, hiring agents desired mandated autistic diversity initiatives and a readily accessible external mediation service supporting the ongoing associated training and facilitation initiatives.

Having a “written company policy specifically addressing” is not enough. Unless it is an actively pro-hire policy with metrics and follow-up, it will likely have little effect. Just writing “we can and do hire candidates with autism” in a handbook will not have a marked impact. HASSQAC Respondent ID # R_22Wzd5dGhbdGZiv


What is important to society will impact hiring practices and hiring agents do not believe that autistics are important. Isolation of the normative taxonomy revealed additional insight into this belief type, F(15, 106) = 34.686, p < .001, adjusted R2 = .807; see Table 3.

Hiring agents believe that society does not feel that employment for autistics is important and that socially accepted processes, such as employment and credit screening, frequently disqualify autistics due to disability-related lapses.

However, their control and behavioral beliefs could mitigate the significant weight of those societal VABEs if public demonstration conveyed differing altercasted expectations.

I think if a candidate can help mitigate a company’s common fears/stereotypes- that would greatly help them. I find it awful that such a burden falls on the Autistic candidate- but - it would help. HASSQAC Respondent ID #R_1P6oZqzH4HubEDG

In addition, hiring agents believe that candidates could offset a portion of those negative associations if they could legally, and ably, identify themselves as autistic and alleviate potential organizational concerns.

Hiring agents believe autistic candidates should relate their potential organizational contributions, in consideration of their autism, to help offset existing poor dependability expectations of autistic workers and organizational needs for job redesign to accommodate them.

Hiring agents believe other employees could not accept instruction from an autistic supervisor and that coworkers would feel inadequate.


Unless it is repetitive work, how do I keep an autistic person focused. We used to say you spend 99% of a manager’s time dealing with that 1% shitbird. HASSQAC respondent ID #R_1IFRR8cqIaDORtf

Put simply, many hiring agents flat-out discriminate.

Isolation of the behavioral taxonomy revealed additional insight into this belief type, F(15, 109) = 11.066, p < .001, adjusted R2 = .549; see Table 4.

Hiring agents doubt the skills of autistic candidates, believe autistics are unproductive and undependable, and related that they blatantly discriminate against autistic candidates.

My analysis indicated that control-related mandates, training, and societal expectations could offset these stereotypes and discrimination; however, the results also demonstrated a significant preference of hiring agents for physical disabilities over autistic and other developmental, mental, emotional, and unseen/not obvious disability types.

Conceptual Analysis

Hiring agents reveal their personal beliefs are responsible for 55% of autistic unemployment but feel that ineffective or nonexistent mandates are responsible for 32% and autistics the remaining 13%.

In Mai and Ferreros’ (2018) post hoc evaluation of the HASSQAC, the scale demonstrated reliability in separating scale items into categories that hiring agents felt were most within their domain to affect.

Mai and Ferreros used forced, three-factor extraction using varimax orthogonal rotation separated items into three categories: demand-side, supply-side, and hiring agents’ VABEs.

Considering factor loadings <.40, there were 14 items hiring agents felt must extend from mandated organizational and public policies (i.e., demand-side, α = .929) and six items hiring agents believe are most affected by the autistic candidates (i.e., supply-side, α = .872).

Of the remaining, hiring agents believed 24 items were most influenced by their own VABEs (α = .945).

This conceptual theoretical model affirmed Ajzen’s theory of planned behavior, and its relationship to ambivalence amplification theory, elaboration likelihood model, expectancy value theory,

Fay’s critical theory, and Pratkanis’ altercasting theory, and it identified which items hiring agents believe most influence their selection of qualified autistic candidates and which items are most effected by policy, candidate, and hiring agent.


At an 83% unemployment rate, qualified and capable autistic candidates face a substantial crisis of workplace rights.

Thus, the central research question that guided this study was:

What is the nature of the relationship among the potential control, normative, and behavioral beliefs of hiring agents and hiring agents’ selection of qualified autistic candidates?