“Always a glass ceiling.” Gender or autism; the barrier to occupational inclusion

Research in Autism Spectrum Disorders 56 (2018) 50–60

Contents lists available at ScienceDirect

Research in Autism Spectrum Disorders journal homepage: www.elsevier.com/locate/rasd

“Always a glass ceiling.” Gender or autism; the barrier to occupational inclusion

T

⁎

Susan M. Haywarda, , Keith R. McVillya, Mark A. Stokesb a b

School of Social & Political Sciences, The University of Melbourne, Victoria, 3010, Australia School of Psychology, Faculty of Health, Deakin University, 221 Burwood Highway, Burwood, Victoria, 3125, Australia

A R T IC LE I N F O

ABS TRA CT

Number of reviews completed is 2

Background: Under- and unemployment adversely affect the economic, health, and social circumstances of people with autism; notably those with a diagnosis of autism spectrum disorder or high autistic traits (HATs). However, little research has been published comparing the experiences of women to men with HATs, and women without autism (i.e., those typically developing; TD) to ascertain if employment issues are a function of gender or autistic traits (ATs). Method: An anonymous online survey was conducted attracting 28 women and 18 men with HATs aged 18–68 years (M = 38.63, SD = 13.12), with a further 21 TD women and 16 TD men aged 23–62 years (M = 38.38, SD = 10.32). Quantitative data were analysed via logistic regression to ascertain the extent to which employment issues were a function of gender or ATs while controlling for confounding variables such as education, and age. Qualitative data were analysed using inductive thematic analysis, then quantitatively using chi-square or Fisher’s Exact Test. Results: It was found that ATs, not gender, was significant to most vocational experiences. Conclusions: It is proposed that employers place greater importance on technical ability than social-communication skills when hiring and supervising women with HATs to reduce barriers and increase workplace diversity.

Keywords: ASD Asperger’s Female Job Sex Work

1. Introduction Affecting approximately one in 189 females and one in 42 males (Baio, 2014), people with neurodevelopmental conditions such as those with High Autistic Traits (HATs), including Asperger’s Syndrome, exhibit greater difficulties with interpersonal skills than their typically developing (TD) peers; i.e., those without autism (Gal, Landes, & Katz, 2015). While challenges with social-communication and interaction are the hallmarks of autism (American Psychiatric Association [APA], 2013), these skills are suggested to be imperative to workplace success (Smith, 2013). This includes gaining (Charney, 2016; Deepa & Seth, 2013; Jones, Baldi, Phillips, & Waikar, 2016), maintaining (Agran, Hughes, Thoma, & Scott, 2016; Lin & Kwantes, 2015), and advancing in employment (Deepa & Seth, 2013; Lin & Kwantes, 2015). However, some employers recognise the unique technical skills possessed by those with HATs compared to those TD (e.g., ASPertise, SAP, Specialisterne, ULTRA Testing, Willis Towers Watson, and Microsoft). Recognition of diversity in employment challenges the social model of disability which stipulates that individuals are not impaired by their disability but by societal barriers

⁎

Corresponding author. E-mail addresses: [email protected] (S.M. Hayward), [email protected] (K.R. McVilly), [email protected] (M.A. Stokes). https://doi.org/10.1016/j.rasd.2018.09.001 Received 23 February 2018; Received in revised form 3 September 2018; Accepted 4 September 2018 1750-9467/ © 2018 Elsevier Ltd. All rights reserved.

Research in Autism Spectrum Disorders 56 (2018) 50–60

S.M. Hayward et al.

(Oliver, 1983). Yet, many with HATs report unfavourable employment circumstances; difficulties gaining and maintaining employment, and skilled underemployment (Hurlbutt & Chalmers, 2004; Müller, Schuler, Burton, & Yates, 2003). These issues can negatively affect well-being (Blustein, Kozan, & Connors-Kellgren, 2013; McKee-Ryan, Song, Wanberg, & Kinicki, 2005; Rosenthal, Carroll-Scott, Earnshaw, Santilli, & Ickovics, 2012). Yet, the extent to which some of these issues might be a function of gender or autistic traits (ATs), while controlling for potenial confounding variables (e.g., education and age) is yet to be investigated. While challenges with social-communication skills might account, in part at least, for unfavourable occupational experiences of people with HATs, only one known paper has compared individuals with HATs to TD individuals which included women (Gal et al., 2015). These authors reported differences in work related aptitude, such as interpersonal and working styles. However, this research did not investigate the potential influence of gender. Here, gender could be an important variable given other research suggests women with HATs might possess ‘superior’ social-communication skills compared to their male counterparts (Brooks, 2016; Head, McGillivray, & Stokes, 2014; Lai et al., 2011). For women, ‘superior’ social-communication ability could moderate the effects of ATs in the labour market. Still, gender could negatively impact labour market experiences for women generally (Hoobler, Wayne, & Lemmon, 2009; Joshi, Jooyeon, & Hyuntak, 2015; McGraw, Kramar, & McGraw, 2011), and especially for women with disabilities (Australian Bureau of Statistics [ABS], 2015; Doren, Gau, & Lindstrom, 2011). Some authors argue that gender inequality (England, 2005; Folbre, 1994) and structural oppression (Ritzer & Goodman, 2008; Young, 2011), i.e., organisations that possess norms favouring men, may contribute to women’s under-representation in the workforce (ABS, 2017b). This includes in full-time employment (Workplace Gender Equality Agency, 2014). Women are also over-represented in part-time and casual work (ABS, 2017a). Thus, they may have limited access to the economic benefits provided by adequate employment (Schofield et al., 2011). In conjunction with gender, it is possible that education and age may moderate employment experiences. Post-secondary attainment (education beyond High School) may decrease unemployment risk (ABS, 2017b; Ohl et al., 2017). While younger aged persons are at increased risk of unemployment (ABS, 2017b). Although skilled underemployment, defined by Duffy (2009) as working in a role where skill, knowledge, and experience are not fully utilised, is problematic among those with HATs (Baldwin, Costley, & Warren, 2014; Hurlbutt & Chalmers, 2004; Müller et al., 2003), another type of underemployment yet to be examined is time-related underemployment. Time-related underemployment is not working as many hours, up to full-time, to which a person is willing and able (Hauser, 1974). Research suggests potential timerelated underemployment issues among people with HATs, as most of these unemployed individuals report being willing to work (Autism Spectrum Australia [Aspect], 2013; Griffith, Totsika, Nash, & Hastings, 2012). Further, over-representation of those with HATs in casual employment (Autism Spectrum Australia Aspect, 2013; Baldwin et al., 2014) may follow similar patterns in the general population where, compared to full-time employees, casual employees are more likely to report wanting to work more hours (ABS, 2009). In addition, age may moderate this relationship as younger individuals, in the general population, are more likely to be casually employed (ABS, 2014). Thus, this research contributes to the current literature by investigating if vocational experiences such as skilled and time-related underemployment are a function of gender or ATs, while controlling for either education or age utilising samples of women and men with HATs and TD individuals. Understanding these occupational experiences will assist addressing areas of concern and support reasonable and effective adjustment, if required, to mitigate potential social and economic impacts. Therefore, based on what is known regarding the employment experiences of women generally and individuals with HATs, the following research questions were posed: are overall unfavourable employment experiences influenced by gender or ATs?; Is gender or ATs predicative of occupational instability?; Does gender or ATs influence difficulty maintaining employment?; Does gender, ATs or education predict unemployment?; Are gender, ATs, or age predictive of skilled and time-related underemployment?; Is casual, part-time or full-time employment predicted by gender, ATs, or age? 2. Method 2.1. Participants The protocol for this project was reviewed and ethics approved by the overseeing university (approval number 1749 897). Participants were obtained via autism related social media sites as well as the websites of autism organisations. Respondents reported themselves to reside in Australia; these were 49 women and 34 men. Participant demographic details are provided in Table 1 where significant differences (p < .05) between women and men with HATs, and women/men with HATs vs TD women/men are indicated. 2.1.1. Participant grouping Participants within the HAT group reported a diagnosis of autism and possessed an Autism Spectrum Quotient (AQ) score (a screening tool for ATs) at or above the published criterion of 32; the recommended score for a diagnostic assessment referral (BaronCohen, Wheelwright, Skinner, Martin, & Clubley, 2001). Conversely, those assigned the TD group did not report a diagnosis of autism and scored below 32 on the AQ. The AQ was completed online by participants at the end of the study given its extensive use in research. This was to maximise participation by not providing a screening tool at the beginning which some might find intrusive. It was thought that potential participants might be more open to completing the AQ after having completed the narrative component of the survey. The tool has good overall fit, discriminating those with HATs from TD individuals (Ruzich et al., 2015). This includes women with HATs from TD women (Lau, Kelly, & Peterson, 2013). Further, the AQ is highly sensitive to HATs at a criterion of 32 (75% and 77% respectively; Broadbent, Galic, & Stokes, 2013; Woodbury-Smith, Robinson, Wheelwright, & Baron-Cohen, 2005). 51

Research in Autism Spectrum Disorders 56 (2018) 50–60

S.M. Hayward et al.

Table 1 Demographic make-up of participants. Demographics

Women

Men

Typically Developing

High Autistic Traits

Typically Developing

High Autistic Traits

Number Age Range (Years) Mean Age (Years) AQ Mean Mean Age of ASD Diagnosis (Years; where applicable)

21 24–62 37.81 (10.62) 16.19 (6.32)* N/A

28 18–60 33.79 (11.09)# 39.43 (4.63)* 28.79#

16 23–61 39.13 (10.20) 16.5 (6.83)* N/A

18 20–68 46.17 (12.69)# 38.67 (4.17)* 39.78#

Highest Level of Completed Education (%) Year 9 or below Completion of Year 10 High School or junior vocational school Completion of High School or advanced vocational school Diploma or Advanced Diploma Bachelor degree including Honours or equivalent Master’s degree Doctoral degree

0.0 4.8 0.0* 14.3 57.1* 9.5 14.3*

0.0 0.0 35.6* 17.9 28.6* 17.9 0.0*

0.0 0.0 25.0 18.8 50.0 6.3 0.0

5.6 5.6 27.7 11.1 44.4 0.0 5.6

AQ = 50 item Autism Spectrum Quotient. ASD = Autism Spectrum Disorder. N/A = not applicable. * Independent sample t-test evidenced a significant difference between women or men HATs vs TD; p < .05. # Independent sample t-test evidenced a significant difference between men and women with HATs; p < .05.

Thus, 66 respondents were excluded from the current analysis due to incongruities between their self-reported diagnostic status, requested at the beginning of the questionnaire, and AQ score. Twenty-five of these participants reported a diagnosis of autism but scored below the cut-off on the AQ. Thirteen participants did not report a diagnosis of autism yet scored above the AQ cut-off. Further exclusions applied to 28 participants who reported a possible diagnosis of autism yet did not possess a formal assessment; 23 scoring above and five below the AQ cut-off. Most women with HATs reported a diagnosis of autism by at least one of the following: Psychologist (64%, n = 18), Psychiatrist (32%, n = 9), and other health professional (11%, n = 3). The majority of men also reported a diagnosis of autism by at least a: Psychologist (83%, n = 15), Psychiatrist (39%, n = 7), Paediatrician (6%, n = 1), or other health professional (6%, n = 1). Women were diagnosed with autism between three and 53 years of age (M = 28.79, SD = 12.75), significantly earlier (t(44) = 2.71, p < .05, d = .82) than men diagnosed with autism who were aged seven to 58 years (M=39.78, SD = 14.45).

2.2. Materials and procedure Considering the social-communication difficulties experienced by those with HATs (American Psychiatric Association APA, 2013), written data collection was chosen to assist with comprehension and self-expression (Benford, 2008; Gillespie-Lynch, Kapp, ShaneSimpson, Shane-Smith, & Hutman, 2014). Therefore, a 16-page anonymous online survey consisting of both multiple choice and short answer questions was developed. The narrative questions were written with the assistance of a woman with HATs, as well as piloted over a sample (n = 4) with HATs and TD individuals to check questions for clarity prior to the study being advertised. Any person over the age of 18 years without an intellectual disability was invited to participate, i.e., those with and without a diagnosis of autism. The advertisement indicated that the researchers were particularly interested in hearing from women.

2.2.1. Education Participants’ highest level of educational attainment was obtained by qualification. For parsimony education was grouped into low (0=no post-secondary education, i.e., completion of High School or lower) and high (1=post-secondary education, i.e., educational attainment higher than High School).

2.2.2. Employment experiences To ascertain employment experiences, participants were asked to “Think about your employment history a) Please briefly describe your employment history. For example, has it been stable/inconsistent, overall a positive/negative experience?”

2.2.3. Unemployment and employment type Unemployment and employment type were assessed by asking participants, “I currently work” (1=as a volunteer, 2=casual, 3=part-time, 4=full-time, 5=I am studying full-time and not working, 6=I am studying full-time and working casually/part-time, 7=I am not working, 8=I am not working and I have never been employed). Unemployment constituted responses seven and eight and was re-coded into a new variable representing 0=unemployed and 1=employed. 52

Research in Autism Spectrum Disorders 56 (2018) 50–60

S.M. Hayward et al.

2.2.4. Underemployment Two types of underemployment were investigated; skilled and time-related. The experience of skilled underemployment was assessed by, “I have been employed in a role/s that was/were below the qualification level I held at the time” (0=no, 1=yes). Experiences of skilled underemployment were verified from respondents who indicated positively (i.e., 1=yes (skilled underemployment experienced)), “Would you please provide some brief comments (dot points are fine) about your employment experience/s? Such as: a) Why you consider/ed your job to be below the level of your education/qualifications? b) What industry you have held a role that was below your educational level/qualifications? c) Why you took the job? d) Why you stayed/left the job?” The verification process compared the skill level provided by the participant’s qualification/s held at the time of reported skilled underemployment with the skill level of the role they advised being underemployed in. These were assessed and compared for discrepancies using the Australian and New Zealand Standard Classification of Occupations (ANZSCO; Trewin & Pink, 2006). Roles that were ambiguous which could be considered across more than one ANZSCO level were coded as belonging to the highest level. If the qualification/s and role held were at the same level of skill, or if the participant provided insufficient information for comparison, the claim of skilled underemployment was coded 0=unsubstantiated skilled underemployment. Conversely, differences in the skill level of the qualification/s held and role/s were coded 1=substantiated skilled underemployment. To ascertain time-related underemployment, those working on a voluntary, casual, or part-time basis (excluding those studying full-time and working) were asked to indicate, “I want to work more hours than I currently am” (0=no, 1=yes). Affirmative responses were taken as evidence of time-related underemployment. 2.3. Analysis 2.3.1. Inductive thematic analysis Only participants that answered the narrative question concerning workplace experiences were included in qualitative analyses, thus one participant who had never been employed was removed. The remaining data were subject to Inductive Thematic Analysis (ITA; Gray, 2014) given the paucity of theoretical perspectives owing to limited literature distinguishing both gender and ATs in employment. Although, Interpretative Phenomenological Analysis was considered. However, as it is grounded in a phenomenological epistemology, ITA was chosen as it is not linked to a pre-existing framework (Willig, 2013). Following guidelines by Braun and Clarke (2006), ITA was conducted using NVivo Pro 11 (QSR International, 2015). As participant responses were closely related to the question asked, themes were named as such where practical. To illustrate, when asking if employment has been unfavourable or favourable; participants usually indicated that their employment history has been unfavourable, favourable, or both. For the latter, the response was coded to both themes. However, if the participant stated opposing responses favouring one, only the favoured was coded. For example, if the response was, “Some good experiences, some bad experiences…Mostly I’ve enjoyed work”, the response was coded only to the favoured “favourable” theme. Likewise, determination of stability in employment was based solely on the participant’s judgement. Saturation was reached after five participant responses were coded. After the primary researcher classified responses to themes, a secondary researcher double coded the first 10% of responses before coding was compared between researchers (κ = 1.0). Member checking was not possible due to the anonymous nature of data collection. Finally, qualitative data were transcribed into the Statistical Package for the Social Sciences (v23; IBM Corp.) and analysed quantitatively where each theme derived from the ITA was coded dichotomously for each participant against each theme (0=theme absent from response, 1=theme present in response). To illustrate, if a participant stated difficulty maintaining employment this was coded affirmatively (1=theme present in response) to the “difficulty maintaining employment” theme for that participant; conversely if a participant did not state this (i.e., 0=theme absent from response). 2.3.2. Comparative analysis Comparative analysis using chi-square or Fisher’s Exact Test were undertaken on the themes emerging from the ITA in answering research questions one to three. To ascertain the effect of gender, the frequency of responses of women with HATs were compared to men with HATs. To determine differences based on ATs, women with HATs were compared to TD women. Owing to multiple comparisons, the alpha was adjusted to .01. 2.3.3. Logistic regression Logistic regression ascertained the relationship between gender (0=woman; 1=man), ATs (0=TD; 1=HAT), and education (0=low; 1=high) or age as appropriate to the research question concerning unemployment and employment type. Interaction effects were considered if apparent. To ensure independence of variables before regression analyses were undertaken, correlations among all variables were reviewed, results are reported in Table 2. 2.3.4. Sample size Post-hoc power analyses using G*Power (Faul, Erdfelder, Lang, & Buchner, 2007) revealed that the sample size range required for regression analyses was 61–81 depending upon assumptions. The following assumptions were made for the higher and lower limit respectively: a) an Odds Ratio (OR) of 18 or 30 (equivalent to R2 = 0.39 and 0.47) given that the study by Gal et al. (2015) comparing individuals with HATs and TD individuals found an effect size measuredby Cohen’s d of 2.85 (equivalent OR = 175.80); b) the probability of a mean difference given a null hypothesis was set to 0.2 or 0.4, and; c) R2 of 0.55 or .30. Further, the proportion of 53

Research in Autism Spectrum Disorders 56 (2018) 50–60

S.M. Hayward et al.

Table 2 Pearson’s Correlations Among Variables for Logistic Regression.

HATs/TD Gender Age Education

HATs/TD

Gender

Age

Education

1 0.001 0.046 −0.186

1 0.305* −0.142

1 0.141

1

HATs = High Autistic Traits. TD = Typically Developing. * Correlation is significant at 0.01 level (2-tailed).

participants with HATs in the sample was set to .55, and power set to 0.8 using an alpha of .01, with a two-tailed binomial distribution. 3. Results For the narrative questions, the core themes induced together with definitions, frequency of responses, and exemplar comments are provided in Table 3. Five themes emerged: overall employment experiences (unfavourable/favourable), stability in employment (unstable/stable), and challenges maintaining employment. The results of the subsequent comparative analysis are reported in Table 4. These apply to research questions one to three inclusively. 3.1. Research question 1: overall employment experiences This study questioned whether an overall unfavourable employment history was influenced by gender or ATs. No differences between women and men with HATs were found. However, there was a significant difference between women with HATs compared to TD women indicating that women with HATs were more likely to report unfavourable employment experiences. For example, “… overall very negative. I tend to … not fit in and becom[e] the target of bullying” (woman with HATs, aged 36); “I had a lot of negative experiences from management. I was bullied by other staff and management (woman with HATs, aged 25). Thus, there were more qualitative reports of ATs, not gender, influencing unfavourable occupational experiences. 3.2. Research question 2: stability in employment It may be anticipated that unfavourable experiences and job instability are related. However, some respondents indicated to the contrary, e.g., “inconsistent yet somewhat enjoyable” (man with HATs, aged 34). Thus, the second research question concerned whether gender or ATs influenced unstable employment. When comparing women and men with HATs no significant differences in frequency of responses were found. However, women with HATs compared to TD women were more likely to report unstable employment. To illustrate, “not as stable as I would have liked. I have taken whatever is available” (woman with HATs, aged 42); “I have had countless jobs. I could never work full time in one. I have tried and each time they have failed. Working in the same group [of people] for anything more than a few months requires navigating complex interpersonal developments that I find overwhelming even though I am brilliant at leading groups at very specific projects/tasks … sustaining it for any lengthy period is very[,] very taxing” (man with HATs, aged 49). Thus, there were more reports of ATs increasing the likelihood of having unstable employment, not gender. 3.3. Research question 3: challenges maintaining employment The third research question asked if ATs or gender influenced challenges maintaining employment. Although no significant difference was found between women and men with HATs, women with HATs were significantly more likely to report difficulties gaining or maintaining work; see Table 4. Exemplar comments included: “I usually last about 5 years in a job although as [I’m] getting older [I] have lasted longer and now [I’m] nearly to my first[,] lol [laugh out loud] 10 years[,] yay[,] but it has been a battle” (woman with HATs, aged 53); “started working for myself at age 22 because it was easier than having to try to fit in a job with other people. I kept trying to get work though, and after being knocked back because I was "not qualified", I decided to focus on getting qualifications. Now I have 2 degrees, 8 certificates, and taking [a] Masters and a 2nd undergrad[uate degree]” (woman with HATs, aged 34). Therefore, there were more reports of challenges maintaining employment being influenced by ATs, not gender. 3.4. Research question 4: unemployment To determine if gender, ATs, or education predict unemployment, these variables were used in a regression but did not fit the data (see Table 5). This suggested that unemployment was not predicted by gender, ATs or education. 54

HATs = High Autistic Traits. TD = Typically Developing.

Overall Employment Experiences

“Think about your employment history… a) Please briefly describe your employment history. For example, has it been stable/ inconsistent, overall a positive/ negative experience?

Challenges Maintaining Employment

Stability in Employment

Major Themes

Survey Question

Table 3 Summary of Core Themes Induced from the Survey Question.

Not always having work, or employment that has not been ongoing

Usually or always having employment

Challenges relating to gaining or keeping work

Stable employment

N/A

Indication of pleasurable/ positive work experiences

Favourable experiences

Unstable employment

Negative or displeasure experienced relating to employment

Definition of Theme

Unfavourable experiences

Minor Themes

Overall Women HATs Men HATs Women TD Men TD Overall Women HATs Men HATs Women TD Men TD Overall Women HATs Men HATs Women TD Men TD Overall Women HATs Men HATs Women TD Men TD Overall Women HATs Men HATs Women TD Men TD

Frequency of Response (%)

“I usually have communication issues with colleagues, I am perceived as too detail-focussed and too questioning, blunt with co[-]workers and "lacking soft skills". I feel isolated…”

“I enjoyed most of my work experiences…” “Overall positive experience…”

“I have never held a proper, stable job (except for some casual, self-employed work).” “It has been very inconsistent and unstable…I have attempted to get support through employment agencies but so far they have been useless.” “I have been constantly and stably employed for approximately 21 years with the same entity. I have changed roles within this time and secured promotions.” “I have had a fairly stable employment history.”

“Employed since I was 14 in part-time roles, never full-time. Didn't manage full-time study well either…difficulties with communication have made it hard at times, especially in the profession for which I studied at uni but only worked in for a very short time” “… usually within a month or two I am beginning to struggle in my role. Even when I'm performing well…”

27 54 33 0 6 35 25 17 43 63 24 43 39 5 0 53 29 28 81 88 22 32 44 0 6

Exemplar Comments

S.M. Hayward et al.

Research in Autism Spectrum Disorders 56 (2018) 50–60

55

Research in Autism Spectrum Disorders 56 (2018) 50–60

S.M. Hayward et al.

Table 4 Comparative Results from Each Theme. Theme

Women HATs vs men HATs

Women HATs vs women TD

Unfavourable Employment History Difficulty Maintaining Employment Unstable Employment History

FET, p≥.01, ϕ = 10 χ²(1) = .71, p≥.01, ϕ = 0.12 FET, p≥.01, ϕ = 0.10

FET, p < .01, ϕ = 0.79* FET, p < .01, ϕ = 0.41* χ²(1) = 15.53, p < .01, ϕ = 0.65*

FET = Fisher’s Exact Test. HATs = High Autistic Traits. TD = Typically Developing. * Significant at 0.01. Table 5 Regression Results. Dependent Variable

Overall Model

Independent Variables in Model

Result

Unemployment

χ²(3) = 2.41, p≥.01

ATs

b = 0.95, SE = 0.75, Wald χ²(1) = 1.60, p = .21, OR = 2.57, CI = 0.60–11.13 b = 0.56, SE = 0.69, Wald χ²(1) = 0.65, p = .42, OR = 0.57, CI = 0.15–2.21 b = 0.02, SE = 0.77, Wald χ²(1) = 0.77, p = .98, OR = 1.02, CI = 0.23–4.58

Gender Education Skilled Underemployment

χ²(3) = 5.55, p≥.01

b=−0.60, SE = 0.46, Wald χ²(1) = 1.69, p = .19, OR = 0.55, CI = 0.23–1.35 b=−0.88, SE = 0.51, Wald χ²(1) = 2.97, p = .09, OR = 0.41, CI = 0.15–1.13 b = 0.03, SE = 0.02, Wald χ²(1) = 1.90, p = .17, OR = 1.03, CI = 0.99–1.08

ATs Gender Age

Time-related

χ²(3) = 4.99, p≥.01

Underemployment

b=−2.34, SE = 1.40, Wald χ²(1) = 2.81, p = .09, OR = 0.10, CI = 0.01–1.49 b = 0.07, SE = 1.32, Wald χ²(1) = 0.003, p = .96, OR = 1.07, CI = 0.08–14.19 b=−0.05, SE = 0.05, Wald χ²(1) = 0.83, p = .36, OR = 0.96, CI = 0.86–1.06

ATs Gender Age

Casual Employment

χ²(3) = 6.94, p≥.01

b = 2.25, SE = 1.09, Wald χ²(1) = 4.27, p = .04, OR = 9.46, CI = 1.12–79.64 b = 0.08, SE = 0.89, Wald χ²(1) = 01, p = .93, OR = 1.08, CI = 0.19–6.21 b = 0.01, SE = 0.04, Wald χ²(1) = 0.16, p = .69, OR = 1.01, CI = 0.95–1.09

ATs Gender Age

Part-time Employment

χ²(3) = 0.78, p≥.01

b = 0.22, SE = 0.71, Wald χ²(1) = 0.10, p = .75, OR = 1.25, CI = 0.31–5.07 b = 0.48, SE = 0.81, Wald χ²(1) = 0.34, p = .56, OR = 1.61, CI = 0.33–7.88 b=−0.03, SE = 0.03, Wald χ²(1) = 0.56, p = .45, OR = 0.98, CI = 0.92–1.04

ATs Gender Age

Full-time Employment

χ²(3) = 16.35, p < .01

b=−1.72, SE = 0.50, Wald χ²(1) = 11.57, p = .001, OR = 0.18, CI = 0.07–0.48 b = 0.06, SE = 0.53, Wald χ²(1) = 0.01, p = .91, OR = 0.94, CI = 0.34–2.66 b=−0.05, SE = 0.02, Wald χ²(1) = 3.51, p = .06, OR = 0.96, CI = 0.91–1.00

ATs Gender Age

ATs = Autistic Traits.

3.5. Research question 5: underemployment Skilled and time-related underemployment were analysed as separate dependent variables to determine whether gender, ATs or age were predictors; seen in Table 5. The regression models concerning both skilled and time-related underemployment were not significant; i.e., neither gender, ATs or age predicted underemployment.

3.6. Research question 6: employment type Exploring if gender, ATs, or age predicted casual, part-time and full-time employment, each were considered separately. The 56

Research in Autism Spectrum Disorders 56 (2018) 50–60

S.M. Hayward et al.

regression models were not significant for either casual or part-time employment. Yet, a significant model for full-time employment was found with a single predictor, ATs. Therefore, while neither gender, ATs nor age predicted casual or part-time employment, HATs decreased the odds of working full-time after controlling for gender and age. 4. Discussion This is the first study to investigate whether employment experiences are influenced by gender or ATs while controlling for confounding variables; such as education and age. This research also contributes to the current literature by examining time-related underemployment among women with HATs. Overall, although no gender differences were found, the findings reported here largely support those of published literature among those with HATs generally without reference to a comparison group. As such, it is suggested that the results reported here could possibly be generalised beyond the sample obtained despite differences between the present and largest known study of employment and people with HATs by Aspect (2013; N=313). The comparable article by Aspect (2013) was conducted in the same country and obtained participants in a similar manner to the present study. While disaggregated data by gender was not reported in the other paper, it appeared that the present sample was slightly older (Mage = 39 vs 34 years). In conjunction with this, by comparison the current study’s participants were better educated, 63% vs 13% respectively had post-secondary education, and were more likely to be employed at the time of the survey (74% vs 55%). In addition, participants with HATs in this study were diagnosed with autism later than the comparable study (33 vs 23 years of age). It is possible that because the other paper is affiliated with support services for people with HATs they might have been more likely to obtain a greater impaired sample. Those diagnosed older may be more likely to have ‘mild’ ATs (Johnson & Joshi, 2016). Alternatively, level of psychosocial functioning may moderate labour market experiences given evidence of adults with HATs who have greater cognitive ability (Happe et al., 2016) and psychosocial function are diagnosed later (Lehnhardt et al., 2012). This may also be why most of the present sample had employment, although psychosocial function was not measured by this study. If the present sample was ‘less affected’ or better able to manage their autism symptomology, this could explain the why unemployment was equally an issue among those with HATs and TD individuals despite differences in post-secondary educational attainment. Those with HATs in this study were less likely to have attained post-secondary education. However, higher education may protect against unemployment (Ohl et al., 2017). Other researchers, not controlling for age of diagnosis, education, and without a control group, have reported unemployment as problematic for those with HATs (e.g., Griffith et al., 2012; Hurlbutt & Chalmers, 2004). Yet, evidence of episodic unemployment for women with HATs was evident in this paper within the theme of unstable employment, defined as not always having work or employment that has not been ongoing. These results relate to Müller et al. (2003) findings among those with HATs generally. HATs were a significant factor to most employment experiences in the present research. Further, social-communication skills are a primary barrier identified in the broader literature as impacting employment (Baldwin et al., 2014; Griffith et al., 2012; Hurlbutt & Chalmers, 2004). This supports the importance of social-communication skills in the labour market (cf. Agran et al., 2016; Charney, 2016; Jones et al., 2016; Lin & Kwantes, 2015). This is further evidenced by responses of people with HATs stating social-communication barriers in relation to unfavourable occupational experiences in the present research. Unfavourability in the labour market by participants with HATs stemmed from inadequate interpersonal skills and workplace bullying; occupational difficulties likewise noted by others (cf. Autism Spectrum Australia Aspect, 2013; Baldwin et al., 2014; Griffith et al., 2012; Hurlbutt & Chalmers, 2004). This supports the social model of disability as a framework for intervention, meaning that the barriers for women with HATs may be a product of the social expectations placed on them by others. If those working along-side people with HATs placed less emphasis on social-communication skills or were understanding and provided support with social-communication challenges these barriers may be overcome. Nevertheless, job instability of women with HATs coupled with challenges gaining and maintaining employment evidenced in this research, and people with HATs generally (cf. Griffith et al., 2012; Hurlbutt & Chalmers, 2004; Müller et al., 2003; Ohl et al., 2017), may impact career progression. Difficulty advancing in employment is also noted for those with other hidden disabilities, i.e., disabilities which cannot be seen (Sang, Richards, & Marks, 2016). Thus, it appears that structural oppression (Ritzer & Goodman, 2008; Young, 2011) might apply to women with HATs where employment advancement impinges upon their capability to hold stable employment to which they appear already disadvantaged. Therefore, job-person fit is imperative given it impacts employment experiences. Vocational support services to achieve this may be required. Adding to this, the present study supports the idea of stunted career growth for those with HATs (versus TD individuals) as they possessed decreased odds to working full-time. According to Human Capital theory (Becker, 1964), opportunities for advancement might be more common in full-time roles as organisations could view these employees as a more worthwhile investment of resources. Although, women with HATs may possess additional hurdles to working full-time and disrupted employment owing to family commitments effectively reducing their human capital potential. While this was not accounted in this research, these issues have been noted for women generally (McGraw et al., 2011) and require further investigation for women with HATs. However, those with HATs might be less likely to choose full-time occupations. This may be evidenced by their narrative responses and lack of time-related underemployment in the present research. Yet, reasons for this requires thorough investigation. Interview methods could aid further inquiry into employment issues which would allow clarity and reasoning of responses to be obtained that were not possible in the current research design. This is also a limitation of the present research, as only salient issues would have been reported by participants. Further, only clear responses were coded and analysed. It is possible that these limitations impacted upon the lack of gender disparities in the present research. However, as no differences were found, this could point to evidence of intersectionality of men with HATs; it was expected that women with HATs would 57

Research in Autism Spectrum Disorders 56 (2018) 50–60

S.M. Hayward et al.

fare worse, like women generally. Thus, it is important to replicate this research with a greater (matched) sample given both size limitations in the present study and differences between women and men with HATs which were not able to be controlled. For example, women with HATs were younger than men with HATs, and diagnosed with autism earlier. As coping strategies are suggested to develop with increased age (Happe et al., 2016) this may have protected the men in this sample. Yet, other authors using predominately men with HATs in their samples have cited skilled underemployment as problematic for individuals with HATs (Baldwin & Costley, 2016; Baldwin et al., 2014; Müller et al., 2003). Conversely, others have not evidenced this (Ohl et al., 2017). The current study supports Ohl et al.’s findings in that there was no indication of skilled underemployment of those with HATs (vs those TD), which may be owing to control group comparisons. To illustrate, a comparable study utilising an Australian sample evidenced skilled underemployment by contrasting a sub-population of women with HATs to Australian women generally (Baldwin & Costley, 2016). As it was not stipulated that the women with HATs were a representative sample, these two populations cannot be readily compared particularly without consideration of confounding factors. Alternatively, perhaps owing to the conservative approach used to compare skill level between qualification/s and role/s, skilled underemployment may have been missed by the present study. Others utilising similar techniques inquired into the role held and duties undertaken (cf. Baldwin et al., 2014) which may have clarified ambiguous skill levels of the role/s underemployed in. Further, as employment challenges could negatively impact health and well-being (Blustein et al., 2013; McKee-Ryan et al., 2005; Rosenthal et al., 2012), this could in turn exacerbate difficulties maintaining employment and possibly contribute to underemployment. Thus, this requires further investigation. 4.1. Limitations The participants that were excluded from analyses on the basis of incongruities between self-reported diagnostic status and AQ score may have resulted in a sample that represented extremes in ATs; high or low. Although it is suggested that the present HAT sample were ‘less affected’ or better at managing their autism symptomology, this group may have been more self-aware of their symptoms compared to those that reported a diagnosis of autism who scored below the AQ cut-off. Hence, these sub-samples may possess differing challenges when compared to their TD peers. Another limitation relates to online methods of data collection which have been criticised for sampling bias (Blasius & Brandt, 2010; Sue & Ritter, 2012); likewise, recruitment of participants using social media (Fowler, 2009; Özguven & Mucan, 2013; Stern, Bilgen, McClain, & Hunscher, 2017). While social media is an effective method for reaching people with HATs (Haas et al., 2016), it has been suggested to attract very young (18–24 years old) or older participants (50 years and above; Stern et al., 2017), as well as those with post-secondary education (Fowler, 2009; Özguven & Mucan, 2013; Stern et al., 2017). These characteristics appear somewhat reflected in the present sample and may have impacted upon the comparisons made. 4.2. Implications The occupational relationships of people with HATs need to be understanding and accepting of their social-communication challenges for employment success. However, job-person fit is imperative for individuals with HATs. This could be facilitated with specialised autism vocational support services which might provide a catalyst to full-time employment. 5. Conclusion No known study has investigated the possible relationship between gender, ATs, and workplace experiences in relation to a reference group while controlling for education or age. It was determined by this research that ATs are more important than gender when considering workplace intervention. Women with HATs (vs TD women) were more likely to report: unfavourable workplace experiences, unstable employment, challenges gaining and maintaining employment, and were less likely to work full-time. The impact of these vocational difficulties could stunt the career prospects for these otherwise skilled individuals while negatively affecting well-being. To increase social inclusion of women with HATs in the labour market, it is proposed that employers place greater importance on technical ability rather than social-communication skills when hiring and supervising women with HATs, and vocational services are provided to achieve job-person fit. However, these issues require further consideration in a larger programme of research. Conflict of interest The authors declare there are no known conflicts of interest. Funding The first author is supported by an Australian Government Research Training Program Scholarship. Acknowledgements The authors thank Dr. Anna-Maria Klas who double coded data for the project. 58

Research in Autism Spectrum Disorders 56 (2018) 50–60

S.M. Hayward et al.

References ABS (2014). Australian labour market statistics, July 2014. (Cat. No. 6105.0). Canberra, Australia: ABS. ABS (2015). Disability and labour force participation, 2012 (Cat. No. 4433.0.55.006). Canberra, Australia: ABS. ABS (2017a). Gender indicators, Australia, September 2017 (Cat. No. 4125.0). Canberra, Australia: ABS. ABS (2017b). Labour force, Australia, November 2016 (Cat. No. 6202.0). Canberra, Australia: ABS. Agran, M., Hughes, C., Thoma, C. A., & Scott, L. A. (2016). Employment social skills: What skills are really valued? Career Development and Transition for Exceptional Individuals, 39(2), 111–120. American Psychiatric Association [APA] (2013). DSM-V diagnostic and statistical manual of mental disorders. Arlington: VA: American Psychiatric Association. Australian Bureau of Statistics [ABS] (2009). Employment arrangements, retirement and superannuation, Australia, April to July 2007 (re-issue). Canberra, Australia: ABS. Autism Spectrum Australia [Aspect] (2013). We belong: The experiences, aspirations and needs of adults with asperger’s disorder and high functioning autism. Baio, J. (2014). Prevalence of autism spectrum disorder among children aged 8 years - autism and developmental disabilities monitoring network, 11 Sites, United States, 2010. Morbidity and Mortality Weekly Report Surveillance Summaries, 63(2), 1–21. Baldwin, S., & Costley, D. (2016). The experiences and needs of female adults with high-functioning autism spectrum disorder. Autism: The International Journal of Research and Practice, 20(4), 483–495. https://doi.org/10.1177/1362361315590805. Baldwin, S., Costley, D., & Warren, A. (2014). Employment activities and experiences of adults with high-functioning autism and Asperger’s disorder. Journal of Autism and Developmental Disorders, 44(10), 2440–2449. https://doi.org/10.1007/s10803-014-2112-z. Baron-Cohen, S., Wheelwright, S., Skinner, R., Martin, J., & Clubley, E. (2001). The autism-spectrum quotient (AQ): Evidence from Asperger syndrome/high-functioning autism, males and females, scientists and mathematicians. Journal of Autism and Developmental Disorders, 31(1), 5–17. Becker, G. S. (1964). Human capital; a theoretical and empirical analysis, with special reference to education. New York: National Bureau of Economic Research; distributed by Columbia University Press. Benford, P. (2008). The use of internet-based communication by people with autism (Doctoral dissertation) fromhttp://search.proquest.com.ezproxy-b.deakin.edu.au/ docview/898749950. Blasius, J., & Brandt, M. (2010). Representativeness in online surveys through stratified samples. Bulletin of Sociological Methodology, 107(1), 5–21. https://doi.org/10. 1177/0759106310369964. Blustein, D. L., Kozan, S., & Connors-Kellgren, A. (2013). Unemployment and underemployment: A narrative analysis about loss. Journal of Vocational Behavior, 82(3), 256–265. Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), 77–101. Broadbent, J., Galic, I., & Stokes, M. A. (2013). Validation of autism spectrum quotient adult version in an Australian sample. Autism Research & Treatment, 2013, 1–7. https://doi.org/10.1155/2013/984205. Brooks, W. T. (2016). Gender differences in social skills, peer relationships, and emotional correlates in adults with high functioning autism spectrum disorders (Doctoral dissertation) fromhttp://ezproxy.deakin.edu.au/login?url=http://search.ebscohost.com/login.aspx?direct=true&db=psyh&AN=2016-17133-254&site=ehostlive&scope=site. Charney, D. (2016). Wanted: Workers with soft skills: Supply chain experience may land a job interview, but it’s the soft skills that will land the job. Material Handling & Logistics, 71(9), 38–40. Deepa, S., & Seth, M. (2013). Do soft skills matter? Implications for educators based on recruiters’ perspective. IUP Journal of Soft Skills, 7(1), 7–20. Doren, B., Gau, J. M., & Lindstrom, L. (2011). The role of gender in the long-term employment outcomes of young adults with disabilities. Journal of Vocational Rehabilitation, 34(1), 35–42. Duffy, F. (2009). Unemployment and underemployment. Tolendo: Great Neck Publishing. England, P. (2005). Gender inequality in labor markets: The role of motherhood and segregation. Social Politics: International Studies in Gender, State & Society, 12(2), 264–288. https://doi.org/10.1093/sp/jxi014. Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39(2), 175–191. Folbre, N. (1994). Who pays for the kids? Gender and the structure of constraint. London: New York: Routledge. Fowler, F. J. (2009). Survey research methods [electronic resource] (4th ed.). Thousand Oaks, California. London: Sage Publications. Gal, E., Landes, E., & Katz, N. (2015). Work performance skills in adults with and without high functioning autism spectrum disorders (HFASD). Research in Autism Spectrum Disorders, 10, 71–77. Gillespie-Lynch, K., Kapp, S. K., Shane-Simpson, C., Shane-Smith, D., & Hutman, T. (2014). Intersections between the autism spectrum and the internet: Perceived benefits and preferred functions of computer-mediated communication. Intellectual and Developmental Disabilities, 52(6), 456–469. https://doi.org/10.1352/19349556-52.6.456. Gray, D. E. (2014). Doing research in the real world (3rd ed.). UK: Sage. Griffith, G. M., Totsika, V., Nash, S., & Hastings, R. P. (2012). ’I just don’t fit anywhere’: Support experiences and future support needs of individuals with Asperger syndrome in middle adulthood. Autism: The International Journal of Research and Practice, 16(5), 532–546. https://doi.org/10.1177/1362361311405223. Haas, K., Costley, D., Falkmer, M., Richdale, A., Sofronoff, K., & Falkmer, T. (2016). Factors influencing the research participation of adults with autism spectrum disorders. Journal of Autism & Developmental Disorders, 46(5), 1793–1805. https://doi.org/10.1007/s10803-016-2708-6. Happe, F. G., Mansour, H., Barrett, P., Brown, T., Abbott, P., & Charlton, R. A. (2016). Demographic and cognitive profile of individuals seeking a diagnosis of autism spectrum disorder in adulthood. Journal of Autism and Developmental Disorders, 5(11), 3469. https://doi.org/10.1007/s10803-016-2886-2. Hauser, P. M. (1974). The measurement of labor utilization. Malayan Economic Review, 19(2), 1–17. Head, A. M., McGillivray, J. A., & Stokes, M. A. (2014). Gender differences in emotionality and sociability in children with autism spectrum disorders. Molecular Autism, 5(1), 1–20. https://doi.org/10.1186/2040-2392-5-19. Hoobler, J. M., Wayne, S. J., & Lemmon, G. (2009). Bosses’ perceptions of family-work conflict and women’s promotability: Glass ceiling effects. Academy of Management Journal, 52(5), 939–957. https://doi.org/10.5465/AMJ.2009.44633700. Hurlbutt, K., & Chalmers, L. (2004). Employment and adults with Asperger syndrome. Focus on Autism & Other Developmental Disabilities, 19(4), 215–222. Johnson, T. D., & Joshi, A. (2016). Dark clouds or silver linings? A stigma threat perspective on the implications of an autism diagnosis for workplace well-being. Journal of Applied Psychology, 101(3), 430–449. https://doi.org/10.1037/apl0000058. Jones, M., Baldi, C., Phillips, C., & Waikar, A. (2016). The hard truth about soft skills: What recruiters look for in business graduates. College Student Journal, 50(3), 422–428. Joshi, A., Jooyeon, S. O. N., & Hyuntak, R. H. (2015). When can women close the gap? A meta-analytic test of sex differences in performance and rewards. Academy of Management Journal, 58(5), 1516–1545. https://doi.org/10.5465/amj.2013.0721. Lai, M.-C., Lombardo, M. V., Pasco, G., Ruigrok, A. N. V., Wheelwright, S. J., Sadek, S. A., et al. (2011). A behavioral comparison of male and female adults with high functioning autism spectrum conditions. PLoS One, 6(6), 1–10. Lau, W. Y. P., Kelly, A. B., & Peterson, C. C. (2013). Further evidence on the factorial structure of the autism spectrum quotient (AQ) for adults with and without a clinical diagnosis of autism. Journal of Autism & Developmental Disorders, 43(12), 2807–2815. Lehnhardt, F. G., Gawronski, A., Volpert, K., Schilbach, L., Tepest, R., & Vogeley, K. (2012). Psychosocial functioning of adults with late diagnosed autism spectrum disorders-a retrospective study. Fortschritte Der Neurologie-Psychiatrie, 80(2), 88–97. https://doi.org/10.1055/s-0031-1281642. Lin, I. Y., & Kwantes, C. T. (2015). Potential job facilitation benefits of “water cooler” conversations: The importance of social interactions in the workplace. The Journal of psychology, 149(3), 239–262. McGraw, P. A., Kramar, R., & McGraw, P. (2011). Women at work: Research, policy, and practice (1st ed.). Prahran, Vic: Tilde University Press. McKee-Ryan, F., Song, Z., Wanberg, C. R., & Kinicki, A. J. (2005). Psychological and physical well-being during unemployment: A meta-analytic study. The Journal of

59

Research in Autism Spectrum Disorders 56 (2018) 50–60

S.M. Hayward et al.

Applied Psychology, 90(1), 53–76. Müller, E., Schuler, A., Burton, B. A., & Yates, G. B. (2003). Meeting the vocational support needs of individuals with Asperger syndrome and other autism spectrum disabilities. Journal of Vocational Rehabilitation, 18(3), 163–175. Ohl, A., Sheff, M. G., Little, S., Nguyen, J., Paskor, K., & Zanjirian, A. (2017). Predictors of employment status among adults with autism spectrum disorder. Work, 56(2), 345–355. https://doi.org/10.3233/WOR-172492. Oliver, M. (1983). Social work with disabled people. Basingstoke: Macmillan. Özguven, N., & Mucan, B. (2013). The relationship between personality traits and social media use. Social Behavior & Personality: An International Journal, 41(3), 517–528. QSR International Pty Ltd (2015). NVivo qualitative data analysis Software; Version 11. Ritzer, G., & Goodman, D. J. (2008). Sociological theory (7th ed.). New York, London: McGraw-Hill. Rosenthal, L., Carroll-Scott, A., Earnshaw, V. A., Santilli, A., & Ickovics, J. R. (2012). The importance of full-time work for urban adults’ mental and physical health. Social Science and Medicine, 75(9), 1692–1696. Ruzich, E., Allison, C., Smith, P., Watson, P., Auyeung, B., Ring, H., et al. (2015). Measuring autistic traits in the general population: A systematic review of the autismspectrum quotient (AQ) in a nonclinical population sample of 6,900 typical adult males and females. Molecular Autism, 6(1), 41–66. https://doi.org/10.1186/ 2040-2392-6-2. Sang, K. J. C., Richards, J., & Marks, A. (2016). Gender and disability in male-dominated occupations: A social relational model. Gender, Work and Organization, 23(6), 566. Schofield, D. J., Shrestha, R. N., Percival, R., Passey, M. E., Callander, E. J., & Kelly, S. J. (2011). The personal and national costs of mental health conditions: Impacts on income, taxes, government support payments due to lost labour force participation. BMC Psychiatry, 11(1), 72–88. Smith, J. (2013). The 20 people skills you need to succeed at work. Retrieved February 13, 2015, fromhttp://www.forbes.com/sites/jacquelynsmith/2013/11/15/the-20people-skills-you-need-to-succeed-at-work/2/. Stern, M. J., Bilgen, I., McClain, C., & Hunscher, B. (2017). Effective sampling from social media sites and search engines for web surveys: Demographic and data quality differences in surveys of google and Facebook users. Social Science Computer Review, 35(6), 713–732. https://doi.org/10.1177/0894439316683344. Sue, V. M., & Ritter, L. A. (2012). Conducting online surveys (2nd ed.). Thousand Oaks, California: Sage. Trewin, D., & Pink, B. (2006). Australian and New Zealand standard classification of occupations: ANZSCO (1st ed.). Canberra, ACT, Wellington: Australian Bureau of Statistics, Statistics New Zealand. Willig, C. (2013). Introducing qualitative research in psychology (3rd ed.). Maidenhead: Open University Press. Woodbury-Smith, M. R., Robinson, J., Wheelwright, S., & Baron-Cohen, S. (2005). Screening adults for Asperger syndrome using the AQ: A preliminary study of its diagnostic validity in clinical practice. Journal of Autism & Developmental Disorders, 35(3), 331–335. Workplace Gender Equality Agency (2014). Gender composition of the workforce: By industry. Retrieved 12 December, 2015, fromhttps://www.wgea.gov.au/sites/ default/files/2014-04-04-Gender%20composition-of-the-workforce-by-industry.pdf. Young, I. M. (2011). Justice and the politics of difference. Princeton, N.J: Princeton University Press.

60