Introduction
Disabled people face considerable barriers to labour market participation, which is reflected in their marginal position in the labour market, where disabled people are about 40% less likely to be in employment than people without a disability (OECD 2022). This enduring and large employment gap has been described by many as an employment puzzle, as many welfare states have, during the past two decades, made considerable investments in active labour market policies and implemented legal protection against the employment discrimination of disabled people. Still, despite these policies and political intentions, employment levels remain largely unchanged (OECD 2022).
When considering the employment gap, an increased recognition of the role of the demand side – the employers – has been observed in the literature. Within this research field, employers’ hiring attitudes towards disabled job seekers have been thoroughly documented. We know from numerous correspondence experiments conducted in various contexts that disabled people meet high levels of discrimination in the recruitment process (Ameri et al. 2018; Baert 2016; Baert 2018; Bellemare et al. 2018; Bjørnshagen 2021; Bjørnshagen and Ugreninov 2021; Hipes et al. 2016; Ravaud et al. 1992). These studies find that when an impairment is described in a job application, the job seeker has a much lower likelihood of being called into a job interview. Disability disadvantage in employers’ hiring attitudes has also been found in survey experiments, although to a varying degree, depending on the type of impairment (Andersson et al. 2015; Berre 2023; Krogh and Breedgaard 2022). Qualitative studies have further explored what can explain this hiring disadvantage and revealed that employers hold low expectations about the abilities of disabled people (Chhabra 2021), concider disabled people as less likely to socially fit into the group (Østerud 2023), and hold stereotypes about disabled people as less productive and costly for the company (Iwanaga et al. 2021; Nagtegaal et al. 2023).
The above research highlights the relevance of policies that target employers’ willingness to employ disabled people, also known as employer-focussed interventions (Derbyshire et al. 2024). Financial incentive schemes, such as wage subsidies and accommodation grants are examples of such policies. However, when employers are presented with these support measures in survey or correspondence experiments, the results tend to find that they have no, or limited impact, on their hiring interest in disabled candidates (Baert 2016; Berre 2023; Krogh and Breedgaard 2022). This is despite the assertions of the employers themselves, when asked directly, which emphasises the importance of these support measures when recruiting disabled people (Mandal et al. 2019). Previous findings, therefore, suggest a discrepancy between what employers say and do when it comes to their attitudes towards support measures in the hiring of disabled job seekers – and it is unclear why this is so.
However, employers operate within vastly different industry contexts, each with unique characteristics. The question of how these differences in context impact their hiring attitudes towards disabled candidates is still underexplored by field and survey experiments. Some growing empirical evidence does suggest that the industry or occupational context matters when measuring disability disadvantage in hiring (Alm Andreassen 2009; Bjørnshagen, Rooth and Ugreninov 2025). However, these studies are few, and no field or survey experiments have yet compared differences between types of disability in how industry impacts the degree of hiring disadvantage measured. Focussing on overall industry differences for just one type of disability might obscure underlying differences between different types of disability. Neither has any study, that this author is aware of, explored industry differences in the impact of public support measures, such as wage subsidies and accommodation grants, on employers’ openness to hire disabled job seekers. Focussing on the general impact of support measures on employers’ hiring attitudes across different industry contexts might obscure underlying differences between them.
Increased knowledge of these underexplored areas of the literature would contribute to a better understanding of disabled people’s hiring disadvantage in working life and the potential role of support measures in decreasing this disadvantage. Building on existing research, this article will, therefore, relying on data from a factorial survey presented to Norwegian employers (N = 951), explore how the industry context impacts employers’ hiring assessments of disabled candidates, compared to their assessments of job seekers without a disclosed disability. It will also explore how the industry context impacts the degree to which informing employers about publicly financed support measures makes them more open to hiring disabled job seekers. Companies operating within the same industry are assumed to have similar characteristics regarding core work tasks, degree of organizational flexibility, productivity demands, etc. Each industry has unique combinations of these characteristics. This makes industry a relevant and important factor to consider in studies on employers’ hiring attitudes towards disabled candidates and how public support measures impact these.
Context: Norway
Norway is considered to have a well-functioning labour market with an overall low unemployment rate among the general population and high investments in active labour market policies. These seemingly good conditions for promoting the labour market integration of disabled people have, however, not been reflected in this group’s employment rate during the past two decades. Disabled people in Norway, which constitute about 17% of the population, have had a stable and low-level employment rate (37.5% in 2021) in comparison to the much higher employment rate of the non-disabled part of the population (78.4% in 2021) (Statistics Norway 2022). The OECD (2006, 14) has described Norway’s main challenge regarding disability and employment as being ‘to understand why the existing frameworks, which look good, are not delivering’. Some have suggested that the Norwegian labour market’s compressed wage structure and high productivity make employers reluctant to hire someone perceived as having lower productivity (Halvorsen, Hvinden and Schoyen 2015; Hauge 2021). Also, the main rule in Norway is that employees must be hired on permanent employment contracts, which is regulated by law, with 82% of all employees being on permanent contracts in 2022 (Statistics Norway 2023). The ‘strictness’ of regulations concerning the dismissal of permanent employees is furthermore above the OECD average (OECD 2024).
Norway’s Working Environment Act and Equality and Anti-Discrimination Act prohibit disability discrimination in employment. Employers are also legally required to provide individual accommodation for disabled job seekers and employees, as long as this does not impose an undue burden on the employer. Norway, which is characterised by tripartite cooperation between the government, employers’ associations, and trade unions, has though relied on a strategy based on ‘consensus, persuasion, and voluntariness’ when engaging employers in the long-standing policy aim of increasing the labour market participation of vulnerable groups (Midtsundstad 2008, 17). The use of job placement programmes, financial incentive schemes, and other support measures have been central in this policy strategy. However, there has been a reluctance to implement more enforced employer engagement and no mandatory quota schemes have therefore been introduced (Halvorsen and Hvinden 2014).
Industry differences in Disabled People’s Hiring Disadvantage
Labour markets consist of employers operating within a wide range of different industries. As Alm Andreassen (2009, 90) points out: ‘The question is not whether differences exist, but whether these are significant for the employment of disabled people and should have consequences for public policy’. Correspondence experiments have thoroughly documented high levels of disability discrimination in the recruitment process. However, in most of these studies, a comparison between industries or professions has not been an area of focus.
A recent correspondence experiment, where 2,048 job applications with randomly assigned information about disability were sent to Swedish employers, is a welcomed exception. This study found that discrimination towards wheelchair users varied considerably depending on the type of occupation, from no discrimination to considerable discrimination in occupations where one could assume that the impairment would not affect productivity (Bjørnshagen, Rooth and Ugreninov 2025). Alm Andreassen (2009) also presented descriptive findings from an employer survey, which showed that the proportion of employers who self-reported that they found it ‘hard’ or ‘impossible’ to employ disabled people varied according to the type of industry. In conclusion, both studies point to profession or industry differences in the level of disability disadvantage in hiring.
Industry differences in the hiring disadvantage of disabled people can also be expected from a theoretical point of view. From what she describes as a materialist feminist understanding of disability, Garland-Thomson (2009, 592) introduces the concept of ‘misfit’ and the situation of ‘misfitting’ as a way of understanding how bodies and minds, that are understood to be impaired, interact with their environments: ‘fitting and misfitting denote an encounter in which two things come together in either harmony or disjunction,’ she describes. The concept of misfitting is arguably useful in understanding both expected and unexpected industry differences in employers’ hiring interest in disabled job seekers. The concept implies that, within all industries, distinct roles have been created with a set of expectations and requirements that must be met for a job seeker to be considered a good ‘fit’ (Lid and Grue 2018). These roles have, however, not been constructed with a diverse population in mind when it comes to people’s functional abilities and work capacity. They have rather been constructed around an ‘ideal worker’ – someone who can handle a diverse set of work tasks, work long hours and be work-ready at any given time (Foster and Wass 2013). This has led to expectations within working life that materially disadvantage people without the majority body – thereby creating enduring structural barriers to employment for disabled job seekers.
However, disabled people’s status as misfits is inherently unstable and context-dependent and can therefore depend on both the type of disability and the industry context (Garland-Thomson 2009). A wheelchair user, as an illustrative example, would be expected to be a larger misfit within industries with physically demanding core work tasks, such as construction and residential care, compared to industries with less physical demanding work, such as public administration and education. However, employers’ hiring attitudes can also be influenced by other factors unrelated to disabled people’s expected ability to perform specific work tasks. This can be why the previously mentioned Swedish correspondence study found that the occurrence of disability discrimination varied considerably across occupations, despite the impairment being assumed to not impact productivity within any of these occupations (Bjørnshagen, Rooth and Ugreninov 2025). Examples of these ‘other factors’ can include an organisation’s management resources, vulnerability to sickness absence, workload, work task flexibility, cultural values, and work environment characteristics. The degree to which these factors are emphasised by the employers when choosing who to hire may vary between seemingly similar industries, in terms of the nature of their core work tasks, and produce differences between them in their assessments of disabled candidates. We can, therefore, assume that the present study finds both expected (based on the expected abilities needed to perform core work tasks) and unexpected (caused by other factors, of which some are mentioned above) differences between industries in their assessments of disabled candidates.
Based on the aforementioned empirical studies and the theory of misfitting, this study expects to find that:
H1) The type of industry impacts employers’ hiring attitudes towards disabled candidates.
H2) How and the extent to which the industry context impacts employers’ hiring attitudes depends on the type of disability.
The impact of support measures on employers’ hiring attitudes
From a policy perspective, there already exist provisions within many welfare states that can, in theory, facilitate the hiring of disabled people (Foster and Wass 2013). The most common among these provisions is the legal requirement to make reasonable accommodations when hiring disabled workers. This requirement directly challenges the match between a job seeker and any given job specification, as it requires employers to demonstrate ‘functional, task, person, and organizational flexibility’ when hiring new employees (Foster and Wass 2013, 706). However, Foster and Wass (2013) find that employers resist imposing such flexibility within their organizations as such a ‘variation to the standardized criteria would inevitably conflict with established organizational logic’ (Foster and Wass 2013, 709).
Other provisions, which can facilitate the hiring of disabled candidates, are publicly financed support measures designed to compensate employers for accommodation costs. In Norway, these include accommodation grants, functional assistance, sign language interpreting services, mentoring schemes, wage subsidy schemes and more. In the design of these support measures lies, implicitly at least, an acknowledgement that a mismatch between the employee and the job design or physical work environment can occur when a person with an impairment enters working life. When asked directly, the employers themselves emphasise the need for these measures (Mandal et al. 2019).
However, when publicly financed support measures are introduced to employers in field or survey experiments, the results tend to show that employers continue to resist hiring disabled job seekers (Baert 2016; Berre 2023; Shamshiri-Petersen and Krogh 2020). A correspondence study conducted in Belgium found that mentioning entitlement to a 20–40% wage subsidy scheme did not influence the likelihood of a positive response from employers when disabled candidates applied for a job (Baert 2016). A vignette experiment revealed that informing Danish employers about the possibility of applying for financial compensation only slightly increased their willingness to hire a wheelchair user (Shamshiri-Petersen and Krogh 2020). In a factorial survey, Berre (2023) found that support measures had limited or no impact on employers’ hiring assessments of disabled job seekers. Evaluation studies in Norway have also found that these support measures have, more than to facilitate the recruitment of disabled job seekers, primarily been used for workers who are already employed in a company and subsequently acquire a disability (Amundsveen and Solvoll 2003; Berre 2023; Econ 2008; Rambøll 2008).
The studies presented above are still few, and no attempt has yet been made to compare the impact of support measures across different occupations or industries. As disabled people’s status as misfits is highly context-dependent (Garland-Thompson 2009), one could expect support measures to have different impacts within different industries. The impact of sign language interpreting services, as an example, would be expected to be larger in industries where a hearing impairment would limit someone from performing core work tasks but enable the same person to fully do the job with an interpreter by their side, and to be smaller within industries where verbal communication is a less important part of the job.
Given the lack of prior studies in this area, this study relies on the materialist feminist concept of misfit to hypothesize that:
H3) the impact of support measures on employers’ hiring attitudes towards disabled job seekers depends on the industry context.
Research design
This study relies on data from a factorial survey (N = 1341) conducted among a stratified random sample of Norwegian employers. The same data was also used in Berre’s (2023) article ‘Exploring Disability Disadvantage in Hiring: A Factorial Survey among Norwegian Employers’, and the present article is a continuation of this work.
Factorial surveys are commonly used in studies of employers’ hiring attitudes (McDonald 2019). They involve presenting fictive descriptions of hypothetical job seekers to a sample of employers, who are then asked to evaluate these on a scale. The employers are however unaware that several job seeker characteristics are simultaneously manipulated in these descriptions, making it ‘possible to identify how specific characteristics, or combinations of these, impact respondents’ assessments’ (Berre 2023, 1092).
When well designed, a factorial survey is believed to decrease social desirability bias (Auspurg and Hinz 2015; Berre 2023; McDonald 2019; Wallander 2009), a known impact in studies where employers self-report their attitudes in surveys or interviews. Another crucial advantage of the method is its multidimensionality (Auspurg and Hinz 2015; Berre 2023), where numerous test variables can be included in the same research design, thereby allowing an exploration of these and the interactions between them – which made it possible to pursue the research questions of the present study.
Construction of vignettes
Nine vignettes (fictive descriptions of job seekers) were presented to each employer participating in the survey. They were asked to give their ‘honest opinion’ of the presented profiles and assume that the job seekers were all ‘highly motivated to do a good job for the company’ (Berre 2023, 1095). They were then asked to evaluate each job seeker on a scale of 0–10, where 0 represented ‘highly unlikely to be hired’ and 10 represented ‘highly likely to be hired’ (Berre 2023, 1095).
The vignettes were constructed based on seven variables, called dimensions, which were: gender, age, education, work experience, disability, support measure, and work percentage. These dimensions represent common characteristics for job seekers to present to employers when applying for jobs and other characteristics believed to impact employers’ assessments (Berre 2023). Tables 1, 2, and 3 describe all the dimensions and their levels in detail.
Table 1
Dimensions and levels of the vignettes.
| DIMENSION | LEVEL |
|---|---|
| Gender | Male |
| Female | |
| Age | 26–28 |
| 40–42 | |
| 54–56 | |
| Education | Primary school |
| Secondary school | |
| Education relevant to the needs of the company | |
| Work experience | Some relevant work experience |
| Many years of relevant work experience | |
| Disability | Non-disabled (nothing mentioned in the text) |
| Wheelchair user | |
| Blind | |
| Intellectual disability | |
| Chronic illness | |
| Mental illness | |
| Hearing impairment | |
| Support measure | No measure (nothing mentioned) |
| Measure specific to type of disability | |
| Work percentage | Full time (nothing mentioned) |
| 20–30% | |
| 50% | |
| 70–80% |
[i] Note: This table has previously been published in Berre (2023): https://doi.org/10.1177/09500170231175776.
Table 2
Full description of the different types of disability.
| Wheelchair user | (Name) is a wheelchair user and could have some accommodation needs related to this. |
| Hearing impaired | (Name) has a hearing impairment. She communicates with speech and can understand what is being said in one-to-one conversations in quiet environments but must use sign language interpreters in contexts with many people present. |
| Blind | (Name) is blind and uses a guide dog. |
| Intellectual disability | (Name) has an intellectual disability/Down syndrome* and needs some extra follow-up and training in work tasks. |
| Chronic/mental illness | (Name) has a chronic/mental* illness and could have periods with reduced work capacity because of this. |
[i] Notes: *The respondent randomly received one of the two descriptions marked in italic. For intellectually disabled candidates, the two different phrasings of the impairment (‘intellectually disabled’ vs. ‘Down syndrome’) did not impact employers’ assessments and have therefore been kept as one category throughout all analyses. Chronically or mentally ill candidates have, however, been kept as two separate categories throughout all analyses due to being very different types of impairments and also because employers responded differently towards these two groups. This table has previously been published in Berre (2023): https://doi.org/10.1177/09500170231175776.
Table 3
Full description of support measures specific to the type of disability.
| Wheelchair user | ‘NAV has support that covers the company’s expenses in case of accommodation needs’ |
| Hearing impaired | ‘NAV organizes and covers all costs associated with the use of sign language interpreters’ |
| Blind | ‘NAV has support that covers the company’s expenses in case of accommodation needs related to the vision impairment’ |
| Intellectual disability | ‘If hired, NAV will offer the company to cover 80% of (Name)’s salary costs, so that the employer pays only 20%’ |
| Chronic/mental illness | ‘NAV will offer the company a permanent wage subsidy that covers 50% of (Name)’s salary costs’ |
[i] Notes: NAV = Norwegian Labour and Welfare Administration. This table has previously been published in Berre (2023): https://doi.org/10.1177/09500170231175776.
In constructing the vignettes, the dimensions of gender, age, education, and work experience varied randomly in whichever level was chosen. The only random component of the disability variable, however, was the order in which they were presented, as each employer evaluated each type of impairment once. This was done to avoid a situation whereby employers would be repeatedly presented with the same impairment description, which could have annoyed the respondents. Due to the very similar descriptions of the mentally and chronically ill job seekers each employer only received one of the two descriptions, which was randomly selected. Support measures were further only presented in vignettes with an impairment description, and it was random whether a support measure accompanied the impairment description or not. Each type of disability had an impairment-specific description of a support measure described in detail in Table 3. Again, this was done to prevent similar texts from being repeatedly presented to employers and to make the vignette descriptions as realistic as possible. It was also random whether the vignettes included a description of part-time work or not. If this was included, which occurred in more or less half of the total vignettes, it was a random level of part-time work (20–30%, 50% or 70–80%) that was presented.
Data collection
Approval by the Norwegian National Committee for Research Ethics in the Social Sciences and the Humanities (NESH) was gained before conducting the factorial survey.1
The data was collected through an online survey between November 2019 and April 2020. A stratified random sample of employers from the Norwegian Brønnøysund Register of Business Enterprises, where all operating businesses in Norway are obliged to register, were contacted and asked to participate in the survey. This stratified random sample consisted of three subpopulations of employers chosen based on their sector and company size: 5,000 public companies, 5,000 private companies with 6–19 employees, and 5,000 private companies with 20 or more employees. They were recruited using physical letters, emails, phone calls and/or SMS text messages. Some of the invited companies turned out to be inactive, and 12,913 employers eventually received an invitation. The invitations included a web address where the employers could log in using their phone, tablet, or computer to participate in the survey. In total 1,490 employers chose to participate in the survey, giving it a response rate of 11.5%. Employers are known to be hard to recruit to surveys due to their busy schedules, and although the response rate is low, it is still within what can be expected and comparable to the response rates of previous employer surveys (Baruch and Holtom 2008; Damelang et al. 2019). Due to partial non-response, a total of 1,341 of the employers ended up evaluating a total of 11,939 vignettes.
A non-response analysis was conducted to test if the employers who participated in the survey systematically differed from those who chose not to participate. In this analysis, the final sample was compared with all the 15,000 companies invited to the survey on three key variables: industry (using the Standard Industrial Classification), company size (fewer than 20 employees/20 employees or more), and sector (private/public). An overall good representation of all branches was found, except for an over-representation of ‘human health and social work activities’ of 8%, and an under-representation of ‘wholesale and retail trade’ of –5%’ (Berre 2023, 1095). The results also showed ‘an over-representation of the public sector of 10% and an equal under-representation of smaller private businesses’ (Berre 2023, 1095). A weighted scale was constructed to adjust for these and other smaller imbalances in the data.
Data analysis
To analyse industry differences in employers’ hiring assessments and the impact of support measures on these, a decision was made to focus only industries represented by more than 50 employers in the data set. The Standard Industrial Classification (SIC) was the starting point in deciding which industries to include. A total of 951 employers from ten industries, who had evaluated 8,404 job seeker profiles, were ultimately included in the final analyses. These industries were public administration, professional, scientific, and technical activities, education, activities of membership organisations, health care services, manufacturing, construction, childcare, the wholesale and retail trade, and accommodation and food service activities. For a full description of these, see Appendix 2.
Employers’ assessments in a factorial survey are not independent of each other but rather nested within each respondent. The data was therefore analysed in linear mixed effects models.2 The models were estimated using Stata 18 software, as described by Mehmetoglu and Jakobsen (2022, 242–267). Since the research questions required to include a cross-level interaction term in all the regression analyses, between disability at level 1 and industry at level 2, a choice was made to specify a random slope for the lower-level variable of the interaction, as recommended by Heisig and Schaeffer (2019).
Two random slope models were made where the coding of the disability variable differed between the two. The dependent variable in both models was employers’ hiring assessment on a scale from 0 (hiring ‘highly unlikely’) to 10 (hiring ‘highly likely’). The independent variables were the vignette dimensions (age, gender, education, work experience, disability, support measure, and part-time work) at level 1 and employer characteristics (industry, sector, and company size) at level 2. An interaction term was created between disability at level 1 and industry at level 2 in both models. This was done to test whether the type of industry impacts the effect of disability on employers’ hiring assessments.
Model 1 was used to test the first hypothesis – whether the type of industry impacts disability disadvantage in general. In this model, the disability variable was recoded into a three-level variable, where the different types of disability were grouped into one category (1 = non-disabled, 2 = disabled job seekers without a support measure mentioned, 3 = disabled job seekers with a support measure mentioned). This was done to measure industry differences in employers’ assessment of disabled candidates in general.
Model 2 was used to answer hypotheses 2 and 3. In this model, the different types of disability were kept at separate levels of the disability variable. This was done to measure the independent effect of each type of impairment, with and without a support measure mentioned in the vignette, on employers’ hiring assessment. The variable was coded: 0 = non-disabled, 1 = chronically ill, 2 = chronically ill with support measure, 3 = mentally ill, 4 = mentally ill with support measure, 5 = wheelchair user, 6 = wheelchair user with support measure, 7 = hearing impaired, 8 = hearing impaired with support measure, 9 = intellectually disabled, 10 = intellectually disabled with support measure, 11 = blind, and 12 = blind with support measure.
Marginal effects were computed to interpret the results, as described by Williams (2012). All marginal effects presented in the results represent by how many scale points (on the hiring assessment scale from 0 = hiring ‘highly unlikely’ to 10 = hiring ‘highly likely’) the predictive margins of the presented group deviate from the predictive margin of the reference group. In answering hypotheses 1 and 2, the predictive margin of non-disabled candidates was used as the reference. In answering hypothesis 3, the predictive margin of each specific type of disability (when no support measure was mentioned in the vignettes) was used as the reference category. Furthermore, pairwise comparisons of marginal effects with significance tests were also performed to test whether the marginal effects were significantly different from each other. This involved calculating contrasts. A contrast is the calculated difference between two marginal effects, and tests were made to check if each contrast was statistically different from zero.
Results
Figure 1 shows industry differences in employers’ overall hiring interest in disabled job seekers: the lowest degree of disability disadvantage was found within public administration, whereas accommodation and food service activities revealed the highest degree of disability disadvantage. Pairwise comparisons of marginal effects also revealed significant differences between industries in the assessment gaps presented. This also included industries that are similar when it comes to being characterised by physically demanding work and the need for communication towards clients/customers: health and residential care, childcare, and accommodation and food service activities all had assessment gaps that were statistically different from each other. The wholesale and retail trade also had a marginal effect that was statistically different from the marginal effects of accommodation and food service activities, and health and residential care. However, no significant difference was found between the marginal effects of manufacturing and construction.
Figure 1
The impact of disability on employers’ hiring assessments, by industry. Random-slope model, marginal effects.
Notes: Average marginal effects with 95% confidence intervals obtained from a two-level random-slope model measuring the impact of disability on employers’ hiring assessment scores. The model includes an interaction term between disability and industry and a random slope for the lower-level variable of this interaction. Level 1 variables: all vignette dimensions presented in Table 1. Level 2 variables: company size/sector and type of industry. Dependent variable: employers’ hiring assessment of candidates on a scale from 0 (‘hiring highly unlikely’) to 10 (‘hiring highly likely’). The marginal effects represent how many scale points lower on the hiring assessment scale disabled people were assessed, compared to non-disabled candidates, when no support measure was mentioned in the vignettes. The vertical line is the reference category (the predictive margin of non-disabled job seekers.)
There were significant differences between many of the marginal effects presented in Figure 1 and the results, therefore, confirm hypothesis H1 – that ‘the type of industry impacts employers’ hiring attitudes towards disabled candidates’.
Figures 2 and 3 present the results from model 2 and illustrate how the assessment gaps varied across industries for the separate groups of disability, using employers’ assessment of non-disabled applicants as the comparison group. The types of disability with more moderate differences between industries (the chronically ill, mentally ill, and intellectually disabled candidates) are presented in Figure 2, while the types of disability with larger differences between industries (the wheelchair user, hearing impaired, and blind candidates) are presented in Figure 3.
Figure 2
The impact of industry on employers’ hiring assessment of chronically ill, mentally ill, and intellectually disabled job seekers. Linear random-slope model, marginal effects.
Figure 3
The impact of industry on employers’ hiring assessment of wheelchair users, hearing impaired and blind job seekers. Random-slope model marginal effects.
Notes to Figures 2 and 3: Average marginal effects obtained from a two-level random-slope model measuring the impact of different types of disability on employers’ hiring assessment scores. The model includes an interaction term between disability and industry and a random slope for the lower-level variable of this interaction. Level 1 variables: all vignette dimensions presented in Table 1. Level 2 variables: company size/sector and type of industry. Dependent variable: employers’ hiring assessment of candidates on a scale from 0 (‘hiring highly unlikely’) to 10 (‘hiring highly likely’). The blue horizontal line is the reference category (the predictive margin of non-disabled job seekers.)
The results show how the blind job seekers and wheelchair users follow similar patterns as presented in Figure 1, where public administration was the most open industry, whilst accommodation and food service activities were revealed as the least open industries. A similar pattern is also found for the intellectually disabled candidate, although childcare turned out to be the least open industry for this type of disability. The variation in assessment gaps between industries were also much larger for wheelchair users and blind candidates, than for intellectually disabled job seekers.
Job seekers with a mental or chronic illness, or a hearing impairment, deviated from the pattern of industry differences described for disabled people as a group. For the mentally ill candidate, it was the professional, technical and scientific activities, and health and residential care that were the most open industries towards hiring, and the wholesale and retail trade that were the least open. Professional, technical, and scientific activities were also the most open industries for the chronically ill candidate. However, construction and activities of membership organisations were the least open. Manufacturing was the most open industry for the hearing-impaired job seeker, while the accommodation and food service activities were the least open. Overall, the results confirm hypothesis H2 – that ‘how and the extent to which the industry context impacts employers’ hiring attitudes depends on the type of disability’.
To test hypothesis H3, it was necessary to first measure the impact of support measures within each industry, for each type of disability separately, and then compare these. The results in Table 4 are presented as marginal effects, representing the change in predictive margin produced for each type of disability when a support measure was introduced to the vignettes.
Table 4
Industry differences in the impact of support measures on employers’ hiring assessments of disabled job seekers. Random-slope model, marginal effects.
| CHRONICALLY ILL | MENTALLY ILL | WHEELCHAIR USER | HEARING IMPAIRED | INTELLECTUALLY DISABLED | BLIND | |
|---|---|---|---|---|---|---|
| Public administration | 1.41* | 0.73 | 0.10 | 0.41 | 0.19 | 0.14 |
| Professional, technical, and scientific activities | 0.05 | 1.20 | 0.37 | 0.60 | 0.40 | 0.10 |
| Education | 0.98* | 1.06 | –0.04 | 0.35 | 0.14 | 0.68 |
| Activities of membership organisations | 1.54** | 0.51 | 0.12 | 0.34 | 0.41 | –0.04 |
| Health and residential care | 0.67 | 1.14** | –0.45 | 0.43 | 0.02 | 1.25*** |
| Manufacturing | 1.77** | 0.36 | 0.95* | 0.24 | 0.72 | 0.03 |
| Construction | 0.77 | 1.78** | 0.40 | –0.14 | 0.16 | 0.60* |
| Childcare | 1.45*** | 1.32** | 0.03 | 0.16 | 1.24*** | –0.12 |
| Wholesale and retail trade | 0.28 | 1.52** | –0.13 | 0.71 | 0.51 | 0.25 |
| Accommodation and food activities | 1.18 | 0.28 | 1.11* | 0.75 | –0.43 | –0.70 |
[i] Notes: ***p ≤ 0.01 **p ≤ 0.05 *p ≤ 0.10. Average marginal effects obtained from a two-level random-slope model measuring the impact of the type of disability (with and without a support measure mentioned) on employers’ hiring assessment. The model includes an interaction term between disability and industry and a random slope for the lower-level variable of this interaction. Level 1 variables: all vignette dimensions presented in Table 1. Level 2 variables: company size/sector and type of industry. Dependent variable: employers’ hiring assessment of candidates on a scale from 0 (‘hiring highly unlikely’) to 10 (‘hiring highly likely’). The marginal effects represent the change in employers’ hiring assessments of each presented group of disabled candidates when a support measure was introduced to the vignettes.
The results in Table 4 show that among all types of disability, except for the hearing-impaired job seekers and the wheelchair users, there was a significant impact (p < 0.5) of the introduced support measure on employers’ hiring assessments within at least one industry. The largest impact was that of wage subsidy schemes for the mentally or chronically ill candidates. Providing information about a 50% wage subsidy scheme led to significant increases ranging from 1.45 and up to 1.77 scale points in the hiring assessment score of the chronically ill candidates within childcare, activities of membership organisations, and manufacturing. No significant impact was found within the other industries. For the mentally ill candidate, the same wage subsidy scheme led to significant increases ranging from 1.14 to 1.78 scale points in the hiring assessment score within health and residential care, childcare, the wholesale and retail trade, and construction. At the same time, no significant impact was found within? the remaining six industries. For the wheelchair user, no significant impact of accommodation grants was found within any of the industries, and no significant impact of sign language interpreting services was found within any industry for the hearing-impaired candidate. For the intellectually disabled candidate, an 80% wage subsidy scheme led to a significant increase in employers’ hiring assessment score of 1.24 scale points within childcare, but no significant impact was found within the other industries. Accommodation grants had a significant impact of 1.25 scale points within health and residential care for the blind candidate. No significant impact was found though within the remaining industries.
Next, pairwise comparisons of the marginal effects presented within the same column of Table 4 were made to test whether the impact of support measures was statistically more concentrated within certain industries than others. This involved calculating contrasts, and testing whether these were statistically different from zero. The results show that for the chronically ill, intellectually disabled, and blind job seekers, there were significant differences (p < 0.05) between the marginal effects of at least two of the industries. In contrast, for the mentally ill, wheelchair user, and hearing-impaired candidates, the impact of the support measure was not found to be significantly more concentrated within certain industries than others. For the chronically ill candidate, the wage subsidy scheme had a significantly lower impact within professional, technical, and scientific activities than within manufacturing and childcare. For intellectually disabled candidate, the impact of wage subsidy schemes within childcare was significantly different from the impact found within food service activities. For blind job seekers, the larger impact of accommodation grants found within health and residential care was significantly different from the impact found within childcare and accommodation and food services.
Although the results found significant impacts of support measures on employers’ hiring assessments and that the degree of impact differed between industries, the overall picture shown by the results in Table 4 is still that the impacts of the support measures were small, limited or non-existent within most industries, especially for the wheelchair user, blind, hearing impaired or intellectually disabled candidates. Based on the results presented above, we can still confirm hypothesis 3, that ‘the impact of support measures on employers’ hiring attitudes towards disabled job seekers depends on the industry context’.
Discussion
This study contributes to the research field with evidence suggesting that the degree of disability disadvantage in hiring, measured when employers make hiring assessments, depends on the industry context. Findings also suggest that the industry context impacts the extent to which information about public support measures makes employers more open towards hiring disabled candidates.
Overall, this study finds that accommodation and food service activities, followed by the wholesale and retail trade, and childcare, were the industries least open towards hiring disabled candidates in general. Accommodation and food service activities were less open than other industries which are also characterised by physically demanding work and customer or client interaction – namely, the wholesale and retail trade, childcare, and health and residential care. The childcare industry was further less open towards hiring disabled job seekers than health and residential care. More research is needed to explore what explains these unexpected differences. However, the ideal worker and misfit concepts suggest that it is not only the job seekers’ lack of perceived ability to perform specific work tasks that contribute to employers’ assessment of disabled candidates as less ‘fit’ for the job but also more under-communicated organizational or structural factors, such as productivity demands, an organisation’s vulnerability to sickness absence, how tasks are organised between employees, etc. One plausible explanation could be that some of these ‘other’ factors are more dominant within certain industries, thereby affecting employers’ openness towards hiring disabled job seekers. Still, the results suggest that it would be easier for disabled people to seek employment within certain industries than others. At the same time, they also suggest that within some industries, there is a particularly strong need for employer-focussed policy interventions to influence employers’ hiring assessments and improve disabled people’s employment chances.
Results also reveal that the different types of disability show different patterns in how employers’ hiring assessments vary across industries. Some of these differences in pattern were expected, as the degree to which a specific type of impairment can limit a job seeker from independently performing core work tasks, without making adjustments to work design, can vary across industries. Other differences were unexpected and must be explained by other factors. One such example is mentally or chronically ill candidates, who were perceived differently within several industries despite being described in an identical manner in the vignette text as ‘having periods with reduced work capacity,’ with the only variations being between the terms ‘mental’ and ‘chronic’ illness, which were randomly selected. The most apparent example of this was within the wholesale and retail trade, where the mentally ill candidates were perceived as being far less employable than the chronically ill candidates. We know from previous research that stigma, discrimination, and employer reluctance have been identified as strong barriers towards employment for job seekers with a history of mental illness (Bjørnshagen 2021; Hipes et al. 2016; Stuart 2006). Our results suggest that, within the wholesale and retail trade, the characteristics associated with having a mental illness do disadvantage candidates to a greater extent than the characteristics associated with having a chronic illness.
The results also revealed that people with certain types of disability were assessed as being especially strong misfits – within all industries. This picture did not change notably when support measures were introduced alongside the impairment description. Whilst wheelchair users appeared to have opportunities within some of the industries, and chronically and mentally ill job seekers, with the help of wage subsidy schemes, appeared to have opportunities within most industries, the other types of disabilities included in the study – the hearing impaired and the intellectually disabled candidates, and the blind candidates in particular – were highly disadvantaged in most industries. Previous studies have found that disabled job seekers experience discrimination and low expectations about their abilities when applying for jobs (Baldwin and Marcus 2006; Chhabra 2021; Thornicroft et al. 2009; Vedeler 2014). However, the present study suggests that the hearing impaired, intellectually disabled, and blind candidates described in this study, encounter especially strong employer reluctance and discrimination when applying for jobs. This implies a need for special attention, both in research and policymaking, towards these specific groups to identify what explains their larger hiring disadvantage and, subsequently, which policies are necessary to implement to improve their employment opportunities.
The results also provide new insights into how the industry context influence the impact that wage subsidy schemes and accommodation grants can have on employers’ openness to hire disabled job seekers. Every industry is characterised by a unique set of characteristics – which can influence how incentivized employers are by these support measures. Childcare, as an example, stood out as the industry where wage subsidy schemes had the most pronounced impact on the employers’ hiring interest in disabled candidates. Kindergartens, which dominate the childcare industry in Norway, are publicly funded and strictly regulated when it comes to staff qualifications and number of children per employee (Drange and Rønning 2020). Sick leave absence is, at the same time, particularly high within this industry – which can help explain why this industry, when no wage subsidy scheme was mentioned, was one of the most hesitant towards hiring disabled job seekers. However, these same combinations of characteristics can also help explain why this industry was the overall most incentivized by wage subsidy schemes. When presented with a wage subsidy scheme, hiring disabled workers was perceived as more possible to do within the framework conditions in which the childcare employers operate.
The data further show that the impact of wage subsidy schemes and accommodations grants depended not only on the industry context, but also the type of disability. Chronically or mentally ill candidates, as an example, were described with a 40% wage subsidy scheme. Yet, this scheme had a more notable impact on employers’ hiring assessments than what an 80% wage subsidy scheme had on their assessments of the intellectually disabled candidates. For mentally or chronically ill candidates, the wage subsidy scheme is suggested to, at least to a notable extent, decrease the assessment gap between them and the non-disabled candidates, whilst this is not the case for intellectually disabled candidates. This lack of impact of wage subsidy schemes for intellectually disabled candidates illustrates how marginalized this particular group is from mainstream employment in Norway, where less than 6% are measured to be in paid employment (Wendelborg and Tøssebro 2018). It further suggests that it is not necessarily the high wage levels (which an 80% wage subsidy scheme would have compensated for) that makes Norwegian employers reluctant to hire these candidates, as suggested by previous scholars (Halvorsen, Hvinden and Schoyen 2015; Hauge 2021). Other factors, where stigma and low expectations about ability stand out as factors which should be explored, appear as equally large barriers to employment for intellectually disabled people.
While wage subsidy schemes and accommodation grants had a substantial impact on employers’ hiring assessments within certain industries for certain types of disabilities, the third support measure included in the research design, namely sign language interpretation services, stood out as the one support measure that did not have any significant impact on employers’ hiring assessments within? any of the industries. There are no plausible explanations for this lack of impact – according to misfit theory, a language interpreter should contribute to creating a material fit for these candidates within most industries. The results suggest, in the same way as the lack of impact of wage subsidy schemes for intellectually disabled candidates, that other factors beyond the costs of making accommodations explain employers’ lower interest in these candidates. The results emphasise the need for particular focus in future research to explore the reasons behind employers’ hesitance towards hiring the hearing-impaired candidates and which additional policies, in addition to language interpretation services, can be implemented to reduce this hesitance.
Norway has relied on a strategy based on ‘consensus, persuasion, and voluntariness’ when engaging employers in policies to increase labour market participation of marginalized groups (Midtsundstad 2008, 17). However, the discouraging finding of the present and prior studies is that publicly financed support measures which are designed to incentivize the employers are, with some exceptions, suggested to have a limited or no impact on their openness towards hiring disabled candidates. An overall plausible reason for these findings, in addition to stigma and employers’ low expectations about ability (Baldwin and Marcus 2006; Chhabra 2021; Østerud 2023; Thornicroft et al. 2009; Vedeler 2014), can be found in employers’ resistance to making adjustments to work design, as such a ‘variation to the standardized criteria would inevitably conflict with established organizational logic’ (Foster and Wass 2013, 709). Financial incentive schemes or services intended to compensate employers for accommodation costs might not be sufficient to challenge this resistance. The open question then becomes which policies could make a stronger impact on employers’ hiring interest and whether more enforced measures with mandatory participation would need to accompany the incentive-based policy measures presented in this study for them to make the desired impact on disabled people’s employment chances.
It is important to recognise that this study does not find that wage subsidy schemes, accommodation grants or sign language interpretation services cannot effectively assist disabled people into employment. The interpretation could rather be that these support measures are highly important, which employers have confirmed in previous studies (Mandal et al. 2019), but that the conditions for them to work as incentives for the employers are not in place. These lacking ‘conditions’ could include employers’ low expectations about disabled people’s abilities, organizational and structural barriers, and/or the lack of more enforced policy measures directed towards the employers. This author’s interpretation of the findings is that incentive-based support measures alone are not sufficient to improve employers’ hiring interest in disabled job seekers and that additional employer-focussed interventions likely is needed to achieve improved employement opportunities for this group of job seekers.
Although this study is located within a Nordic context, its main findings – that contextual factors impact employers’ assessments of disabled job seekers and how incentivized they are by support measures – can still be transferable to other socio-political contexts. However, characteristics such as unemployment rates, the ‘tightness’ of employment regulations, wage compression, disability discrimination legislation, cultural attitudes towards disability, access to assistive technologies and more – do vary across sociopolitical contexts and can produce different patterns in how disability disadvantage varies across industries. However, disability disadvantage and how employers respond to support measures must still be expected to be highly context-sensitive within any socio-political context. Future research on employers’ hiring attitudes towards disabled job seekers should be particularly aware of this when constructing their research design and when interpreting their findings.
Limitations
Although the factorial survey has many advantages, such as its reduced social desirability bias, and its multidimensionality, it also has the weakness of low external validity that must be addressed. When presented with job seeker profiles in a factorial survey, the employers’ assessments might not be representative of how they would respond to similar profiles in real-life hiring situations. The results must, therefore, be considered as showing employers’ preferences within a certain context rather than measures of their actual recruitment behaviour (McDonald 2019).
A second weakness is that, although the factorial survey is designed to decrease social desirability bias, we cannot rule out the possibility of employers being aware that they were being measured on their openness towards hiring disabled candidates. It can be easier to be positive towards hiring the presented candidates in a survey, where their responses are non-binding, compared to real-life hiring situations. Therefore, social desirability bias is likely to have impacted the assessments made, and employers may have assessed the disabled candidates more favourably than they may have in, for example, a field experiment. Therefore, this study’s results cannot be used to draw conclusions about the exact gap in employers’ hiring interest between disabled and non-disabled candidates – for that, other methods are more suitable, such as correspondence study design. However, the present study’s aim has instead been to measure industry differences in disabled peoples’ hiring disadvantage. For this objective, the study design has arguably proven to be suitable, as we have no reason to believe that social desirability bias is more present within specific industries than others. Instead, it is equally likely to have impacted the measured hiring disadvantage within all of the industries.
A third limitation is the decision to focus on industry differences when studying the impact of context on employers’ hiring attitudes. Although employers within the same industry share many similar characteristics, this specific way of grouping the employers may also obscure further differences between them within narrower categories – for example, between different professions. However, industry is used by this study as an example of how context can shape disability disadvantage – and it does not rule out the possibility that further differences may exist within narrower categories of employers.
An additional limitation is that this study has only focused on the impact of specific incentive-based support measures on employers’ hiring interest in disabled job seekers – but has not measured their impact on employers’ decisions to retain employees who acquire a disability after entering employment. This is another reason why the results should not, as also emphasised in the discussion section, be used to conclude about the importance of these support measures in preventing labour market exclusion of disabled people.
A final weakness is that no pre-study power calculations for the interaction between type of industry and disability were conducted before performing this study. Therefore, statistically insignificant marginal effects, or differences between these, cannot be used to conclude that there is, in fact, no impact of industry, or differences between industries, on employers’ assessments of disabled candidates.
Conclusions
This study contributes to an increased understanding of how context shapes disability disadvantage in hiring. It documents that employers’ openness to hire disabled job seekers, compared to non-disabled job seekers, vary across industry. The industry context also impacts the extent to which public support measures make employers more open towards hiring disabled people. The reasons behind these findings are not well understood, and future research would benefit from in-depth studies of specific industries to gain a clearer understanding. The results still reveal that job seekers with certain types of disabilities, especially blind candidates, but also those who are intellectually disabled or hearing-impaired, are highly disadvantaged within all industries. There is a need for special attention in future research towards these groups of job seekers to identify the reasons behind their larger hiring disadvantage and, subsequently, which policies can be implemented to improve their employment chances. Additionally, there is a need for future research to gain an increased understanding of why employers, in most settings, are not incentivized by information about publicly financed support measures in their hiring assessments of disabled job seekers. Importantly, the findings highlight a need for future research on disability disadvantage in hiring, and the role of support measures in reducing this disadvantage, to be particularly aware of how findings can be highly sensitive to the chosen research context.
Additional File
The additional file for this article can be found as follows:
Notes
Funding Information
This study is funded by the Research Council of Norway. Project ID: 273259.
Competing Interests
The author has no competing interests to declare.