Predictors of Employer-Sponsored Disability Accommodation Requesting in the Workplace

Katherine Breward, MBA, PhD
Assistant Professor, Department of Business and Economics
University of Winnipeg
k [dot] breward [at] uwinnipeg [dot] ca

Abstract

Employer-sponsored disability accommodation is contingent upon employees being willing to request such accommodation. This paper examines individual, organizational, and institutional predictors of accommodation requests among adult workers with disabilities using data collected from 5,418 respondents to a Statistics Canada post-census survey. The Theory of Planned Behavior provides a frame of reference to develop a series of hypotheses about how each type of predictor influences accommodation requests. One key finding is that different predictors are significant for each category of accommodation. Another important finding is that individual variables directly related to disability accounted for greater variance in requesting than other aspects of personal identity, organizational factors or institutional considerations. There was some evidence of decision-making based on attitudes, specifically fear of stigmatization. However, the data suggest that norms in the form of industry and occupation-specific logics are also salient influencers. Meanwhile institutional forces meant to act as behavioral controls, such as legislation and union protection, do not seem to have the intended influence on accommodation requesting. This suggests that other forms of intervention, such as community education, may be required to encourage the requesting of needed workplace accommodation.

Keywords

Acknowledgments

The author would like to acknowledge the insight, direction, and support offered by Richard Ivey School of Business faculty members during the development of this research. She would like to note the contributions of Prof. Chris Higgins, Prof. Debbie Compeau, and most especially Prof. Alison Konrad. In addition, this research would not have been possible without the staff of the Statistics Canada Research Data Center in London, Ontario. Statistician Bradley Corbett in particular contributed valuable insights and advice during the analysis process. Finally, Prof. Michael Breward from the University of Winnipeg offered commentary, personal support, and an endless amount of understanding throughout the project.

Introduction

Equal access to the labor market has become an important social justice goal all over the world. Many people continue to experience disadvantage in the labor market. For instance, people with disabilities are more likely to be unemployed (Burkhauser, Daly, & Hottenville, 2000; Schur, 2002; Schur, Kruse, & Blanck, 2005; Yeager, Kaye, Reed, & Doe, 2006) or underemployed (Jones, 2007; Kaye, 2009). They are also more likely to experience turnover (Schur, 2002), work part time (Bruyère, Erickson, & VanLooy, 2000; Schur, Kruse, Blasi, & Blanck, 2009), be stuck in entry level jobs (Kaye, 2009), and experience job insecurity (Kaye, 2009; Schur et al., 2009).

Employer-sponsored accommodation can assist workers with disabilities to overcome challenges that limit their full participation in the labor force. For example, employee accommodation is associated with job retention in both competitive and supported employment contexts (Campolieti, 2005; Campolieti, 2009). Employee accommodation is also associated with increased morale, productivity, and ability to complete job-related tasks (Schartz et al, 2006; Yeager et al, 2006).

In the past three decades, governments have devoted numerous political resources to maximizing labor force participation through the provision of employer-sponsored accommodation for workers with disabilities. The success of these rules and regulations depends, however, on several factors including people’s willingness to request needed accommodations, the probability that rules will be enforced, resource availability, and organizational culture. This study focuses on one of these factors: employees’ willingness to reveal their disabilities and actually request accommodation. Despite the importance of this precondition, the current understanding of accommodation requests is limited. Prior studies have focused largely on the need for accommodation or the provision of accommodation without explicitly assessing whether employees actually request accommodation (Campolieti, 2004; Campolieti, 2009; Rumrill, Schuyler, & Longden, 1997). Most studies that address accommodation requests rather than accommodation needs and employer behavior used small samples representing only one type of disability or one intervention (Allaire et al., 2003; Cook et al., 2007; Baldridge and Swift, 2011). One American study involving 780 workers with disabilities did look at requesting patterns based on the ethnicity, gender, education, and disability severity of the requestor (Von Schrader, Malzer, and Bruyere, 2014).This study, however, considers a broader set of variables, uses data from Canadian respondents, and involves a larger, much more representative sample.

Baldridge and Veiga (2001) proposed a framework of accommodation requests in which attributes associated with the requester (such as the severity of disability), with the accommodation (such as the cost), and with the work context (such as the organizational culture) collectively influence accommodation requesting patterns. However, only small components of the framework have been empirically validated (Baldridge and Veiga, 2006). Thus, while insightful, Baldridge and Veiga’s framework fails to identify the most salient predictors of accommodation requesting. Categories such as “attributes of the requestor,” “attributes of the accommodation” and “work context” are too broad and vague. They lack empirical insight into which specific requester attributes, accommodation attributes and work context attributes are most influential. Their framework also fails to address the underlying psychological processes that are responsible for accommodation requesting patterns.

The present research, therefore, takes inspiration from Baldridge and Veiga’s framework while seeking to identify which variables related to the individual and individual decision-making processes have the greatest predictive power in determining requesting rates. It also assesses a wider range of predictors than Baldridge and Veiga and relates them to a broad, inclusive psychological process theory: the Theory of Planned Behavior.

Theoretical Perspective

The Theory of Planned Behavior is an extension of Rational Choice Theory. Rational Choice Theory posits that an individual seeks to maximize benefits over detriments. More specifically, the Theory of Planned Behavior posits that an individual’s behavior is influenced by their behavioral intention, which is in turn influenced by their attitude toward the behavior, subjective norms, and behavioral controls (Ajzen 1991). The Theory of Planned Behaviour focuses on individual decision-making and as such it does not explore the full range of contextual variables that may also impact decisions. Despite this limitation, Ajzen (1991) has provided strong evidence that the theory has significant explanatory and predictive value. The three factors instrumental to this theory consistently explain a minimum of 0.43 and a maximum of 0.94 of the variance in intention to perform a given action across multiple contexts and disciplines.

According to this theory, therefore, requesting accommodation would only be a rational choice if the requestor anticipated receiving their accommodation and felt that the benefits of the accommodation would outweigh any negative consequences associated with making the request. Beliefs about anticipated gains in productivity and ease of task completion, for example, would need to be evaluated against anticipated negative consequences (or even the mere possibility of negative consequences) associated with making a request. This approach was supported by survey results from 780 American workers with disabilities, many of whom indicated reluctance to request accommodations due to fears of being treated differently by supervisors and peers (Von Schrader, Malzer, and Bruyere, 2014). This strongly suggests that accommodation requesting is a deliberate rather than subconscious process and that it requires active decision-making. This paper explores the variables that are relevant to that active decision-making, dividing them up into categories representing attitudes, subjective norms, and behavioural controls.

Hypothesis Development

Hypotheses 1-7, Attitudes Component of the Model

One of the negative consequences that inhibit accommodation requesting among workers with disabilities is the fear of experiencing stereotyping and stigmatization. Internalization of negative stereotypes about disability and associated fear of being stigmatized have been documented among people with physical disabilities associated with aging (Levy, Hummert, and Zebrowitz, 2003), psychiatric impairments (Lysader, Tunze, Yanos, et al., 2012; Watson and Corrigan, 2007), learning disabilities (McDonald, Keys, and Balcazar, 2007), obesity (Zettel-Watson and Britton, 2008), and bowel disease (Taft, Keefer, Arte, et al., 2011). There is evidence that these fears are justified. For example HR professionals given hypothetical scenarios in a pre-employment context gave lower suitability rating to candidates who were identified as disabled (Hazer and Bedell, 2006). Managers asked to grant accommodations made decisions driven by perceived obligation and the work attitude of the requestor; performance instrumentality and perceived fairness were less influential (Florey and Harrison, 2000). Negative social consequences, including coworkers resentment and exclusion, have also been associated with fear of accommodation requesting (Baldridge and Veiga, 2006; Von Schrader, Malzer, and Bruyere, 2014). Personal attributes that might induce fear of stigmatization include those directly related to being disabled and/or attributes related to other aspects of identity such as gender, ethnicity and age. Research by Baldridge and Swift (2011) found evidence of both disability stigma and personal attribute stigma among hearing impaired workers where age and gender also influenced accommodation requesting rates.

Stigma and associated discrimination can take place on a conscious or unconscious level and can be compounded by negative stereotypes. Awareness and consideration of stereotypes may therefore inform the “rational choice” of workers with disabilities when they are deciding whether to make disability accommodation requests. The fear of being stigmatized will impact persons with disabilities differently because, among other things, different disabilities are stereotyped differently. For instance, workers with invisible mental and psychiatric impairments experience more employment-related prejudice than those with physical disabilities (Cook et al., 2007; Dunn, Wewiorski, & Rogers, 2008; Sanderson & Andrews, 2006; Scheid, 2005). Moreover, workers with these forms of invisible impairments are often aware of these prejudices and internalize them (Lysader, Tunze, Yanos, et al., 2012; Watson and Corrigan, 2007). People with mental impairments therefore rationally feel inhibited from requesting accommodation. By contrast, physical impairment, such as loss of hearing, vision, and agility, are less stigmatized (Unger and Kregel, 2003), mitigating potential negative impacts on accommodation requesting patterns.

The perceived difficulty of providing an accommodation may also impact accommodation requests. Field evidence suggests that managers believe (often erroneously) that mobility impairments are especially difficult to accommodate (Lee, 1996; Lowman, West, & McMohan, 2005; McMahon, Shaw, West, & Waid-Ebbs, 2005). Moreover, many mobility-impaired workers are aware of these managerial biases (Balser, 2007), which should negatively impact employee willingness to request accommodations for mobility impairment, even though such impairments have comparatively low levels of stigma in the broader community. Thus, the greater the perceived difficulty of accommodation, the more likely an employee is to refrain from seeking accommodation.

Disability severity is measured by self-reported degree of functional limitation. It would be rational that people with very mild disabilities would be less likely to request accommodation because they have fewer functional limitations and are therefore more likely to be able to function adequately without accommodations. In addition, those with milder forms of disability are the most likely to be able to effectively hide their disabilities from their employers and colleagues. People who request accommodations run the risk of exposing themselves to stigma by drawing attention to disabilities that had previously been hidden or ignored. This perspective was supported in a study of 780 American workers. People with more severe disabilities were significantly more likely to make accommodation requests than people with milder impairments, and they did so earlier in their relationship with the employer. (Von Schrader, Malzer, and Bruyere, 2014).

Studies demonstrate greater labor market impacts on people who become disabled as adults (Brown & Emery, 2010; Wilkins, 2004). Adult onset is also associated with decreased self-esteem and increased anxiety (Kessler & Milligan, 2004), which may contribute to and enhance specific anxieties regarding labor force marginalization. Therefore, the older someone becomes disabled, the less likely s/he is to request accommodation.

The expectation of experiencing workplace stigmatization among workers with disabilities may be heavily influenced by other aspects of their identity, such as gender, being an immigrant or being a member of a visible minority. Such considerations may further inhibit accommodation requesting among individuals who anticipate experiencing multiple or compound forms of discrimination. That is, factors such as gender, immigration status and visible minority/ethnicity have strong associations with workplace discrimination independent of disability (Sargeant, 2006; Eagly and Carli, 2007; Ozbilgin, 2009), thereby decreasing the likelihood of an accommodation request. Such factors therefore will have an independent and negative effect on a request for disability accommodation in the workplace.

Attitudes are also heavily influenced by past experiences. Experience of past discrimination creates negative affect associated with subsequent avoidant behaviors (Richman and Leary, 2009; Hofstede, Hofstede, and Minkov, 2010; Von Schrader, Malzer, and Bruyere, 2014). Prior experience of negative discrimination may therefore result in a lessened probability of requesting needed accommodations.

Hypotheses 8-11: Subjective Norms Component of the Model

Having set out seven hypotheses associated with “attitudes towards behavior”, the first component of the Theory of Planned Behavior-based model, this paper now sets out four hypotheses associated with “subjective norms,” the second theoretical component. Subjective norms help establish how accommodations are culturally supported or rejected (Scott, 2001). Individual decision-makers will apply shared “logics of appropriateness” to lessen the possibility of being punished for accommodation related decisions (Meyer & Rowan, 1977). In other words, when employees consider potential reactions to disability accommodation requests, their expectations may be influenced by commonly applied logics of appropriateness with respect to who “deserves” organizational support.

There are well-established norms of increased supportiveness for loyal workers (Shore et al., 2008). For example workers who return to their previous employers after becoming disabled were less likely to experience wage reductions than those who change jobs (Campolieti, 2009; Gunderson and Hyatt, 1996; Thun, 2007). Given this norm, it seems likely that workers with longer tenure would be more likely to anticipate a positive reaction to a request for accommodation.

Receiving job training is also positively associated with perceived organizational support (Johlke, Stamper, & Shoemaker, 2002), in part because training represents a direct financial investment by the employer in the employee. Therefore, workers who received recent job training would be more likely to think their employer values and supports them and will be likelier to anticipate success when requesting accommodation.

There is a well-established and well-known norm in many workplaces of diminished supportiveness for temporary and part-time workers (Galarneau, 2005; Marshall, 2003; Gaskell and Rubenson, 2004). Both Balser (2007) and Zwerling et al. (2003) found that full-time employees with a disability were more likely to receive accommodations as compared to their part-time peers. Full-time and permanent employees are therefore likelier to make a formal request for disability accommodation.

Hypotheses 12-14: Behavioral Controls

Having considered attitudes and norms, this paper now turns to consider the third and final component of the Theory of Planned Behaviour: behavioral controls. Three hypotheses are set out which relate to behavioural controls, meaning externally prescribed factors that formally regulate granting patterns. The first of these hypotheses relates to a piece of employment equity legislation that is specific to Canada. As such the subset of findings related to this legislation may not be generalizable in countries without comparable legal protections. Most countries in the developed world, however, have similar legislation that would also act as a behavioural control. Comparative effectiveness would be expected to vary based on enforcement norms, economic vulnerability of worker populations, and national culture influences.

Employment equity legislation encourages the removal of discriminatory employment barriers for certain designated groups, including people with disabilities (Canadian Human Rights Commission, 2011). However, the Employment Equity Act applies only to those industries that are federally regulated under the Canadian constitution.1 Under this Act, employers have a statutory “duty to accommodate” workers with disabilities. It seems highly probable that employees who fall under employment equity legislation, (labeled “federal industries” for the remainder of this paper) will be likelier to request accommodations since they have the law on their side.

Whether an individual is a member of a union is a second potential behavioral control. Unions’ raison d’etre is to protect the interests of their members. Therefore, union contracts frequently include provisions related to the accommodation of workers with disabilities. Moreover, unionization is generally associated with increased access to formal benefits (Renaud, 1998), including disability accommodation. One Canadian field study found that unions had a “unique and positive” role in supporting accommodation, although it also acknowledged that issues of seniority and poor management/union relations could sometimes interfere with that process (William-Whitt, 2007). A union’s strong focus on formal rules, regulations and process may in fact inhibit the provision of disability accommodation since regulation can constrain behavior and encourage a focus on bureaucratic compliance rather than on genuine accommodation (Renaud, 1998). For example, one study of 186 healthcare workers in Ontario found that being unionized was positively associated with having employer-sponsored disability case management. However, unionization was also negatively associated with having a culture that truly meets the needs of disabled employees (Williams, Westmorland, Shannon, & Amick, 2007). Overall, however, it seems more probable that the average individual will anticipate support from their union and therefore be more confident in requesting workplace accommodation, even if the quality of the granted accommodation is not as effective in a unionized workplace.

In addition to legislation and union contracts, macroeconomic factors can also act as a behavioral control on employers. All organizations operate within a broader labor market and employers are generally willing to provide greater levels of support to employees with skills that are in high demand in an effort to minimize turnover and maintain access to their expertise (Kuttner, 1997). In addition employees with skills in areas of skills shortage have increased power in the employee-employer relationship and are thus better able to positively influence the decisions made by their supervisors and managers (Ringer and Boss, 2000).

In summary, fourteen hypotheses are organized into three categories – attitudes towards behavior, norms and behavioral controls –each representing one component of the Theory of Planned Behavior. In the next section of this paper, these hypotheses are tested.

Methodology

Sample

To test the hypotheses set out above, data were drawn from the 2006 Participation and Activity Limitation Survey (PALS), a post-census survey conducted by Statistics Canada. Survey respondents were selected on a stratified basis according to age, geographic location and activity limitations. People from all provinces and territories within Canada were included on a representative, stratified basis. The data were collected using a telephone administered survey from late 2006 to early 2007. A sample of 22,513 respondents was determined to meet Statistics Canada’s classification criteria of disability. (See http://www.statcan.gc.ca/pub/11-008-x/2009002/article/11024-eng.htm#a6 for extensive details on the definition of disability that was used.) For the purposes of this study, only data on adults over 18 years of age who held paid jobs at the time of the survey were included. Self-employed persons and individuals who had not reported needing at least one workplace accommodation were excluded from the analysis. The total sample size was 5,418 people and included the following characteristics:

Measures

Accommodation Type Categorization

The accommodations being studied had different characteristics that may influence requesting rates; as such they were divided into four subtypes based on these characteristics.

Dependent Variable of “Requested Accommodations”

Creating the dependent variable was challenging because of limitations associated with the wording of the PALS questions. Respondents were not directly asked about accommodation requesting. Therefore, “Requested Accommodations” was derived by combining data from several PALS questions, each of which is indicated below with PALS’s protocol of using an “HH” and a number/letter.

Respondents who reported needing accommodations but not receiving at least one of them had their responses to HH3a-b analyzed, which led to the following conclusions:

Respondents who had received all their accommodations were also categorized into people who requested and people who did not. The wording of questions HH2a-m implies that the accommodations are employer-sponsored. There is a possibility, however, that some respondents who received their accommodation actually did so without employer involvement. To address this, the type of accommodation was considered to determine if it was possible to self-provide. Some accommodations would be extremely difficult to self-provide such as job redesign, modified work hours, a modified work station, elevators and accessible washrooms. In such cases, therefore, it was assumed that the employer was involved in provision.

The remaining respondents reported receiving accommodation that could be employee- or employer-provided (such as technological aids, specialized computers, communication aids and specialized chairs). These respondents had their responses cross-referenced with JJ5, which asked if they have been denied a workplace accommodation in the last 5 years. If they answered “no” to JJ5, they were categorized as requesting their accommodation, since the cumulative evidence from multiple questions pointed strongly in that direction. Those who answered yes to JJ5 were removed from the sample since they could not readily be categorized.2

Please see Figure 1 for a visual representation of the variable creation process. The image shows the data categorization flow that was just described above. To review, step one was determining whether or not an accommodation was required. Step two examined whether it was provided or not. Step three determined whether the accommodation was actually requested or was likely to have been self-provided, ultimately resulting in a sample of workers who had requested accommodations. It is important to note that respondents were only included in the initial sample if they reported needing a given accommodation in the first place. As such the effects observed during data analysis are due to factors other than differences in initial need.

Figure 1: A Graphic Representation of the Variable Creation Process for “Requested Accommodation” Accommodation Required? (HH1) Accommodation Received? (HH2) Did not request. (HH3) Accommodation Requested Did request. (HH3) Employer Provided (acc type + JJ5) Non-Employer Provided (acc type + JJ5) NO YES NO YES

Figure 1: A Graphic Representation of the Variable Creation Process for “Requested Accommodation”

The “HH” number refers to the question number in the PALS questionnaire.

Independent Variables

Unless otherwise indicated, all binary variables were coded such that “1” indicated group membership (i.e. having a particular disability type, being a permanent employee, being an immigrant, working in an industry facing labor scarcity, etc.), while “0” indicated a lack of group membership. For demographic questions (gender, immigration status, visible minority status, age) responses were sourced from the regular census, which was cross-referenced with the PALS questionnaire using unique identifiers common to both surveys and embedded for that purpose.

Disability Type

Respondents were asked whether they had any of eight types of disabilities, specifically, hearing, seeing, mobility, agility, learning, memory, developmental, and emotional disabilities. The sample size (i.e. Ns) ranged from 236 to 3,737. For a full and highly detailed description of the definition of disability used, including dozens of qualifying questions tailored for each disability type, please see the Statistics Canada website at http://www.statcan.gc.ca/pub/11-008-x/2009002/article/11024-eng.htm#a6 .

Disability Severity

The severity score is derived by Statistics Canada and is reported on a 4 point scale with 4 indicating the highest degree of severity. Mean 2.08, SD .932.

Age at onset of disability

Respondents were asked their age when they acquired their disability. Mean was 34.02, SD 29.98.

Prior experience of discrimination

Respondents provided yes or no answers about whether they perceived experiencing each of eight forms of disability related discrimination in the past five years. The measure indicated the proportion of “yes” responses (Kuder-Richardson reliability measure = .73).

Permanent

Respondents were asked if their employment was permanent. 728 said “no”, while 4.690 said “yes”.

Full-time

Full-time was defined as working 30 hours per week or more. 1,135 were part-time and 4,283 were full-time.

Trained on the job

1,817 respondents had reported receiving training in the past year while 3,601 did not receive any training in the same time period.

Tenure

Tenure was rounded up to the nearest month based on the date they started their job.

Federal industry

Industries were documented by North American Industry Classification system codes (NAICs), which are standard classifications used by government agencies such as Statistics Canada and Industry Canada. 460 worked in industries beholden to federal equity legislation. 4,958 did not.

Union Member

1,699 respondents reported being members of a union. 3,719 were not.

Scarce Occupation

131 respondents reported working in an occupation that was experiencing scarcity, 5,287 were not. The occupations were categorized based on province-specific lists of scarce occupations. These lists (based on standard government National Occupation Classification codes, or NOCs), are used by Citizenship and Immigration Canada to inform the selection of immigrants for entry into Canada. They were published between the years of 2005-2007 under the Temporary Worker Program.

Control Variables

All control variables were taken from responses to the PALS survey:

Industry type

Consisted of eight groupings based on North American Industry Classification (NAIC) codes, which are standard classifications used by various Canadian government agencies. Groupings included agriculture, natural resources, manufacturing, trade, business services, professional services, tourism, and personal services. The n for industries ranged from 75 to 1,194.

Occupation type

Consisted of seven groups based on NOC codes, including management, professional, technical, clerical, sales/service, trades, and laborers. The n for occupations ranged from 266 to 1,440.

Analysis Method

Logistic regressions were conducted using SPSS v.18 in order to test the hypotheses. This technique produces odds ratios (ExpB) and a Nagelkerke R2, a pseudo-R2 measure that indicates the strength of the relationship between the predictors and the dependent variable (UCLA, 2012). Unlike OLS regression the R2 is not a measure of variance explained but rather an assessment of how well the model fits. The larger the Nagelkerke the better the model fit (UCLA, 2012). Regressions were run separately for each of the four categories of accommodation (job changes, technical aids, human support and structural changes). The control variables (industry and occupation) were always entered in the first step of the regression. The second step controlled for three of the four categories of predictors being considered [attitudes related to disability-related variables (H1-H5), attitudes related to other aspects of identity (H6a-H7), subjective norms (H8-H11), and behavioural controls (H12-H14)]. The third step included the predictor variables that were not controlled for in the second step. The model was therefore run four times in separate regressions for each type of accommodation, with the sole difference between them being the order of variable groups in the second and third steps. This enabled a comparison of the extent to which the goodness of fit was enhanced by each category of predictor. To guard against the possibility of inflating the degrees of freedom in the regression analysis, fractional weights (i.e., sampling weight for the individual รท average of all sampling weights) were used. Unweighted analyses showed that the weighting had no effect on the primary conclusions of the study.

Results

Overall Model Fit and Comparative Variance Explained by Category of Predictor

The model chi-square fit was significant (p<.001) for every model presented here, as were the Nagelkerke scores (p<.001) unless otherwise indicated. Overall there was strong evidence that the variables directly related to being disabled (disability type, severity, and age of onset) were much more substantive predictors of accommodation requesting than other variables. Some other aspects of identity (such as gender and age) were also statistically significant predictors of requesting. This supports the importance of attitude-related predictors in requesting. The organizational variables, which represent norms, showed varied relationships with requesting depending on the type of accommodation needed. Legal, contractual, and market-based behavioral controls were either insignificant or not substantively related to requesting. See Table I for details.

Table I: Nagelkerke Scores by Variable Type

Unless indicated otherwise all were significant at p < .001
Type of Request Overall model with controls Model fit contributed by disability variables Model fit contributed by identity variables Model fit contributed by org variables Model fit contributed by institutional variables
Job Change .18 .07 .02 .03 not sig
Tech Int .25 .08 .01 .009 .002
Human .33 .15 .02 not sig not sig
Structure .21 .10 .01 not sig not sig

See Table II for a detailed summary of results, including ExpB and p value statistics.3 H1 suggested that mental impairments would be negatively associated with accommodation requesting. This received mixed support in the statistical analysis since learning, emotionally, and developmentally impaired people were indeed less likely to ask for some types of accommodations. By contrast, emotional impairments were positively associated with requesting human support and structural changes. These results are highly counter-intuitive since psychiatric impairments are generally the most misunderstood and heavily stigmatized of all disabilities (Hinshaw, 2007; O’Mahoney and Donnelly, 2007).
H2, which suggested that physical impairments would be positively associated with accommodation requesting, received mixed support in the statistical analysis. Hearing and vision impairments showed positive associations with requesting some types of accommodations. Agility impairments, however, showed patterns that failed to support the hypothesis. It is possible that agility impairments are perceived more like mobility impairments, explaining the negative findings, but more research is required before this conclusion can be drawn.
Mobility impairments were negatively associated with requesting structural changes, providing partial support in the statistical analysis for H3.
Disability severity was positively associated with requesting job changes, providing partial support in the statistical analysis for H4.

Table II: Binary Regressions Predicting Accommodation Requesting

*** = regression step significant at p < .001, ** = p < .01, * = p < .05

Note that the significance level noted for Exp(B) is actually the significance level of the Wald statistic. Since the actual Wald statistic does not provide useful information to aid in interpretation it has not been included in these tables.
Job Change
N = 2,391		 Technical Int.
N = 2,183		 Human Sup
N = 210		 Structural
N = 468
Variable Exp(B) Exp(B) Exp(B) Exp(B)
Controls
Agriculture .77 1.12 2.38 1.06
Nat. Resource 1.00 1.70* .91 .90
Manufacturing .91 1.93** .81 1.30
Trade .89 2.36*** 1.15 1.31
Business Svcs 1.05 1.24 1.21 .86
Pro. Svcs .92 2.32*** .44 1.09
Tourism .96 2.78*** 1.54 1.83
Personal Svcs 1.09 2.34** .63 1.41
Management .85 .62* .79 .80
Professionals .67 .44*** .72 .56
Technicals .76 .68 1.39 .98
Clericals .74 .50** 1.62 .75
Sales/Svc .70 1.59* 1.13 1.16
Trades .79 1.84** 1.14 1.74
Laborers .83 1.56 1.66 1.32
Disability Related
Hearing 1.62*** 1.30 1.25 1.70*
Seeing 1.24 .96 .68 1.83**
Memory .79 1.07 1.46 1.65
Learning 1.19 1.14 .54* 1.04
Emotional .76* 1.28 2.24** 2.14**
Developmental .65* 1.09 .28*** .99
Agility .77** .50*** .71 .58**
Mobility .91 .85 1.57 .40***
Severity 2.86*** 1.42 2.55 1.67
Age at Onset 1.01*** .99 .96*** .98*
Intersectional
Female .89 .66*** 1.34 .83
Age .99* .99 .98 1.01
Visible Minority 1.18 1.29 .77 1.09
Immigrant .92 1.00 .77 1.09
Dis. Experience 1.31*** 1.24*** 1.39*** 1.23***
Organizational Norm
Permanent Job .74* .76* .85 1.06
Full-time 2.40*** .69** 1.85* .91
Trained in Job 1.19* .82 1.06 .84
Tenure 1.02*** 1.01** 1.00 1.00
Behavioral Control
Union Member .92 .79* 1.04 .94
Federal Industry .98 .91 .84 .81
Scarce Occ. 1.15 2.02 1.89 .58

Age at onset of disability was positively associated with requesting job changes but negatively associated with requesting human support and structural changes, providing mixed support for H5. It is worth noting that “age of onset” effects are marginal since the odds ratios are very close to 1. For example, results indicate that each one year increase in age results in a mere 1.01 increase in the likelihood of requesting needed job changes.
The next set of hypotheses (H6a through d) focused on other aspects of identity. These hypotheses received some limited support. Women with disabilities were 1.52 times less likely to request technical interventions than men with disabilities. Immigration status and visible minority status were not statistically significant predictors of accommodation requesting. Age was negatively associated with requesting job changes, although the odds ratios of .99 are so close to 1 that the effect is marginal. H7, which considered the impact of prior experience with disability-related discrimination, was thoroughly disconfirmed and strong evidence of the opposite relationship emerged. Each unit increase in prior experience of discrimination was associated with an increased likelihood of requesting accommodations of between 1.23 and 1.39 times depending on the accommodation type. Further research into responses to discrimination may help explain this finding.
The next series of hypotheses, H8-H11, considered the role of organizational norms surrounding who receives support. As expected, H8 confirmed that tenure was positively associated with requesting accommodations, in particular job changes and technical interventions. The other hypotheses received mixed support. H9, which considered worker training as a factor in disability accommodation requests, showed that workers who had received training were 1.19 times more likely to request job changes than untrained workers but they were 1.20 times less likely to request technical interventions. Full-time employees were more likely than part-time workers to request job changes and human support but they were also less likely to request technical interventions, providing mixed support for H10. Most surprisingly, H11 was disconfirmed since permanent employees were less likely to request job changes and technical interventions as compared to temporary employees. The final series of hypotheses considered the role of behavioral controls in accommodation requesting. Surprisingly, being in a federal industry or a scarce occupation were not significant predictors of accommodation requesting, disconfirming H12 and H14. Unionized employees were 1.25 times less likely to request needed technical interventions compared to non-unionized employees, demonstrating the opposite of what was expected for H13.

Requesting Accommodations Supplementary Analysis

Some unexpected findings emerged during the analyses that warranted further investigation. One of the most interesting unexpected findings is the extent of the variation in the models predicting requests for each type of accommodation (see Table II). It is clear that no single model can explain requesting for all types of disability accommodations. Technical interventions in particular are an outlier (see Table II). Since industry and occupational control variables had significant and substantive associations with requesting technical interventions, further analysis was warranted. Separate regressions (identical to the primary regression except for the sample selected) were run, predicting “requesting technical interventions” for individual industries with a sufficient sample size and for each occupational category. While there were no obvious overarching patterns to the differences in requesting technical interventions by occupation and industry, the extensive differences that did emerge nevertheless suggest that some currently unidentified institutional logics may be influencing requesting patterns. Further assessment could include researching each occupational and industrial subculture to determine respective attitudes towards disability and their associated logics of appropriateness. That information could explain the puzzlingly inconsistent industry and occupational patterns.

Discussion

One of the most notable findings of this study is the extent of the variation in the regression results by accommodation type. This difference reinforces the need to be careful when generalizing from studies using limited samples or studies that focus on one particular intervention. Because such studies currently constitute the majority of the published work in this area, this represents a serious limitation that should be addressed in future research programs. Interesting patterns emerge in the overall model of requesting accommodations (see Table I). Behavioral controls, set out in H12-14, were non-significant predictors for most types of accommodations and, when significant, they explained very little of the model fit and/or had negative associations with requesting (see Table II). One interpretation for this is that policy interventions at the legislative and group contract level have little impact on individual accommodation requesting patterns. This would imply that workers with disabilities are either unaware of their legislated and/or collective agreement-mandated rights or that they anticipate that these rights are difficult to enforce and/or are poorly protected. A couple of Canadian studies have supported this speculation, reporting that employees and employers alike find the formal regulations confusing and their associated rights and obligations unclear (Jain, 1992; Hucker, 1997). This is true both for federal regulations (the ones studied here), and for similar provincial legislation such as the AODA, also known as the Accessibility for Ontarians with Disabilities Act. If the interpretation presented here is correct then the extent to which this finding would be generalizable outside of Canada is questionable since it would depend on the availability of comparable legislation in other countries, the clarity of associated information, and their record of enforcement. The negative association between being a union member and requesting technical interventions also offers some support for the idea that workers lack confidence in their ability to enforce their rights. A perceived lack of individual voice as opposed to group voice may explain why behavioral controls such as union membership are not as influential as expected. This interpretation is consistent with emerging work that takes a more nuanced view of rights enforcement and voicing in union settings (see Benson and Brown, 2010; Callus, Morehead, Cully, and Buchanan, 2011). It is also worth noting that, to date, behavioral controls have not proven sufficient to address subtle social consequences associated with making special demands or having unique needs. Baldridge and Veiga (2006) found that the perceived imposition on others and attendant fear of social consequences heavily influenced willingness to make recurrent requests. Similar social effects may explain the lack of results associated with formal behavioral controls.
As for the variables related to organizational norms of supportiveness, set out in H8-11, some types of accommodations – full-time, trained, and long-tenured – were positively associated with accommodation requesting, as expected. There were also some instances of negative associations between accommodation requesting and being permanent, full-time, and trained on the job. One possible explanation is that permanent and full-time workers may simply have more to lose by revealing their limitations than temporary or part-time workers, inhibiting requesting. Anticipated social and self-esteem costs are relevant when asking people for help (Anderson and Williams, 1996; Galdas, Cheater, and Marshall, 2005). Temporary and part-time employees may have less to lose if they are stigmatized because they spend less time in the workplace, are less defined by their work role and may have more time for alternate esteem-generating activities. Temporary employees may also simply not care about stigma, since they will depart a given workplace in the short term, leaving behind their stigmatized reputation.
The analysis about other aspects of identity (H6a-H7) suggests that those variables are less substantive predictors of accommodation requesting. Canada has an especially equity-focused and inclusive workplace culture, so this finding may not generalize well in other countries where more overt forms of identity-based discrimination are tolerated. As such caution is in order and more research should be done outside of Canada. The discrimination index, by contrast, showed surprisingly robust results, albeit in contrast to the hypothesis. A post hoc search of the literature revealed that a small number of studies suggest that experiences of discrimination may motivate people with disabilities and foster activism rather than withdrawal (see Hofsetter and Schultze, 1989; Schur, 2003). Qualitative study in this area is warranted to determine the conditions under which experiencing discrimination will lead to avoidant versus activist behavior since prior research offers empirical support for both responses. For example people with disabilities could be presented with fictional scenarios involving discriminatory behaviours and their subjective responses (activism, avoidance, or other) could be documented. Scenarios could vary based on a range of variables including whether the discrimination seemed to be inadvertent or deliberate, whether it occurred publically or privately, the immediate consequences of the discrimination, resource availability, and perceived personal empowerment. The greatest contributor to model fit in accommodation requesting was clearly attitudes towards disability, set out in H1-5. At first blush, this pattern of findings seems to be attributed to accommodation needs being associated with certain types of disability. However, this would be erroneous since only those respondents who reported needing the accommodation were included during the variable creation process. There was support for the hypothesis that people with heavily stigmatized disabilities are less likely to request accommodations than people with other types of disabilities. Emotional impairments, however, had inconsistent relationships with requesting. Emotional impairments are generally the most misunderstood and heavily stigmatized of all disabilities (Hinshaw, 2007). The strong positive associations between emotional impairments and requesting certain types of accommodations are therefore difficult to explain. While this finding may indicate a lack of support for the stigma avoidance perspective, other explanations are possible. Some of the symptoms associated with emotional impairments may make awareness of other’s emotions, self-monitoring and impression management more difficult. People with emotional impairments may therefore be less likely to recognize the potential for stigmatization and/or be less likely to alter their behavior due to anticipated stigmatization (Brotman, Skup, Rich, Blair, and Pine, 2008; Downing, Johnston, Hansen, Schembri, and Stough, 2008; Lamanna, 2001; McClure, Pope, Hoberman, Pine, and Leibenluft, 2003; Pavuluri, O’Connor, Harral, Sweeney, 2007). Since two of the other forms of mental disabilities (learning and developmental) did show the anticipated relationship with accommodation requesting, the emotional impairment finding may be due to the symptoms associated with the disabilities themselves rather than a rational decision that the disability is unlikely to be stigmatized. This interpretation, however, does not fully eliminate the inconsistencies in the research findings. Using a stigmatization avoidance perspective, one would expect people with physical impairments to be more willing to request accommodations. In fact, this was true for hearing and vision impaired people. It was not true, however, for the agility impaired (who may be confused with mobility impaired individuals due to similar types of activity limitations). Further study is warranted. In particular it would be useful to conduct qualitative interviews with able-bodied people that attempt to surface, identify, and describe the specific stereotypes associated with various disabilities. This could then be followed up with interviews with workers with disabilities that assess personal awareness of disability-related stereotypes and associated impacts on behaviour. In the meantime, one firm conclusion is that disability type matters and is suggestive of stigma avoidance in some contexts. However, it is probably also driven by some other elements not currently understood, perhaps relating to culture-specific logics of appropriateness in particular disability communities.

Applied Implications

The most important thing for practitioners and policy makers to take from this research is that among the variables tested, disability type is the single largest determinant of accommodation requesting. (It is worth noting that contextual variables not addressed in this study may also make a significant contribution.) The substantive impact of disability type is a concern because it implies that fear of being stigmatized and the internalization of stereotypes about particular disabilities pose a barrier to accommodation requesting. The results of our analysis suggest that additional legislative or legalistic measures may not be effective. Education aimed at dispelling stereotypes and increasing awareness of the many abilities of people with disabilities may therefore be helpful since it could reduce the internalization of stereotyping. Education should also improve society-wide attitudes towards people with disabilities, easing fears about marginalization. At the organizational level, this could include formal diversity training, articles aimed at dispelling specific disability stereotypes in employer-sponsored health and wellness newsletters, and/or public speakers discussing stereotyping at company-wide meetings. At the government level, numerous public education options exist, from school programming to advertising. For example, in 2012 the Government of Canada sponsored television advertisements which attempted to dispel stereotypes by introducing real people with real impairments who perform well in jobs not traditionally associated with disability. Though it is impossible to help employees who do not identify their needs, managers should be mindful to avoid inadvertently creating barriers to requesting accommodations. Though it may seem ironic, given the research discussed above, permanent workers and union members need enhanced reassurance that accommodations are available. Workers transitioning from other industries or occupational categories may also have very different expectations regarding accommodation due to differing institutional logics. Managers should be particularly mindful of the need to establish appropriate expectations with employees undergoing a career transition in order to encourage accommodation requesting. Finally, all managers should note the results related to perceived discrimination. There is strong evidence that perceived workplace discrimination leads to increased activism. Employers wishing to avoid the negative outcomes associated with critical media attention should ensure that discrimination, deliberate or otherwise, neither impacts accommodation decisions nor is perceived as impacting accommodation decisions. Consistent, transparent accommodation decision processes can help protect against perceptions of unfairness, although the regression results for equity industries and union members suggest that formal policies are not sufficient in and of themselves to encourage requesting.

Limitations and Future Directions

One weakness of this study is its inability to assess the role of direct financial costs associated with providing an accommodation on requesting patterns. Since actual and perceived resource limitations may inhibit requesting, this is a limitation. The data in PALS simply does not enable valid cost-based analysis so field data will need to be collected to better understand the role of resource issues on requesting. That said, some categories of accommodation represent lower average costs than others. This may be one of the reasons that the comparatively low cost technical interventions show different patterns than other types of accommodations. Another weakness of this study is the small chance of classification errors as described in the variable creation section under the dependent variable description. See Footnote 2. Finally, as noted above, there was much more variation in accommodation requesting by accommodation type, industry and occupation than anticipated. While the rational choice perspective was supported in some ways, counter-intuitive findings also emerged. These inconsistent findings are difficult to interpret. Further research, especially qualitative research, is required in order to more fully interpret these findings and fully identify appropriate moderator and mediators. Open-ended interviews with workers with disabilities from a range of industries and occupations are likely to be the most effective way of moving this research forward, especially if analyzed in tandem with interviews conducted with managers that attempt to surface any implicit associations between disability and work performance that are industry and occupation specific.

Conclusions

Overall, the findings failed to clearly support or disconfirm a rational choice perspective influenced by stigmatization avoidance. Some results were consistent with the Theory of Planned Behavior. Many other results, however, were inconsistent. Most of the findings that failed to support the hypothesis were suggestive of the mediating and/or moderating role of individual affect, self-esteem maintenance and voice in the requesting process. For example, the negative relationship between union membership and requesting could be attributed to problems in exercising individual voice in a group voice context. The negative findings related to being permanent and full-time may relate to self-esteem maintenance since the consequences of stigmatization are greater when workers are more fully engaged in and defined by job roles. The discrimination index findings also have clear links to the role of political activism in voicing. Taken together, this suggests that predisposition to voice and social self-esteem needs should be added to models of accommodation requesting, perhaps as mediators between behavioral controls and organizational norm-based variables and requesting. The extent of the variation in requesting patterns for technical interventions is also highly informative. The differences between industries and occupations are difficult to explain with current models and are highly suggestive of sub-cultures, each with its own logics of appropriateness influencing accommodation requesting. These logics need to be researched and included in models of accommodation requesting, perhaps as moderators between disability type, severity and requesting. Taken together, it is not clear if these findings disprove the rational choice perspective or simply clarify the criteria used to determine what is considered rational. The supplementary assessment of technical interventions, which suggested that industry and occupation-specific logics were more relevant than generalized norms related to organizational supportiveness, is an excellent example of “clarifying criteria used to determine rationality.” This still represents a rational decision, just not the type of rational calculation that was originally anticipated since the workers making requests appear to be focused on different criteria than expected (i.e. specific industry norms rather than broader organizational supportiveness norms). This sheer number of surprising findings suggests the theoretical limitations of this research. Interpretation of some of the counter-intuitive results, although well supported by prior research and theory, is difficult because multiple interpretations are possible and multiple behaviors could be contextualised as rational in some way. The post hoc explanations for surprising findings that appear in the Discussion section should therefore be considered speculative until validated by qualitative interviews.

Footnotes

  1. The Canadian Human Rights Act also mandates accommodation, though in less specific terms. It applies to all industries.↑
  2. There is a slight possibility of classification error in this last sub-group. The risk of such errors impacting the analysis is mitigated due to the large sample size, the fact that this sub-group represents less than 6% of the total sample and that membership in this sub-group is not associated with significant differences in correlations with other predictors. Out of an abundance of caution, regressions were run once, eliminating this 6% from the sample. No substantive differences in results were observed so they were retained within the sample.↑
  3. The term “partial support” is being used when some accommodation types (i.e. job changes, human support) show the anticipated relationship with a predictor but others are not significantly related to the predictor. The term “mixed support” is being used when some accommodation types show the anticipated relationship with a predictor but others show the opposite relationship.↑

References