Abstract
BACKGROUND:
Several studies have shown that the structure of employer attitudes toward employing individuals with psychiatric disabilities was multidimensional, but none has clarified the causal relationships within the employer attitude structure.
OBJECTIVE:
The purpose of this study was to formulate causal models based on data from two previously published studies of employer attitudes toward employing individuals with psychiatric disabilities, and to examine and verify the models. The models were either for all types of industry or specifically the transport sector.
METHODS:
The sample was selected to infer the influence factors for employment of individuals with psychiatric disabilities (for all types of industry) and specifically promotional factors (for the transport sector). An analysis of covariance structure was used.
RESULTS:
For all types of industry, prejudice and fear were shown to be indicators of activity limitation, attention distribution, and employer disincentive for hiring. In the transport sector, recruitment standards based on capacity were shown to be an indicator toward employer motivation for hiring and for making preparations to hire, self-efficacy for managing employment, and prejudice and fear.
CONCLUSIONS:
The results suggest that having recruitment standards based on capacity might reduce prejudice and fear and increase employer motivation for hiring individuals with psychiatric disabilities.
Keywords
Introduction
In studies mostly performed in the United States, employer attitudes toward employment of individuals with psychiatric disabilities tended to be more negative than attitudes toward individuals with other disabilities (Corrigan et al., 2008; Dalgin & Bellini, 2008; Diksa & Rogers, 1996; Farina & Felner, 1973; Glozier, 1998; Haj-Yahia, 1999; Manning & White, 1995; Melchiode & Jacobson, 1976; Ozawa & Yaeda, 2007; Tsang et al., 2007). Employer attitudes varied depending on the industry (Dightman & Marks, 1968; Diksa & Rogers, 1996; Ozawa & Yaeda, 2007), the age of the employer, and the experience of the employer in employing individuals with disabilities (Ozawa & Kikuchi, 2009; Ozawa & Yaeda, 2007). These studies have also illuminated additional research issues including ensuring the reliability and validity of studies, extracting samples objectively, determining factors that affect positive and negative attitudes, and improving response rates in mail questionnaire surveys.
There are two main dimensions along which attitude definitions seem to vary: (a) a dimension represented by a continuum of abstractness, ranging from concrete (or operationally defined) to abstract (or theoretically conceptualized); and (b) a dimension represented by the extensiveness of the components included in the definition (Antonak & Livneh, 1988). In this study, the structure of attitude refers to the extensiveness of the components (cognitive, affective, behavioral, or conative) included in Triandis’s (1971) definition of attitude, “An attitude is an idea [cognitive component] charged with emotion [affective component] which predisposes [conative component] a class of actions [behavioral component] to a particular class of social situations” (p. 2).
Several studies have shown that the structure of employer attitudes toward employing individuals with psychiatric disabilities was multidimensional, using exploratory factor analysis to investigate the structure that lies hidden behind the correlation of variables (Diksa & Rogers, 1996; Ozawa & Kikuchi, 2009; Ozawa & Yaeda, 2007). Employer attitudes toward employing individuals with psychiatric disabilities differ in their concerns (e.g., anxiety, worry; Crowther et al., 1995) by industry type and by prior experience employing individuals with disabilities, consisting of multiple elements and dimensions. The factors formulating employer motivation to employ individuals with psychiatric disabilities have been clarified in two studies. These two studies are the core research the current study is based on, so a detailed presentation follows.
Ozawa and Yaeda (2007) developed the Attitudes toward Employment of Psychiatric Disability scale (ATEP), a 27-item instrument composed of four subscales, each rated on a six-point Likert type scale ranging from “totally disagree” to “totally agree.” The coefficient alpha reliability estimates for the subscales ranged from +0.68 to +0.95. Estimates of test-retest reliability for the subscales ranged from +0.65 to +0.83. Content validity of the ATEP was established through evaluation by a panel of judges (counselors and researchers in the field of vocational rehabilitation) to determine the scaling values of the items. The construct validity of the scale was established by conducting factor analysis and confirming that the scale was composed of four factors. Ozawa and Yaeda (2007) used the ATEP to survey 358 employers in all types of industry in Japan and conducted multiple regression analysis to explain the factors formulating the dependent variable, “Employer Motivation” toward employing individuals with psychiatric disabilities, making the explanatory variables the respondent characteristics and the ATEP’s three subscales. The significantly related factors were prior experience employing individuals with disabilities and “Prejudice and Fear.” Transport sector employers ranked the highest in the industry-classified response rate (30.8%) among the response rate of all types of industry (22.4%), and they seemed to be highly interested in employment of individuals with psychiatric disabilities. The proportion of transport sector employers who achieved the legally required employment rate of workers with disabilities was the fourth highest (51.1%) among the 14 sectors examined (range from 20.4% [information-communication] to 56.2% [medical and welfare sector]) as of June 1st, 2007 (Ministry of Health, Labour and Welfare, 2007). The transport sector was considered an appropriate sector for further analysis in the identification of promotional factors of employment of individuals with psychiatric disabilities because employers in the transport sector were considered to be highly interested in employment of individuals with psychiatric disabilities and the achievement rate of the legally required employment rate of workers with disabilities was relatively high.
To determine attitudes in greater detail, Ozawa and Kikuchi (2009) therefore developed the revised ATEP II, a 52-item instrument composed of nine subscales, each rated on a six-point Likert type scale ranging from “totally disagree” to “totally agree”. Alpha reliability coefficients for the subscales ranged from +0.81 to +0.93, and test-retest stability coefficients ranged from +0.67 to +0.86. Content validation of the ATEP II was performed by a panel of judges, as was done for the ATEP. Construct validation of the scale was rendered by conducting factor analysis and confirming that the scale was composed of nine factors. They surveyed 478 transport sector employers in Japan. Again, respondent characteristics and the ATEP II’s eight subscales were considered explanatory variables. The results showed that “Recruitment Standards based on Capacity”, “Advantages of Hiring” individuals with psychiatric disabilities, “Prejudice and Fear”, and whether or not employers run businesses other than transport were significantly related to “Employer Motivation” for hiring individuals with psychiatric disabilities. Although the results of multiple regression analysis revealed linear relationships between “Employer Motivation” toward employing individuals with psychiatric disabilities, the dependent variable, and the explanatory variables in both studies, the causal relationships among the employer attitude structure were not examined, suggesting the need for alternate analyses. It is hypothesized that multiple causes intertwine to influence “Employer Motivation” to employ individuals with psychiatric disabilities in various forms.
The advantages of applying an analysis of covariance structure have been indicated in the field of psychology because the structure clarifies the causal relationship among factors and enables a comparison of the relative merits of multiple causal models (Kano & Miura, 2002; Toyoda, 1998; Toyoda, Maeda, & Yanai, 1992). In the current study, considering the causal relations among the employer attitude structure will be useful for helping potential employers, because it can elucidate how disincentive and promotional factors influence “Employer Motivation” to employ individuals with psychiatric disabilities.
Based on the issues identified above, the purpose of this study was to formulate causal models, focusing especially on “Employer Motivation” to employ individuals with psychiatric disabilities, based on data from two previously published studies of employer attitudes toward employing them (Ozawa & Kikuchi, 2009; Ozawa & Yaeda, 2007). The models were either for all types of industry or specifically the transport sector.
The primary research question for this study was: What are the causal relationships of the employer attitude structure toward employment of individuals with psychiatric disabilities in all types of industry and the transport sector?
Methods
Construction of hypothetical model for Ozawa and Yaeda (2007)’s data
In Ozawa and Yaeda (2007), the following four factors associated with employment of individuals with psychiatric disabilities were extracted: “Employer Motivation for Hiring”, “Activity Limitation”, “Prejudice and Fear”, and “Attention Distribution” (see Table 1 for factor definitions).
With these four factors as latent variables, an inter-latent variable causal model was formulated as follows (see Fig. 1). The hypothetical model presumed that “Activity Limitation”, “Prejudice and Fear”, and “Attention Distribution” have an impact on “Employer Motivation for Hiring”.
Construction of hypothetical model for Ozawa and Kikuchi (2009)’s data”
In Ozawa and Kikuchi (2009), the following nine factors associated with employment of individuals with psychiatric disabilities were extracted: “Employer Motivation for Hiring”, “Activity Limitation”, “Trustworthiness”, “Employer Motivation for Making Preparations for Hiring”, “Attention Distribution”, “Prejudice and Fear”, “Self-efficacy for Managing Employment”, “Advantages of Hiring”, and “Recruitment Standards based on Capacity” (see Table 1 for factor definitions).
With these nine factors as latent variables, an inter-latent variable causal model was formulated as follows (see Fig. 2).Of these nine latent variables, four were virtually the same in content as “Employer Motivation for Hiring”, “Activity Limitation”, “Prejudice and Fear”, and “Attention Distribution” from Ozawa and Yaeda (2007). The other five constructs were also taken as latent variables. These were: “Trustworthiness”, “Employer Motivation for Making Preparations for Hiring”, “Self-efficacy for Managing Employment”, “Advantages of Hiring”, and “Recruitment Standards based on Capacity”.
The hypothetical model presumed that: (a) eight variables (all except “Employer Motivation for Hiring”) probably influence “Employer Motivation for Hiring”; (b) “Recruitment Standards based on Capacity”, “Self-efficacy for Managing Employment”, “Trustworthiness”, and “Advantages of Hiring” probably influence “Employer Motivation for Making Preparations for Hiring”; (c) “Self-efficacy for Managing Employment” and “Prejudice and Fear” probably influence “Trustworthiness”; (d) “Prejudice and Fear”, “Self-efficacy for Managing Employment”, and “Trustworthiness” probably influence “Advantages of Hiring”; (e) “Prejudice and Fear”, “Trustworthiness”, and “Advantages of Hiring” probably influence “Activity Limitation”; (f) “Prejudice and Fear”, “Trustworthiness”, “Advantages of Hiring”, and “Activity Limitation” probably influence “Attention Distribution”; (g) “Recruitment Standards based on Capacity” probably influence “Self-efficacy for Managing Employment”; and (h) “Recruitment Standards based on Capacity” and “Self-efficacy for Managing Employment” probably influence “Prejudice and Fear”. The causal model of employer attitude formation shown in Fig. 2 was constructed by combining these hypotheses.
Analytical data
The analytical data were those reported in Ozawa and Yaeda (2007) and Ozawa and Kikuchi (2009). The first author of the present article conducted both of those previous studies on employer attitudes of all types of industry and of the transport sector specifically toward employing individuals with psychiatric disabilities, applying the use of exploratory factor analysis.
From the viewpoint of the extensiveness of the components (Antonak & Livneh, 1988) in the ATEP and the ATEP II, the following nine factors are considered: “Employer Motivation for Hiring”, “Employer Motivation for Making Preparations for Hiring”, and “Recruitment Standards based on Capacity” are the conative components; “Trustworthiness”, “Self-efficacy for Managing Employment”, and “Advantages of Hiring” are the positive cognitive components; and “Activity Limitation”, “Attention Distribution”, and “Prejudice and Fear” are negative cognitive components. Affective and behavioral components do not seem to be included in either scale.
The participants from whom the data were obtained were as follows: Ozawa and Yaeda (2007), 358 employers with more than 56 employees representing 14 Japanese industries classified by the Japan Standard Industrial Classification from all over Japan; and Ozawa and Kikuchi (2009), 478 employers with more than 56 employees representing seven Japanese transport industries classified by the Japan Standard Industrial Classification from throughout Japan. In Japan, all employers must employ a legally required number of individuals or more with physical or intellectual disabilities. This legally required number of employees with disabilities is calculated based on the sum of employees at all places of operation of a single business enterprise. Employers in Japan with more than 56 employees had been subject to an Employment Quota System for Individuals with Disabilities from July 1998 to March 2013. Because of a revision of the Employment Quota System for Individuals with Disabilities in April 2013, the employer’s size for obligation to employ individuals with disabilities was changed from 56 or more employees to 50 or more employees (Ministry of Health, Labour and Welfare, 2012). Individuals with psychiatric disabilities are not covered by this scheme, but are planned to be covered from April 2018 by the new Employment Quota System for Individuals with Disabilities (Ministry of Health, Labour and Welfare, 2013).
Analytical methods
Separate analyses were conducted for the data used in Ozawa and Yaeda (2007) and in Ozawa and Kikuchi (2009). For each data set, latent variables (i.e., all four factors of the ATEP; Ozawa & Yaeda, 2007 and all nine factors of the ATEP II; Ozawa & Kikuchi, 2009) were extracted based on the exploratory factor analysis, and a causal model was formulated. Analysis of covariance structure was applied to the causal models constructed; examination and verification were performed for each model. The effect indicators towards the observed variables were calculated for each of the latent variables, and the validity of the latent variables was examined. The values of the causal coefficients among the latent variables were used to examine the intensity of the causal relations. In cases where the causal coefficients among the latent variables were not statistically significant, the respective path was removed, and an attempt was made to correct to a more appropriate model.
The latent variables and observed variables (i.e., question items of the ATEP and ATEP II) used for analysis in this research are shown in Tables 2 and 3.
In evaluating the models, emphasis was placed on the use of chi-square value, Comparative Fit Index (CFI), Akaike’s Information Criterion (AIC), and Root Mean Square Error of Approximation (RMSEA). Other indicators such as the Goodness of Fit Index (GFI) and Adjusted Goodness of Fit Index (AGFI) could not be calculated because of missing values.
Results
Analysis of the Ozawa and Yaeda (2007) data
The maximum-likelihood method was used for parameter estimation for the model (see Fig. 1) that used the Ozawa and Yaeda (2007) data. The results of the analysis are shown in Fig. 3 and Table 4.
The causal coefficients shown in Fig. 3 were solved for standardization. All the coefficients were significant at the level p≤0.05. In Fig. 3, the arrows from d1, d2, and d3 indicate error variables.
As shown in Fig. 3, the effect indicators towards the observed variables for each of the four latent variables express high values on the whole, and all indicators are statistically significant at the 0.001 level. Thus, each observed variable will be valid as an indicator of the four latent variables. The Fit Index of this model was χ2 (318, N = 364) = 783.42, p < 0.001, CFI = 0.91, RMSEA = 0.06. Although the null hypothesis that this model fits data well was rejected, the Fit Index indicated that the fit of this model to data was not problematic on the whole.
This model was outlined as follows by paying attention to the statistically significant path coefficients shown in Table 4. First, significant path coefficients from “Prejudice and Fear” were found to the latent variables (a) “Activity Limitation” (path coefficient = 0.51, p < 0.001); (b) “Employer Motivation for Hiring” (path coefficient = –0.55, p < 0.001); and (c) “Attention Distribution” (path coefficient = 0.24, p < 0.01).
Second, significant path coefficients from “Activity Limitation” were found to (a) “Employer Motivation for Hiring” (path coefficient = –0.19, p < 0.05) and (b) “Attention Distribution” (path coefficient = 0.47, p < 0.001). Additionally, a significant path coefficient from “Attention Distribution” was seen to “Employer Motivation for Hiring” (path coefficient = 0.28, p < 0.01).
Analysis of the Ozawa and Kikuchi (2009) data
Figure 4 and Table 5 show the results of an analysis of covariance structure using the maximum-likelihood method for the data from Ozawa and Kikuchi (2009).
The following paths were removed in verifying the model, because those causal coefficients were not statistically significant: (a) from “Self-efficacy for Managing Employment”, “Trustworthiness”, “Activity Limitation”, and “Attention Distribution” to “Employer Motivation for Hiring”; (b) from “Trustworthiness” and “Advantages of Hiring” to “Activity Limitation”; (c) from “Prejudice and Fear” to “Trustworthiness”; and (d) from “Advantages of Hiring” to “Attention Distribution”.
The effect indicators from each latent variable towards the observed variables express high values on the whole. All indicators are statistically significant at the 0.001 level. Thus, each observed variable will be valid as an indicator of the nine latent variables. The Fit Index of this model was χ2 (1358, N = 478) = 4070.40, p < 0.001, CFI = 0.84, RMSEA = 0.07. Again, although the null hypothesis that this model fits data well was rejected, the Fit Index indicated that the fit of this model to data was not problematic on the whole.
The results of multiple regression analysis in all combinations of independent variables revealed that possibility of the existence of multicollinearity is not high, because there are not any multiple correlation coefficients R≥0.90. This model was outlined as follows by paying attention to the statistically significant path coefficients shown in Table 5.
First, significant path coefficients from “Recruitment Standards based on Capacity” were found to the variables: (a) “Self-efficacy for Managing Employment” (path coefficient = 0.51, p < 0.001); (b) “Employer Motivation for Making Preparations for Hiring” (path coefficient = 0.62, p < 0.001); (c) “Employer Motivation for Hiring” (path coefficient = 0.63, p < 0.001); and (d) “Prejudice and Fear” (path coefficient = –0.17, p < 0.01).
Significant path coefficients from “Self-efficacy for Managing Employment” were found to (a) “Trustworthiness” (path coefficient = 0.63, p < 0.001); (b) “Employer Motivation for Making Preparations for Hiring” (path coefficient = 0.22, p < 0.01); (c) “Advantages of Hiring” (path coefficient = 0.52, p < 0.001); and (d) “Prejudice and Fear” (path coefficient = –0.21, p < 0.01). Additionally, significant path coefficients from “Trustworthiness” were found to (a) “Employer Motivation for Making Preparations for Hiring” (path coefficient = –0.14, p < 0.01); (b) “Advantages of Hiring” (path coefficient = 0.29, p < 0.001); and (c) “Attention Distribution” (path coefficient = 0.17, p < 0.01).
From “Advantages of Hiring”, path coefficients were significant to the following latent variables: (a) “Employer Motivation for Making Preparations for Hiring” (path coefficient = 0.21, p < 0.01); and (b) “Employer Motivation for Hiring” (path coefficient = 0.09, p < 0.05). The coefficient from “Activity Limitation” to “Attention Distribution” was significant (path coefficient = 0.30, p < 0.001), as was the path coefficient from “Employer Motivation for Making Preparations for Hiring” to “Employer Motivation for Hiring” (path coefficient = 0.22, p < 0.001). Finally, significant path coefficients were seen from “Prejudice and Fear” to the latent variables: (a) “Advantages of Hiring” (path coefficient = –0.12, p < 0.01); (b) “Activity Limitation” (path coefficient = 0.53, p < 0.001); (c) “Employer Motivation for Hiring” (path coefficient = –0.08, p < 0.01); and “Attention Distribution” (path coefficient = 0.16, p < 0.01).
Discussion
Ozawa and Yaeda (2007) model
Results in the employers of all types of industry suggest that having prejudice and fear against individuals with psychiatric disabilities heighten the employer concerns about activity limitation and attention distribution of individuals with psychiatric disabilities, and lower the employer motivation for employing individuals with psychiatric disabilities. The results also suggest that having concerns about activity limitation lowers employer motivation for employing and heightens employer concerns about attention distribution of individuals with psychiatric disabilities.
Providing potential employers with accurate information on the nature of psychiatric disabilities may reduce employer concerns about coping with potential maladjustment behavior of and employing individuals with psychiatric disabilities. One apparent inconsistency is that employer concerns about attention distribution may heighten the employer motivation for employing individuals with psychiatric disabilities. A possible reason is that attending to details is a vital part for executing the work, and completing every task without any mistakes is a prime requirement for every industry sector.
Ozawa and Kikuchi (2009) model
Results of applying this analysis in transport sector employers suggest that having recruitment standards based on capacity raises the employer self-efficacy for managing employment of individuals with psychiatric disabilities along with the motivation for preparing to employ and for actually employing. In addition, such capacity-based standards lower the employer prejudice and fear against individuals with psychiatric disabilities. The results also suggest that self-efficacy for managing the employment of individuals with psychiatric disabilities raises the employer faith in competency and integrity of individuals with psychiatric disabilities. Such self-efficacy also strengthens the employer motivation for preparing to employ and the employer perception of the advantages of employing individuals with psychiatric disabilities. It lowers the employer prejudice and fear against individuals with psychiatric disabilities.
The results from this model also suggest that having prejudice and fear against individuals with psychiatric disabilities lowers the employer perception of advantages of and the employer motivation for employing individuals with psychiatric disabilities. Moreover, prejudice and fear heighten the employer concerns about activity limitation and attention distribution of individuals with psychiatric disabilities, as is indicated in Ozawa and Yaeda (2007) model. Conversely, having faith in the competency and integrity of individuals with psychiatric disabilities heightens employers’ perception of the advantages of employing them and strengthens employers’ attitudes about attention distribution.
Japan ratified the Convention on the Rights of Persons with Disabilities in January 2014 (Cabinet Office, Government of Japan, 2014). The convention prescribes the prohibition of discrimination for disabilities and the guarantee of equal rights for individuals with disabilities in all fields of society, such as education, labor, and social security (United Nations, 2006). In Japan, the government efforts toward realization of the rights of individuals with disabilities will be further strengthened. Individuals with psychiatric disabilities will be covered from April 2018 by the new Employment Quota System for Individuals with Disabilities. Together, these efforts should lead to an improvement in environmental factors for individuals with psychiatric disabilities, a decrease in participation restrictions that they face when applying for jobs, and the ability for employers to consider the competence and activities of individuals with psychiatric disabilities more fully.
Since transport sector employers appear to regard recruitment standards based on capacity as important, a key point for acceptance of individuals with psychiatric disabilities as on-the-job trainees or employees is how rehabilitation professionals present the capability and efficiency to potential transport sector employers. On the other hand, explaining weaknesses to and providing follow-up service to potential transport sector employers would also be a key factor for consideration of whether they can continue employing individuals with psychiatric disabilities. Telling potential transport sector employers how recruitment standards based on capacity differ between individuals with and without psychiatric disabilities would also be important, referring to the best practice of employing workers with psychiatric disabilities, and additionally to the characteristics of psychiatric disabilities.
Limitations and future tasks
In these analyses, employer attitudes toward employment of individuals with psychiatric disabilities were targeted. It is important to understand how these attitudes differ in regard to individuals with disabilities other than psychiatric disabilities to avoid overemphasizing the importance of these results for this one population of individuals with psychiatric disabilities.
Replications of this study, particularly in sectors other than the transport sector, are needed to establish generalizability. We need to better understand factors influencing employer motivation to employ individuals with psychiatric disabilities. Future studies will be necessary to formulate a more sophisticated model, because this study was based on published studies that had not originally aimed at formulating causal models. In these future studies, care should be taken to avoid missing values in the attitude scale during the scaling process, so that all the indicators can be calculated during the analysis of covariance structure. In addition, action research, especially participatory action research (PAR), will be helpful for researchers and employers to collaborate in defining the research problems and choosing research methodology.