The Influence of Perceived Discrimination on the Labor Market Status
of Disabled Individuals
23th of June 2017
Manon van der Stam - 10500146
MSc Business Administration - Leadership and Management Track Amsterdam Business School - University van Amsterdam
Statement of Originality
This document is written by Student Manon van der Stam who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Abstract
This study aimed at examining the influence of discrimination as perceived by disabled individuals on a daily basis on their labor market status. Additionally, the influence of self-esteem and well-being on this relationship has been studied. A negative direct relationship between perceived everyday discrimination and labor market status was expected, as well as a partial mediation effect of psychological well-being and self-esteem on this relationship. In addition, self-esteem was expected to moderate the relationship between perceived everyday discrimination and self-esteem. Data was collected from 154 Dutch individuals, with any kind of disability, illness or condition limiting them from executing daily activities, such as working, optimally. Results did not show a significant direct effect of perceived discrimination of the labor market status of disabled individuals. Additionally, the indirect effect of psychological well-being and conditional effect of self-esteem that were expected based on previous literature were not supported by the data analysis. However, the hypothesis that self-esteem mediates the relationship between perceived discrimination and labor market status is supported by the results.
Key words: perceived discrimination, disability, self-esteem, psychological well-being, employment
Table of Contents
1. Introduction ... 5
2. Theoretical Framework ... 9
2.1 Disabled Individuals and Labor Force Participation ... 9
2.2 Perceived Discrimination and Employment of Disabled Individuals ... 11
2.3 Perceived Discrimination and Well-Being ... 13
2.4 The Role of Self-Esteem ... 17
3. Method ... 21 3.1 Sample ... 21 3.2 Procedure ... 23 3.3 Measures ... 23 3.4 Data Preparation ... 25 4. Results ... 27 4.1 Correlations ... 27 4.2 Regression Analysis ... 29 4.2.1 Direct effect. ... 29
4.2.2 The mediating effect of psychological well-being. ... 29
4.2.3 The mediating effect of self-esteem. ... 33
4.2.4 Conditional effect. ... 35
5. Discussion ... 36
5.1 Theoretical Implications and Suggestions for Future Research ... 38
5.2 Practical Implications ... 40
5.3 Limitations ... 40
6. Conclusion ... 42
6. References ... 43
1. Introduction
In 2015, 1.7 million people in the Netherlands had to deal with a disabling condition restricting them from gaining or successfully executing a job. This is around 10% of the Dutch population. To illustrate the difference between disabled individuals and non-disabled individuals in terms of participation in the labor market, the age group of people between 25 and 45 years old will be used as an example. In general, this age group is the most active group in terms of labor market participation, as 92% of them participate in the labor market. However, of disabled people within this category of age, only 42% is participating in the labor market, which means that more than half of them have recently not been seeking employment, or have not been available for it (CBS, 2016). Additionally, data shows that from those people with a disabling condition who are actively seeking employment, 14% was unemployed, in contrast to 5% of people without a disability (CBS, 2016). This data provided by the Central Bureau for Statistics in the Netherlands indicates that people with disabilities are less likely to participate in the labor market than people without disabilities. Furthermore, it shows that of those people with disabilities who are participating in the labor market, a large portion is unemployed in comparison to people without disabilities. Although this unavailability could be due to the fact that the disability or illness of these individuals does not allow them to have a paid job, the possibility that there are other factors influencing the participation of disabled or ill individuals in the labor market needs to be taken into consideration. Therefore, the question is, what other factors could influence the labor market participation of disabled individuals?
In previous literature, an extensive amount of research has considered employment of marginalized groups, mainly focusing on demographic factors such as gender diversity or ethnic minorities (Snyder, Carmichael, Blackwell, Cleveland, & Thornton III, 2009). The research on the employment of disabled individuals usually focuses on employers’ motives
not to hire disabled individuals, for example, because they are afraid that disabled employees require accommodations that are costly, or because of discrimination and prejudice about their productivity and other factors (Snyder et al., 2009; Stone & Colella, 1996). Additionally, the integration of disabled employees in organizations has been studied, particularly focusing on how disabled employees experience their job and how their colleagues and managers treat them (Ali, Schur, & Blanck, 2010; Nota, Santilli, Ginevra & Soresi, 2013; Schur, Kruse, Blasi, & Blanck, 2009; Stone & Colella, 1996). Furthermore, the perceptions disabled people have about discrimination and injustice in the workplace have been studied (Snyder et al., 2009).
Despite the fact that the topics that have been studied are all relevant research topics, the fact that a high percentage of disabled individuals is still not participating in the labor market or is unemployed should not be ignored (CBS, 2016). A limitation of the extant literature is that most of it is assuming that people with disabilities are participating in the labor market, since it focuses on the application process to obtain a job, or on the integration within organizations. Although part of the employment gap between people with and without disabilities might be due to lower productivity associated with many disabilities, prejudice and discrimination are also likely to play a role (Schur, 2002). As research by Ali et al. (2010) has revealed, the unemployment of people with disabilities is not due to their unwillingness to work, since they are as likely as people without disabilities to want a job. Research has also shown that the majority of non-employed people with disabilities would prefer to be working (Schur et al., 2009). However, they are less likely to actively seek employment, because they are pessimistic about finding a job that suits them (Ali et al., 2010). In addition, since literature about the employment of disabled individuals and other marginalized groups has found that discrimination often forms a barrier to employment for these groups, we should investigate how the unemployment of disabled individuals is related to the extent to which
people with disabilities perceive to be discriminated in their daily lives as a result of their disabling conditions. In comparison to previous literature on this topic, studying this relationship will provide novel insight as to how the participation of people with disabilities in the labor market is influenced, before they are actually participating in the labor market by applying for and executing jobs. Therefore, this study aims at providing more insight into how the degree of discrimination as perceived by disabled individuals influences their labor market status. In doing so, the role of self-esteem and psychological well-being on this relationship will be studied.
Research into the factors influencing employment of disabled individuals is relevant for both practical and theoretical reasons because it will provide an insight into possible ways to overcome these problems. In a practical sense, the results of this study will be useful for individuals with disabilities, for organizations, but also for society as a whole. Firstly, an increase in employment of people with disabilities will improve their overall life satisfaction, and decrease the level of social exclusion, which is a very common issue for disabled individuals without employment (Schur, 2002). In addition, employment can increase the level of social interaction experienced by disabled individuals, which gives them the chance to improve their social skills, so that they can integrate into society more successfully (Schur, 2002). Secondly, organizations can profit from the employment of disabled individuals through the benefits of having a diverse workforce, resulting from the different points of view people with differing backgrounds may have (Shen, Chanda, D'Netto, & Monga, 2009). Additionally, organizations can benefit from hiring disabled individuals in terms of performance, as research has found that disabled employees performed as well as, if not better, than their non-disabled counterparts (Stone & Colella, 1996). Thirdly, the Dutch population is aging, as many workers from the baby-boom generation are soon retiring. This will result in labor shortages that will rise up to a level where organizations will not be able to
hire enough employees, which will happen at the expense of economic growth (CBS, 2015). In case of such shortages, an increase in the employment rate of disabled individuals could be a solution to fill labor shortages, apart from the other positive effects resulting from it (CBS, 2015; Schur, Nishii, Adya, Kruse, Bruyère, & Blanck, 2014).
This study will continue with a detailed review of the existing literature, which will be used to explain what this study refers to when discussing disabled individuals, discrimination, and other core concepts. Also, the extant literature will provide the basis for the expected relationships in the present study. Next, the method for executing the empirical part of the study will be discussed. Finally, the results will be presented and discussed, together with their implications.
2. Theoretical Framework
2.1 Disabled Individuals and Labor Force Participation
Before explaining the variables in the model and presenting the hypotheses tested in the study, it is important to discuss what group of individuals is referred to as disabled individuals in the study. Since the present study is about employment, it is essential to understand that many people with some kind of disability, long-term illness, or other condition are able to perform a paid job, and that this is the kind of disability referred to in the study. Therefore, the definition of disabled individuals as used in this study will be further explained using existing literature.
Stone and Colella define disability as ‘‘a physical or mental impairment that substantially limits one or more major life activities’’ (1996; p. 4). In their research they assume that disabled individuals, in the situation of employment, are qualified and able to perform the job in question well (Stone & Colella, 1996). Ali et al. (2010) identified four major categories of impairment, which are visual, hearing, mobility and mental impairment. Additionally, Burkhauser, Daly, Houtenville, & Nargis (2001) and Schur (2002) use the term ‘work disability’, which does not refer to a severe disability, but rather to a health condition that limits the kind or amount of work individuals can do. A combination of these definitions captures what is meant by disability in the present study, in which by disabled individuals we refer to individuals who have some kind of disability, long-term illness or other condition that limits them to execute particular activities in life as easily as they would be able to without their condition, but who are mentally and physically able to perform a paid job. This implies that many people with a work disability are still able, or at least to some extent, to contribute to the labor market. Research has already shown that there are many jobs that people with disabilities can do (Ali et al., 2011). Although there might be some exceptions of people who are currently or permanently unable to work, another portion of the many disabled individuals
who are unemployed is able and willing to have a paid job, but has a lot of difficulty in finding one (Ali et al., 2010; Schur et al., 2009). As Ali et al. (2010) found, this can be due to the rather pessimistic attitude of many disabled individuals towards finding a job that suits them, for example because of their often lower educational level, or because they feel like they will not be accepted as a result of discrimination.
This is unfortunate, because employment can be of great importance to disabled individuals, for several reasons. According to the literature about disabled individuals, they usually have low incomes, are likely to live alone, less likely to be married than people without disabilities, have low educational levels and are likely to be socially isolated (Ali et al., 2010; Schur, 2002, Schur et al., 2009). Taking this into consideration, it is not surprising that individuals with disabilities have overall lower levels of life satisfaction than those who do not have a disability (Schur, 2002). Therefore, employment could have positive implications for disabled individuals. Not only does it increase their overall life satisfaction, it may also reduce the economic and social gaps that exist between disabled and non-disabled individuals (Schur, 2002; Schur et al., 2009). Employment will enhance the economic security of disabled individuals (Schur, 2002). Furthermore, by providing disabled individuals with the ability to gain more social contacts through employment, they can develop social skills and decrease the degree of social isolation (Putnam, 2000). This also has psychological benefits for disabled individuals (Schur, 2002).
So, employment is beneficial for people with disabilities in multiple ways, and therefore it is relevant to find out why a large portion of disabled individuals is still unemployed or not even seeking employment. Different studies have already discussed factors influencing barriers to employment faced by disabled individuals. One barrier that has a particularly strong influence on the employment of disabled individuals is discrimination (Stone & Collela, 1996).
2.2 Perceived Discrimination and Employment of Disabled Individuals
Before discussing perceived discrimination and how it affects disabled individuals, it is necessary to define what is meant by discrimination in this study. Many different concepts relating to discrimination are used in the literature, such as stigma, prejudice, and discrimination. Thornicroft, Rose, Kassam and Sartorius (2007) studied these three concepts and discussed if and how they are related. Based on their findings, we can say that, although the three concepts are sometimes used either separately or together, they actually describe a process.
First of all, stigma refers to ignorance, which is a problem of knowledge. Then, prejudice refers to the negative attitudes people have based on a lack of knowledge about a minority group. Finally, discrimination refers to the behavior that results from these negative attitudes (Thornicroft et al., 2007). In other words, this means that discrimination can be referred to as the behavior of people towards minority groups that is based on the negative attitudes they have, due to a lack of knowledge about this group. For the present study this means that people may build attitudes towards disabled individuals due to a lack of knowledge about the disability or condition and its consequences. However, this study will not measure actual discrimination, but perceived discrimination, which refers to a behavioral manifestation of a negative attitude, judgment, or unfair treatment toward members of a group as experienced by members of that group, which, in this case, is the group of disabled individuals (Banks, Kohn-Wood & Spencer, 2006). Important to note is that this is not an objective measure of discrimination, since it is about the feelings and perceptions individuals have about being the victim of discrimination, rather than a measure of the actual degree of discrimination (Paradies, 2006; Verkuyten, 1998). Araújo and Borrell (2006) explain that there are two different types of perceived discrimination. The first type consists of chronic discriminatory stressors, which are daily hassles of discrimination and exclusions. Acute
discriminatory stressors, on the other hand, are events that occur sporadically throughout one’s life. In the current study, by perceived discrimination we refer to everyday experiences of discrimination, so the chronic type of stressors. This type of discrimination was chosen because we are interested in the effect of everyday discrimination experienced by disabled individuals on their participation in the labor market rather than sporadic events of discrimination, in order to provide insight in the impact prejudices and discriminatory of society as a whole can have on individuals with disabilities.
Earlier research about the treatment of disabled individuals in organizations and employment has discussed the influence stigma, prejudice, and discrimination have on the employment of disabled people (Ali et al., 2010; Schur et al., 2009; Stone & Colella, 1996) For example, Stone and Colella (1996) developed a model of factors that affect the treatment of disabled individuals in organizations. In their study they found that employers’ stereotypes and attitudinal biases are an important source of the employment problem of disabled individuals. This effect of employers’ discriminating attitudes towards disabled individuals has been recognized in many studies and might influence not only those individuals who are employed, but might even be a barrier to seek employment (Baldwin & Johnson, 2006; Blanck, 2001; Schur et al., 2009; Ali et al., 2010). The fact that disabled individuals are less likely to actively seek employment than nondisabled individuals, is due to the fact that disabled individuals are less optimistic about their chances to find a suitable job (Ali et al., 2010). This can have different reasons, for example, these individuals might feel that their disabilities lower their productivity level, so that they are not able to perform certain jobs. However, their more pessimistic attitude towards looking for employment is also likely to be due to the fact that disabled individuals perceive that prejudice and discrimination lowers their chances at finding a job (Ali et al., 2010).
So, the literature suggests that discrimination is related to employment of disabled individuals in different ways, which implies that discrimination has a large influence on disabled individuals (Stone & Colella, 1996). Additionally, disabled individuals are less likely to seek employment, although they want to have a job, which is likely to be influenced by the discriminatory behavior of others that they experience in their daily lives (Ali et al, 2010; Baldwin & Johnson, 2006; Blanck, 2001; Schur et al., 2009). Combining these different findings from previous literature, the first hypothesis is:
Hypothesis 1: everyday discrimination as perceived by disabled individuals is negatively related to their participation in the labor market.
2.3 Perceived Discrimination and Well-Being
The first hypothesis suggests that perceived discrimination has a direct effect on the participation of disabled individuals in the labor market. However, this relationship might also be partially influenced by other variables, such as psychological well-being. Perceived discrimination implies rejection and exclusion of targeted groups and their members, and thus can undermine psychological well-being by threatening the fulfillment of needs for inclusion and acceptance (Baumeister & Leary, 1995; Wirth & Williams, 2009).
Well-being is a very broad construct that can be defined in many different ways and can be deconstructed into many components, which makes it quite difficult to provide a general definition (Biddle & Ekkekakis, 2005). According to Summer (1996), the original meaning of well-being is literally the condition of being well. In addition, the term well-being is often used interchangeably with happiness (Diener & Lucas, 2000), and is regularly indicated by life satisfaction (Ryff & Keyes, 1995). However, the concept of well-being is of subjective nature, which means that individuals determine for themselves whether they are
being well or not, so what ‘being well’ truly means, can differ among individuals (Diener & Lucas, 2000; Summer, 1996). Although there is no clear, general definition of well-being, it is important to specify which aspects of well-being are and which are not included in the current study. Van der Zee and Sanderman (2012) describe eight dimensions of well-being, namely physical functioning, social functioning, role limitations due to a physical problem, role limitations due to an emotional problem, mental health, vitality, pain, and General Health Perceptions, and health change. In the present study, some of these dimensions are more relevant and useful than others. For example, Schulz et al. (2006) found that everyday experiences of discrimination causally affect poor mental health and physical health outcomes. However, this relationship is clearer for mental health than for physical health, for which the relationship with discrimination is more complex and inconsistent.
The present study will focus on the psychological, emotional and mental aspects of well-being, because these aspects are most likely to be influenced directly by the perception of discrimination (Schulz et al., 2006). As Van der Zee and Sanderman (2012) also indicate, different dimensions of well-being are related so that people with physical health issues may be affected by this mentally or emotionally, and the other way around. However, since the direct effect of discrimination on psychological well-being is more clear, and the physical consequences often appear later (Schulz et al., 2006), it is more likely that discrimination perceived by disabled individuals will directly affect psychological well-being than physical well-being. In previous literature this is usually referred to as psychological well-being or mental health, therefore these terms may be used throughout the current study. Psychological well-being is characterized by the mental and emotional conditions of individuals, so that psychologically well people are more prone to experience positive emotions and less prone to experience negative emotions (Wright & Cropanzano, 2000). Important to note here, is that the relationship between perceived discrimination and mental health outcome has been tested
in cross-sectional studies, in order to prove that the relationship is not due to the tendency of people with mental health issues to perceive themselves as being treated poorly or discriminated against (Brown, Williams, Jackson, Neighbors, Torres, Sellers, & Brown, 2000).
Many studies have examined the relationship between perceived discrimination and psychological well-being or mental health outcomes. Overall, the direct negative relationship between perceived discrimination and psychological well-being has been supported by multiple researchers in different contexts (Araújo & Borrell, 2006; Flores, Tschann, Dimas, Bachen, Pasch, & De Groat, 2008; Kessler, Mickelson, & Williams, 1999; Mays & Cochran, 2001; Pascoe & Richman, 2009; Schmitt, Branscombe, Postmes, & Garcia, 2014). Pascoe and Richman (2009) have done a meta-analysis of 192 studies that provided relevant insights about the influence of perceived discrimination on the health of individuals. They found that persistent perceived discrimination could lead to chronic stress, which enhances individuals’ vulnerability to mental illness. The study concludes that increased levels of discrimination are associated with more negative mental outcomes, such as depressive symptoms and psychiatric distress. In addition, Schmitt et al. (2014) found that mistreatment and relatively poor life events resulting from discrimination represent threats to psychological well-being. Consequences are especially harmful when disabled individuals perceive this kind of mistreatment and disadvantages as discrimination (Schmitt & Branscombe, 2002). Goffman (as cited by Schmitt et al., 2014) explained this by the fact that perceiving the self as a target of discrimination threatens the self-concept, which emerges through social interactions, such that others’ view of the self is. Major, McCoy, Kaiser, and Quinton (2011) add to this that psychological well-being is dependent on inclusion by others, and the perception that one is valued by others. So, if one perceives to be discriminated and thus not valued for whom they
are because of, in this case, their disability, this harms the psychological well-being. Based on this evidence from extant literature, we hypothesize:
Hypothesis 2a: the level of everyday discrimination as perceived by disabled individuals is negatively related to the psychological well-being of those individuals.
Based on the literature that has been examined in the previous chapters, a lot of support has been found to assume that perceived discrimination is related to the labor market status of disabled individuals. Since perceiving the self as a victim of discrimination has serious implications for mental health, as discussed previously, it seems likely that the relationship between perceived discrimination and LMS (labor market status) is partially mediated by psychological well-being. However, in order for this to be true, a positive relationship should exist between psychological well-being and LMS. This relationship has been studied before, but usually the causality of the relationship is expected to be the other way around. However, the possibility that poor psychological well-being leads to long periods of unemployment has left the question of causality unresolved (Diette, Goldsmith, Hamilton, & DarityJr., 2012). As already mentioned, persistent perceived discrimination of individuals could result in mental and physical illness (Pascoe & Richman, 2009). Also, it has been shown that many disabled individuals, who have some kind of visual, hearing, mobility, or mental impairment, do not participate in the labor market although they want to (CBS, 2016; Ali et al., 2010). Ali et al. (2010) have found that people with disabilities are less likely to actively seek employment because they are pessimistic about finding a suitable job. Since discrimination has been found to affect several employment outcomes for disabled individuals, and even has been found to be a barrier to employment for disabled individuals, it is very likely that disabled individuals who perceive to be discriminated by others at a constant basis, do not participate in the labor market because this perception of discrimination worsens their psychological well-being.
Schmitt et al., (2014) argue that discrimination can harm psychological well-being by making it hard to find a good job or apartment, for example. However, since the causality in the relationships suggested has not been confirmed in research (Diette et al., 2012), it might as well be the case that discrimination does not harm well-being through the inability to find a job, but that perceived discrimination makes it hard to find a job because it harms psychological well-being. This leads to the next hypotheses:
Hypothesis 2b: psychological well-being of disabled individuals is positively related to their participation in the labor market.
Hypothesis 2c: psychological well-being of disabled individuals mediates the relationship between perceived everyday discrimination and participation in the labor market, such that perceived everyday discrimination leads to lower psychological well-being, which in turn increases the likelihood of disabled individuals to be unemployed. In other words, perceived discrimination has an indirect effect on LMS through psychological well-being.
2.4 The Role of Self-Esteem
Besides well-being, the existing literature shows that other factors might influence the direct relationship expected in the research model. One of these factors is self-esteem, which refers to an attitude, specifically toward oneself (Coopersmith, 1967; Rosenberg, 1965). This attitude is based on an evaluation of oneself in the light of the beliefs one has about the self (Leary et al., 1995). This kind of self-esteem can also be referred to as personal self-esteem. For the current study it is essential to distinguish two types of self-esteem typically used in the existing literature, namely personal, and collective or group self-esteem. Collective or group self-esteem refers to the value attached to the membership of a particular group (Ellemers,
Kortekaas & Ouwerkerk, 1999). It is important to specify the difference between the two types of self-esteem for this study, because the kind of self-esteem measured has implications for the effect of self-esteem on the outcomes resulting from, in this case, perceived discrimination (Corning, 2002; Bourguignon, Seron, Yzerbyt, & Herman et al., 2006; Major et al., 2011). In the current study, self-esteem will be addressed at the personal rather than at the collective level, because the kind of self-esteem measured is about disabled individuals feeling personally discriminated against in their daily life, and more specifically in their past and current work experiences.
In previous literature, the relationship between perceived discrimination and self-esteem has already been studied. However, most of these studies are directed to another target group than disabled people, often based on demographic factors, such as ethnicity or gender differences (Bourguignon et al., 2006; Corning, 2002, Snyder et al., 2010; Panchanasdeswaran & Dawson, 2011). Multiple studies have found a negative relationship between perceived discrimination and self-esteem, in different targeted minority groups (Crocker & Major, 1989; Bourguignon et al., 2006; Verkuyten, 1998; Waddell, 2006; Panchanasdeswaran & Dawson, 2011). Major et al., (2011) confirm this negative relationship, and suggest that the degree to which individuals respond to discrimination is dependent on a variety of personal, situational and structural factors. In addition, if an individual perceives his or herself as a victim of discrimination, this has a direct, negative effect on their personal self-esteem. The negative relationship between perceived discrimination and self-esteem has been found in studies sampling youth (Verkuyten, 1998; Waddell, 2006), ethnic minority groups, and women (Panchanasdeswaran & Dawson, 2011). The study by Major et al. (2011) is discussing the topic in general rather than for a specific target group. Because the relationship between perceived discrimination and self-esteem has been found for several different target
groups, it is expected that the positive relationship between perceived discrimination and self-esteem also occurs for disabled individuals. Consequently, the following is hypothesized:
Hypothesis 3a: The level of everyday discrimination as perceived by disabled individuals is negatively related to their self-esteem.
Furthermore, some previous literature has examined the relationship between self-esteem and labor market status. For example, Waddell (2006) and Carroll (2007) found that self-esteem and employment are significantly related, so that people with low self-esteem are less likely to be employed. In addition, research has provided evidence that causality is running from esteem to labor market outcomes (Waddell, 2006). According to the literature, the self-esteem system monitors reactions of others and alerts the individual to the possibility of social exclusion, through events that lead to either rejection or acceptance. If people are rejected or excluded for personal reasons (as is the case when one is perceiving to be discriminated because of their disability), self-esteem declines (Leary, Terdal, Tambor & Downs, 1995). Self-esteem possibly produces this kind of effects because it is associated with confidence and high expectations of success, and therefore also associated with optimism and lowered anxiety.
When self-esteem is high, people’s willingness to strive towards desired goals and to persist in the face of obstacles and setbacks is enhanced. Consequently, according to Leary et al. (1995), people with high self-esteem often work harder and perform better after initial failure than people with low self-esteem. Relating this to the labor market status of disabled individuals, it can be expected that those individuals with high levels of self-esteem are more likely to seek employment. According to the theory proposed by Leary et al. (1995), they should be more optimistic, less anxious and more willing to seek employment, despite possible failures or negative experiences that they have encountered in their lives, such as
perceiving to be the victim of discrimination. Based on this literature, the following is hypothesized:
Hypothesis 3b: Self-esteem of disabled individuals is positively related to their participation in the labor market.
Hypothesis 3c: Self-esteem mediates the relationship between perceived everyday discrimination and labor market status of disabled individuals, such that when perceived everyday discrimination is high, self-esteem is low, and the participation of disabled individuals in the labor market is also low.
Moreover, existing research has also looked at self-esteem as being a moderator between discrimination and other outcomes that might result from it. For example, Corning (2002) studied the moderating effect of self-esteem on the relationship between discrimination and psychological distress among women. The study showed that, although discrimination is usually connected to negative outcomes, these negative outcomes are weakened by a high self-esteem. This effect might be explained by the self-protecting role of self-esteem, which means that higher self-esteem serves as a protective function for discrimination, so that when self-esteem is high, the individual can protect itself from negative outcomes from perceived discrimination (Corning, 2002). In addition, self-esteem moderates responses to rejection, because it decreases the extent to which individuals translate negative social evaluations into negative self-evaluations (Ford & Collins, 2010). Based on this, it is expected that perceived discrimination is negatively related to LMS, but less so when the individual has a high degree of self-esteem. This leads to the following hypothesis:
Hypothesis 4: the relationship between perceived everyday discrimination and labor market participation of disabled individuals is moderated by self-esteem. More specifically, when disabled individuals perceive a low degree of discrimination, their self-esteem is higher, and they will be more likely to be employed than when perceived discrimination is high, which leads to low self-esteem and consequently they are more likely to be unemployed.
3. Method
In the following chapter, the empirical part of this study will be discussed. First, the characteristics of the sample will be presented. Next, the variables and the items and scales used to measure them are discussed, including their reliabilities. Finally, the approach of analysis and the expected results coming from this analysis will be outlined. In the appendix the complete questionnaire (in Dutch) can be found.
3.1 Sample
The sample of this study consists of Dutch people who suffer from any kind of disability, long-term or chronic illness, or any other disabling conditions, such as learning difficulties. A total of 214 people started filling out the questionnaire, of which 154 fully completed it. 83.6% of the sample was female, and the participants were between 16 and 70 years old, with an average of 39.3 (SD=.37). 40.5% of the sample was unemployed, 59.5%% was employed. Of those participants who were employed, 28.7% has been working at their current job for less than 1 year, 14.8% for one year, 23.1% between two and five years, and 33.6% has been working at their current job for more than five years. 69.7% of them is moderately to very satisfied with their job, and only 11.4% of them is somewhat to very dissatisfied with their job. Of those unemployed, 19% has been unemployed for less than a year, 14.3% for one year, 26.1% between two and 5 years, and 34.5% for more than five years. Of the participants
who indicated to be unemployed, 40.5% is seeking employment, whereas the other 59.5% is not seeking employment. Those not seeking employment indicate that this is in 2% of the situations due to family duties, in 2% of the situations due to being victimized by discrimination, 70.6% of those not seeking employment indicate that they are incapacitated and are unable to work because of that. 25.5% of participants indicated that the reason for not seeking employment is not any of these, but is due to another reason, such as studying, low chances for people with disabilities, temporary inability to work. Examples of this are study, inability to stand or sit for a long time, poor chances in the labor market, no experience. In terms of education, 0.9% of the sample completed only primary education, of 11.7% of the sample secondary school is their highest level of education, 27.6% has an intermediate vocational education, 42.5% has a higher vocational education, and 17.3% of the sample has a university degree. Furthermore, 31.8% has a household size of one person, and thus lives alone. 36% of participants has a household size of two to five persons, and 32.2% has a household size larger than two persons. Additionally, the household income relative to the average Dutch income per household is low for 46.1% of the participants, average for 40.2%, and high for 13.7%. In order to be able to explore the differences between disabled individuals who are employed and those who are unemployed, no distinction has been made in the sampling between employed and unemployed people. Additionally, people with different kinds of disability conditions or illnesses could self-selected to participate in the study by filling out the questionnaire. These conditions can vary from visual or hearing impairment, to mobility impairment, but also mental or emotional conditions, including depression, autism, learning difficulties such as dyslexia. A general question applicable to different kinds of disabilities ask participants whether they have difficulty with independently undertaking daily activities, which was answered yes by 29.9% of the sample. Furthermore, 6.3% of the sample was deaf or had difficulty hearing, 11.9% was blind or had difficulty
seeing. A total of 37.6% indicated to have mobility impairment, of which 17.8% had difficulty with getting dressed or bathing by themselves. 23.3% indicated to have difficulties with moving, even when specialized equipment is used. Moreover, 25.2% of the sample indicated to have learning difficulties, 22.9% indicated to have a mental or emotional condition, and 9.4% had difficulties with speaking..
3.2 Procedure
To reach the relevant population to fill out the survey, the survey is shared through social media channels Facebook and LinkedIn so sample members can identify themselves and decide to fill out the survey, which is a convenience sampling method. On Facebook, the survey has been posted in several Facebook groups for people with disabling conditions. Also, a total of 27 organizations working with or for disabled people, or helping other organizations in becoming more inclusive in terms of hiring disabled employees, have been contacted. These companies have been found through an extensive online search. Of these 27 companies 26% did not respond, even after a follow-up e-mail. Only 15% of the organizations responded positively, willing to post the survey on their website and/or social media channels, or granting permission to post the survey on a forum for people in the relevant target group. Those who responded negatively did so for various reasons, such as lack of time, lack of capacity, but also for the protection of their employees or customers to prevent them from being overloaded with survey requests.
Before taking the survey, participants were ensured that their participation is anonymous. Also, the contact details of the researcher were included in case participants had any questions about the survey or the study as a whole.
3.3 Measures
The survey asked respondents for their demographics such as gender (nominal variable), age (ratio variable), and educational level (ordinal level), which are the control variables in the
analysis. Additionally, the kind of disability of the respondent will be assessed through some questions, indicating whether they have a hearing, visual, mobility or cognitive impairment. For the other constructs existing and validated scales will be used differing from 4-point, 5-point and 7-5-point scales (completely disagree – completely agree) at interval level. Also, some items are rated on the basis of yes/no (nominal level). The survey items were translated into Dutch because the survey will be filled out by Dutch people and will be better understandable for them that way. After translating, the Dutch translations were translated back into English to check if this translation was similar to the original items. For the SF-36 well-being scale, an existing translation was readily available (Van der Zee & Sanderman, 1993).
Labor Market Status is measured through one item, at nominal level, indicating whether the respondent is either employed or unemployed (Cole et al., 2009).
Perceived everyday discrimination is measured using a scale for everyday discrimination (Williams et al., 1997) consisting of 9 items is used (α = .88). Only 8 items were used, since one item, namely: ‘Do you feel like people act as if you are dishonest on a daily basis?’ was also more particularly about people from different races and therefore not directly relevant for this study. Participants were asked how frequently they experienced certain acts of discrimination on a daily basis, such as: being treated with less courtesy than others, people who act as if you are not smart, being threatened or harassed’. Another two items (r = .45) are used from Bourguignon et al. (2006) to measure perception of discrimination. This scale was used by Bourguignon et al. (2006) to measure the perceived discrimination of women, but in this study the items have been adjusted to disabled individuals. An example item is: ‘To what extent do you agree with the statement: Sometimes I feel left out because of my disability?’. Since two different scales are combined to measure
perceived discrimination, an exploratory factor analysis was done to test whether all items load on one construct. The results of the factor analysis can be found in Chapter 3.4.
Self-esteem is measured through the widely used Rosenberg Self-Esteem Inventory (as found in Bourguignon et al., 2006). This scale consists of 10 items on a 4-point scale (α = .86), indicating to what extent participants agree with each statement. Examples of statements are: ‘Generally, I am satisfied with myself’, ‘Sometimes I feel like I am worth nothing’ and ‘I would want to have more respect for myself’.
Well-being is measured through 9 items from the SF-36 Health Survey, of which a Dutch translation was already available (α = .84) (Van der Zee & Sanderman, 1993). The 9 items used from this scale measure the psychological and vitality aspects of well-being by asking participants to rate how often in the past 4 weeks they felt, for example, nervous, energetic, and tired on a scale from 1 (All the time) to 5 (Rarely). Four of the items were counter-indicative, and were therefore recoded.
Control variables used in the present study are age, gender and educational level. Age is measured at a continuous level, and gender is indicated as 0 (Female) and 1 (Male). Educational level was measured at five levels, 1 being primary educational level and 5 being university level. Since the first category contained only 2 responses, the first and second categories (secondary education) were combined so that four categories remained, being 1 (Low educational level), 2 (Medium educational level), 3 (High educational level), 4 (Very high educational level).
3.4 Data Preparation
In order to prepare the data for analysis, counter-indicative items were recoded in order to be able to interpret the data correctly. Furthermore, missing values were coded so that SPSS could recognize them and exclude them from the analysis in a pairwise manner in order to keep as much of the data as possible. Also, the data was checked for any errors and outliers by
running frequencies for all items using Z-scores. No errors or outliers were found. Moreover, since participants could select themselves to participate in the study, it was needed to make sure that all participants responded positively to at least one of the questions indicating disability or illness. Therefore, cases responding to all these items negatively were selected, which resulted in zero cases to select. This means that all participants had some kind of disability or illness.
Additionally, the internal consistency of the variables perceived everyday discrimination, psychological well-being and self-esteem were checked through Cronbach’s alpha, which assesses if the items combined to one scale truly measure the same construct, or if some items should be excluded from the analysis to improve the reliability of the scale. All variables have a Cronbach’s alpha >.8, which indicates a high level of internal consistency, as can be seen in Table 1. Moreover, because perceived everyday discrimination was measured using a combination of two different scales, an exploratory factor analysis has been done in addition to the reliability test to test whether all items load on one factor. The Kaiser-Meyer-Olkin measure verified the sampling adequacy for the analysis, KMO = .872. Bartlett’s test of sphericity Chi2 (45) = 897.035, p<.001, which indicates that correlations between items are sufficiently large for factor analysis. Total variance explained is 52%, and all factors are loading on the component with a score above .5. Based on the factor analysis, we can conclude that the items can be used together to measure the variable perceived everyday discrimination.
Finally, after conducting these preliminary steps, variables were created of the items used to measure the constructs. In order to do so, the means of all items measuring the same construct were used to describe a variable. When interpreting the means, the scales on which the variables are measured need to be taken into account. After this, Pearson correlations were generated and analyzed to obtain an initial overview of the relationships between the
variables. A significance level of p < .05 is applied to all analyses. Means, standard deviations, correlations and reliability are presented in Table 1.
Before analyzing the data, the variables were tested for normality. Firstly, the independent variable perceived everyday discrimination is quite strongly skewed towards the left (S=.48, SE=.19), but not particularly pointy or flat (K=-.004, SE=.37).. Next, the distribution of psychological well-being showed a slight tendency towards the right (S=-.32, SE=.20) and is flatter than normal (K=-.42, SE=…). Finally, self-esteem has a slightly negatively skewed distribution (S=-.37, SE=.19), which means it has a tendency towards the right. Additionally, the distribution is very flat (K=-.77, SE=.38). In addition to the skewness and kurtosis, histograms and Q-Q plots were checked to confirm the form of the distribution based on the skewness and kurtosis. Although the distributions of these three variables show some skewness and kurtosis, Tabachnick and Fidell (2001) have stated that skewness and kurtosis are acceptable within a range from -1 to 1.Additionally, they argue that with large sample sizes, skewness and kurtosis do not make a substantive difference.
4. Results
In the following chapter, the results from the analyses will be discussed in relation to the predicted effects. Firstly, the results from the correlations (see Table 1) will be discussed. Next, the analysis of the direct relationship between perceived discrimination and LMS will be discussed, followed by the indirect and conditional effects.
4.1 Correlations
In Table 1, the Pearson’s correlations between the variables are presented, which were found through SPSS. This table also includes the means, standard deviations, and reliability scores of the variables. This table shows moderate relationship between self-esteem and LMS has been found (r=.18, p<.05). The relationship is positive, meaning that when self-esteem
increases, LMS increases as well, but not a lot since the correlation is relatively low. In this case, this indicates that when self-esteem is high, individuals with disabilities are more likely to be employed, which is in line with the expectations based on the literature. Furthermore, a moderate, negative relationship was also found between self-esteem and perceived everyday discrimination, which is again in line with our expectations (r=-.20, p<.05). This relationship is also relatively low, but it is significant. Furthermore, psychological well-being and perceived discrimination are also moderately strongly related, and, as expected, the relationship is negative so that a high degree of perceived discrimination indicates a low degree of psychological well being (r=-.32, p<.01). Interesting to note is that self-esteem and psychological well-being are strongly, and positively related (r=.58, p<.05). Against our expectations, no statistically significant correlation was found between psychological well-being and LMS.
Table 1. Means, Standard Deviations, Correlations and Reliability (n=162)
Variables M SD 1 2 3 4 5 6 7 1. Age 36.29 13.36 - 2. Gender .16 .37 .31** - 3. Educational level 3.64 .93 -.09 -.05 - 4. LMS .60 .49 -.07 .08 .04 - 5. Psych Well-being 3.38 .82 .11 .14 .16* .10 (.88) 6. Self-esteem 5.03 1.18 .09 .01 .14 .18* .58* (.87) 7. Perc Discrimination 2.97 1.27 .04 .01 -.03 -.15 -.32** -.20* (.89) **. p<0.01, *. p<0.05
4.2 Regression Analysis
4.2.1 Direct effect.
Hypothesis 1 has been tested through a hierarchical binary logistic regression analysis in SPSS, of which the results are presented in Table 2. In this regression analysis, the direct effect of the independent variable perceived everyday discrimination on LMS has been tested. In the regression analysis, three control variables were added: gender, age, and educational level. These variables were added to the model in the first step, and in the second step the independent variable perceived everyday discrimination was added to the model to measure the direct effect of this variable on the LMS of disabled individuals. Hypothesis 1 states that perceived everyday discrimination is negatively related to LMS. The logistic regression analysis shows that adding perceived everyday discrimination does not significantly improve the ability of the model to predict the outcome variable (Chi-square=2.65, df=1, p>.05). Although the coefficient of perceived discrimination suggests a negative relationship between perceived discrimination and LMS, this relationship is not significant (p>.05). Therefore, Hypothesis 1 is rejected.
4.2.2 The mediating effect of psychological well-being.
Next, the mediation analyses will be discussed. Firstly, the indirect effect of perceived discrimination on LMS through psychological well-being is tested, using Model 4 of the Process Macro for SPSS by Andrew Hayes (2013). This mediation analysis tests three hypotheses simultaneously, which are hypothesis 2a, 2b and 2c. The results of the mediation analysis are presented in Table 3 and Table 4, of which Table 3 shows the direct effects needed for the calculation of the mediating effect, and Table 4 shows the direct, total, and indirect effect.
Table 2. Hierarchical Logistic Regression Model of Labor Market Status Step 1 B SE Wald OR CI Age -.02 .013 2.49 .98 Gender .94 .498 3.60 2.57 Educational Level 3.15 Low .05 .60 .01 1.05 [.32-3.40] Medium -.41 .49 .70 .66 [.25-1.73] High .31 .46 .46 1.37 [.55-3.36] Step 2 Age -.02 .01 2.44 .98 Gender .97 .50 3.69 2.64 Educational Level 3.01 Low .04 .61 .00 1.04 [.32-3.42] Medium -.39 .50 .62 .68 [.26-1.79] High .32 .47 .47 1.38 [.55-3.43] Perceived Discrimination -.21 .13 2.56 .81
Note. Reference category for educational level was Very High. R² = .07 (Cox & Snell), .09 (Nagelkerke). Model χ² (7) = 11.43, p >.05. OR = Odds ratio, CI = confidence interval.
The effect of perceived discrimination on psychological well-being a1 = -0.20 means that two disabled individuals that differ by one unit on perceived discrimination are estimated to differ by -0.20 units on psychological well-being. The sign of a1 is negative, meaning that those relatively higher in perceived discrimination are estimated to be lower in their psychological well-being. This effect is statistically different from zero, t = -4.08 p < .001, with a 95% confidence interval from -0.30 to -0.10. This means that Hypothesis 2a, stating that perceived
discrimination is negatively related to psychological well-being, is confirmed by the data. The effect b1= 0.09 indicates that two disabled individuals who experience the same level of perceived discrimination but that differ by one unit in their level of psychological well-being are estimated to differ by b1 = 0.09 units in LMS. The sign of b1 is positive, meaning that those relatively higher in psychological well-being are estimated to be more likely to be employed than those lower on psychological well-being. This effect is not statistically different from zero, Wald χ² = 0.37 , p = .711, with a 95% confidence interval from -0.37 to .54, which includes zero. Therefore, Hypothesis 2b, suggesting that psychological well-being is positively related to LMS, cannot be confirmed by the present study.
The indirect effect of -0.02 means that two workers who differ by one unit in their reported perceived discrimination are estimated to differ by -0.02 units in their reported labor market status as a result of the tendency for those who perceive to be discriminated on a daily basis to have lower psychological well-being, which in turn translates into a decreased likelihood of being employed. This indirect effect is not statistically different from zero, which is also revealed by a 95% BC bootstrap confidence interval that contains zero (-0.13 to 0.08). This means that Hypothesis 3c should be rejected, so that we cannot confirm the expected indirect effect of perceived discrimination on LMS through psychological well-being.
The direct effect of perceived discrimination, c′ = .20, is the estimated difference in LMS between two disabled individuals with the same level of psychological well-being, but who differ by one unit in their reported perceived discrimination, meaning that the person perceiving more discrimination but who has equal psychological well-being is estimated to be 19.5% more likely to be employed. This direct effect is not statistically different from zero, Wald χ² = -1.34, p = .181, with a 95% confidence interval from -0.48 to 0.09.
The total effect of perceived discrimination on LMS is c = -.22, meaning two disabled individuals who differ by one unit in perceived discrimination are estimated to be differ in LMS in a way that if one has perceived discrimination of 1 unit more than the other, that person is 22.0% less likely to be employed, which is indicated by the negative sign. This effect is not statistically different from zero, Wald χ² = 2.59, p = .108, or between -.48 and 0.09 with 95% confidence.
Table 3. Mediation Analysis for Well-Being as a Mediator of the Relationship Between Perceived Discrimination and LMS (n=138)
Consequent Well-being (M)
LMS (Y) Antecedent Coeff. SE p Coeff. SE p
Perceived Discrimination (X) a1 -.20 .05 .00 c' -.20 .15 .18 Well-being (M) b1 .09 .24 .71 constant i1 3.33 .31 .00 i2 .65 1.15 .57 R² = .17 R² = .07 F(3, 139)= 6.86 , p<.001
Model log likelihood = 8.05, p>.05
Note. In logistic regression for LMS Nagelkerke’s R² is used.
Table 4. Regression Model for Well-Being as a Mediator of the Relationship Between Perceived Discrimination and LMS
Effect SE p LLCI ULCI Direct effect c' -.20 .15 .18 -.48 .09 Total effect c1 -.22 .13 .11 .62 1.05
Boot SE Boot LLCI Boot ULCI Indirect effect a1*b1 -.02 .05 -.13 .08
4.2.3 The mediating effect of self-esteem.
Furthermore, the indirect effect of perceived discrimination on LMS, through self-esteem is tested. In order to test this effect, three different hypotheses need to be tested. This mediation analysis is again tested through the Process Macro for SPSS, using Model 4 (Hayes, 2013). The results of the mediation analysis are presented in Table 5 and Table 6 of which Table 5 shows the direct effects needed for the calculation of the mediating effect, and Table 46shows the direct, total, and indirect effect.
The effect of perceived discrimination on self-esteem a1 = -0.19 means that two disabled individuals that differ by one unit on perceived discrimination are estimated to differ by -0.19 units on self-esteem. The sign of a1 is negative, meaning that those relatively higher in perceived discrimination are estimated to be lower in their self-esteem. This effect is statistically different from zero, t= 8.43, p = .01, with a 95% confidence interval from -0.34 to -0.04.
The effect b1= 0.31 indicates that two disabled individuals who experience the same level of perceived discrimination but that differ by one unit in their level of self-esteem are estimated to differ by b1 = 0.31 units in LMS. The sign of b1 is positive, meaning that those relatively higher in self-esteem are estimated to also be more likely to be employed. This effect is also statistically different from zero, Wald χ²= 2.07, p = .038, with a 95% confidence interval from .02 to .60.
The indirect effect of -0.06 means disabled individuals with a one unit increase in their reported perceived discrimination are estimated to be less likely to be employed as a result of the tendency for those who perceive to be discriminated on a daily basis to have lower self-esteem, which in turn translates into a decreased likelihood of being employed. This indirect effect is not statistically different from zero. However, the 95% BC bootstrap confidence interval from -0.17 to -0.01 does not contain zero, which does indicate a significant direct
effect. Because both Hypothesis 3a and Hypothesis 3b are confirmed by the data, and the 95% BC bootstrap confidence interval does not contain zero, the indirect effect is supported by the results. This is despite the fact that the p-value is insignificant, because the distribution of the interaction effect is often not normal. Bootstrapping is a nonparametric procedure that does not make any assumptions regarding the normality of the distribution of the indirect effect (MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002; MacKinnon, Lockwood, Williams, 2004). All in all, this means that the mediation effect is supported by results.
The direct effect of perceived discrimination, c′ = -.14, is the estimated difference in labor market status between two disabled individuals with the same level of self-esteem, but who differ by one unit in their reported perceived discrimination, meaning that the person perceiving more discrimination but who has equal self-esteem is estimated to be 14.0% less likely to be employed. This direct effect is not statistically different from zero, Wald χ² = -.99, p = .318, with a 95% confidence interval from -0.42 to 0.14.
The total effect of perceived discrimination on LMS is c = -.22, meaning two disabled individuals who differ by one unit in perceived discrimination are estimated to differ by -0.22 units, but since LMS is a binary variable, this should be interpreted as a 22.0% difference in the likelihood of being employed, so that a higher degree of perceived discrimination leads to a decrease in the likelihood of being employed. The negative sign means the person perceiving greater levels of discrimination is less likely to be employed. This effect is not statistically different from zero, Wald χ² = 2.59, p = .108, or between -.48 and 0.09 with 95% confidence.
Table 5. Mediation Analysis for Self-Esteem as a Mediator of the Relationship Between Perceived Discrimination and LMS (n=153)
Consequent
Self-esteem (M)
LMS (Y) Antecedent Coeff. SE p Coeff. SE p
Perceived Discrimination (X) a1 -.19 .08 .01 c1' -.14 .14 .33 Self-esteem (M) b1 .31 .15 .04 constant i1 4.71 .46 .00 i2 -.33 1.08 .76 R² =.08 R²=.09 F(4, 148)= 3.06 , p<.05
Model log likelihood = 10.70, p>.05
Note. in logistic regression for LMS Nagelkerke’s R² is used.
Table 6. Regression Model for Self-esteem as a Mediator of the Relationship Between Perceived Discrimination and LMS
Effect SE p LLCI ULCI Direct effect c' -.14 .14 .32 -.42 .14 Total effect c1 -.22 .13 .11 .62 1.05
Boot SE Boot LLCI Boot ULCI Indirect effect a1*b1 -.06 .04 -.17 -.01
4.2.4 Conditional effect.
Finally, in order to test Hypothesis 4, data were analyzed using Model 1 from the Process Macro for SPSS (Hayes, 2013). Before running the analysis, the independent variable perceived discrimination and the moderating variable self-esteem were centered around zero by subtracting their means (Aiken, West, & Reno, 1991). In order to test for the moderating relationship, Process ran a hierarchical logistic regression, of which the results are presented
in Table 6. The regression coefficient for perceived discrimination*self-esteem is c3=-.104 and is not statistically different from zero, Wald χ² = -.91, p>.05. Thus, the effect of perceived discrimination on LMS does not depend on the participant’s self-esteem. Moreover, this model accounts for 9.71% of variance in labor market status according to Nagelkerke’s R2. Altogether, since the interaction term is not significant, the expected moderation effect is not found, which means that Hypothesis 4 cannot be confirmed.
Table 7. Moderation Analysis for Self-esteem as Moderator of the Relationship between Perceived Discrimination and LMS (n=153)
Coeff. SE Wald p Intercept i1 .81 .74 1.09 .28 Age -.02 .01 -1.47 .14 Gender .85 .49 1.72 .09 Educational Level .04 .19 .20 .84 Perceived Discrimination (X) c1 -.17 .14 -1.14 .25 Self-esteem (M) c2 .31 .15 2.07 .04 Perceived Discrimination* Self-esteem (XM) c3 -.10 .11 -.91 .36 R²: Cox&Snell=.07, Nagelkerke=.10 p>.05 Model log likelihood = 11.52, p>.05
5. Discussion
The current study aimed at investigating whether perceived everyday discrimination by disabled individuals influences their employment status, and if this relationship was indirectly or conditionally affected by psychological well-being or self-esteem. Expectations were a negative relationship between perceived discrimination on labor market status, mediated by both psychological well-being and self-esteem. Additionally, self-esteem was expected to
have a moderating effect on the relationship between perceived discrimination and labor market status.
The results do not support the expected negative relationship between perceived discrimination and labor market status. The expected negative relationship between perceived everyday discrimination and labor market status was based on the extant literature that found a significant negative relationship between perceived discrimination and employment outcomes for people with disabilities, for example in gaining employment, but also in integrating in organizations (Stone & Colella, 1996). Because the effect on job-related outcomes has been found in different studies it is also seems very likely that perceiving to be the victim of discrimination can be a barrier to seek employment (Ali et al., 2010; Baldwin & Johnson, 2006; Blanck, 2001; Schur et al., 2009). Unfortunately, this cannot be confirmed in the current study.
However, as expected, the indirect effect of perceived discrimination on LMS was supported by the data. This means that, although no statistically significant direct effect of perceived discrimination on LMS could be a found, there is a relationship between perceived discrimination and LMS, which is carried by self-esteem. However, this effect is relatively small. Unfortunately, although the direct relationship between perceived discrimination and psychological well-being was supported, the relationship between psychological well-being and LMS was not. Overall, the indirect effect of perceived discrimination on LMS through psychological well-being was not supported. Furthermore, no significant moderating effect was found of self-esteem on the direct relationship between perceived discrimination and labor market status. This expected moderating effect was based on the fact that negative outcomes of discrimination are weakened by high self-esteem (Corning, 2002). The unexpected results could be due to a number of reasons. Firstly, the fact that the expected moderation effect was not supported can be explained by the fact that self-esteem in itself is
influenced by perceived discrimination and is also significantly related to LMS. Secondly, perceived discrimination is a very broad construct that can be measured in many different ways and used for many different kinds of discrimination. In this study, perceived everyday discrimination is measured, but other kinds of discrimination, for example the more acute discriminatory stressors (Araújo & Borrell, 2006) could have a different effect on LMS than the chronic stressors investigated in the current study. The items used to measure the degree of perceived discrimination were not obtained from studies about disabled individuals, but rather about other minority groups based on gender, race, ethnicity or other demographic factors (Araújo & Borrell, 2006; Bourguignon et al., 2006; Waddell, 2006).
5.1 Theoretical Implications and Suggestions for Future Research
The findings of this study have several implications for theory. The fact that no significant direct effect of perceived discrimination on labor market status was found could indicate that the effect is different for disabled individuals than for other marginalized groups, such as women, ethnic minorities, and gay individuals (Bourguignon et al., 2006; Corning, 2002, Snyder et al., 2010; Panchanasdeswaran & Dawson, 2011). The fact that not all expected relationships were supported could be due to the fact that the current study was based on one group of people with many different kinds of impairments, including visual, hearing, mobility and mental impairments. Future research should study the effects for separate groups based on different kinds of impairments, because it might be the case that results differ among these groups, which could explain why not all relationships were supported in the study. Unfortunately, this was beyond the scope of the present study because a much larger sample size would have been necessary to examine the differences among groups.
The finding that self-esteem mediates the relationship between perceived discrimination and labor market status contributes to literature because no previous literature has studied this mediating effect. However, the direct relationship that was found for other
