STEREOTYPING ELDERLY EMPLOYEES:
WHAT ARE THE CONSEQUENCES OF A NEGATIVE ATTITUDE?
Master Thesis, MScHRM
University of Groningen, Faculty of Economics and Business
Stereotypes are ideas formed about groups of people based on common characteristics. This study investigates the effects of negative stereotypes as experienced by elderly employees on their commitment to the company and their intention to continue working. Using the Cognitive-Affective Personality System and literature on stereotyping we hypothesized that the negative effect stereotyping has on commitment and intentions to continue working, is mediated by the quality of exchange relationships employees have with their colleagues at work, measured by the Co-worker exchange (CWX) scale. We tested hypotheses using data gathered among 108 employees of a Dutch transportation company. Results indicate a strong negative relationship between the amount of negative meta-stereotypes with commitment and the intention to continue working. Partial
mediation by CWX was found. Theoretical and practical implications of these results are discussed and recommendations for future research on stereotyping and ways to influence it are made.
INTRODUCTION
The aging of the population is a problem which has received a lot of attention lately; the average age of the labor force is slowly rising. The United Nations (2002) have projected that the median age of employees will raise from 29.2 in 1950 to 49.5 years old in 2050. This increase in average age leaves us as a society with a variety of issues. One of them is addressed in the Lisbon targets by the European Union in 2000; the rising percentage of elderly and the declining percentage of younger employees will result in a big increase in social security costs and difficulty in the payment of pensions. To prevent these problems, the EU specified in the Lisbon targets that each member state should increase the
employment rate of the elderly. Unfortunately, not a single member state succeeded (EurActiv, 2009). The Dutch government has tried to raise the percentage of older employees by rising the obligatory retirement age from 65 to 67. This solution has found a lot of resistance among employees and unions. To increase the problem; even though the official retirement age is currently still set at 65, the majority of older employees retire years before reaching this age (CBS, 2009). Where governments are trying to increase the inclusion of the elderly, companies mostly do not see the advantage of pursuing the same goal (CBS, 2009). Many companies replace their older employees by younger ones, based on the common stereotype that older employees are slower to learn, less adaptable and overall more expensive (e.g. Nelson, 2002; Scholl & Sabat, 2008). In reality the reverse is often true; elderly are more experienced, committed to their job and require less guidance and mentoring (e.g. Hess et al., 2003). Yet, these negative
stereotypes exist, and they affect a companies’ workforce, not just at the managerial level, but also on a personal level. Research into the effects of stereotyping has shown that elderly employees in an open culture of acceptance perform better than in one where they are negatively stereotyped (e.g. Chiu et al., 2001; Sinclair et al., 2006; Scholl & Sabat, 2008).
The personal basis of stereotyping can be divided into two different concepts; self-stereotyping (Hess et al., 2003) and meta-self-stereotyping (Vorauer et al., 1998). Self-stereotypes are the Self-stereotypes a person holds concerning his own group,
outcome variables such as commitment, the intention to retire, performance, etc. (Desmette & Gaillard, 2008; Scholl & Sabat, 2008; Chiu et al., 2001). Desmette and Gaillard (2008) did research into the effects of self-categorization as an older worker, a concept very similar to self-stereotyping, on several outcome variables such as the desire to retire early and found an inverted relationship. Scholl and Sabat (2008) did research on memory performance of Alzheimer patients, and found that negative stereotyping, or merely the threat of being negatively stereotyped had adverse effects on the performance of healthy elderly people. Furthermore, research into concepts similar to
meta-stereotyping, such as a supportive climate for older workers, concluded that a negative climate resulted in a higher intention to leave the company (Armstrong-Stassen & Schlosser; 2007). These negative stereotypes exist and it is affecting the well-being elder employees experience at work. In this study we will try to find how stereotyping affects the commitment elderly employees feel towards the organization and how it affects elderly employees in their decisions to continue working.
As mentioned before, the personal basis of stereotyping exists of self-stereotyping and meta-stereotyping. Both concepts find their origins in the social context, since they can only exist in a situation where people interact with others. Concurrently, self- and meta-stereotyping can shape the relationships people enter at work; Koudenburg and Gordijn (2011) did research into the effects of stereotyping and role-expectancy, and found that stereotypes greatly influence the way people respond to certain behaviors, at and off work. Other researchers, such as Vorauer, Main and O’Connell (1998), who introduced the concept of meta-stereotyping, immediately connect it to intergroup relations.They conclude that the existence of negative stereotypes can deteriorate the relationships between the in-group and the out-group, but one-on-one interactions can partially remove these stereotypes. Simon and Hamilton (1994) make a connection between
self-stereotyping and the social context, and conclude that the social context can increase or decrease self-stereotyping, albeit that the causality of this relationship could not be proven.
social exchange relationships at work and organization-relevant outcome variables. The present research will try to unravel the black box of the relationship between stereotyping and organization-relevant outcome variables, by including the social context, in the form of the quality of social exchanges between an older employee and his co-workers. We expect to find a mediating role for the quality of social exchanges with co-workers in the relationship between self- and meta-stereotyping and organizational commitment and the intention to continue working.
Taken together, the present study aims to advance the understanding of the personal basis of negative stereotyping for older employees’ commitment and intentions to continue working. Specifically, a mediation model is examined that proposes self- and meta-stereotyping to be negatively related to organizational commitment and intentions to continue working through the intervening mechanisms of deteriorated exchange relationships with direct co-workers. This model is graphically represented in figure 1. Our study combines features from theories on stereotyping (e.g. Vorauer et al., 1998; Hess et al., 2003), and quality of social exchange (e.g. Seers, 1989) with ideas from literature on commitment (e.g. Meyer and Allen, 1991) and the intention to continue working (Armstrong-Stassen et al., 2007). As such, we suppose to make three unique contributions; firstly we differentiate self- and meta-stereotyping as two aspects of stereotyping to examine whether these two distinct aspects have differential relationships with older employees’ organizational commitment (affective and normative) and
intention to continue working; secondly, we integrate theoretical perspectives on
stereotyping and social exchange theory to examine whether and how the two aspects of stereotyping are related to the quality of exchange relationships older employees hold with their co-workers; and thirdly we combine insights from previous research on
THEORY AND HYPOTHESES Personal stereotypes
A lot has been written on stereotyping, and opinions differ, but most writers seem to come to the same broad definition of stereotyping; it’s a way of simplifying the world around you (e.g. Levy, 1996; Buelens et al., 2002; Cuddy & Fiske, 2002; Scholl & Sabat, 2008). While this is a sound definition, it is too broad for this research, especially since its focus lies on self- and meta-stereotyping. Yet, in order to understand these concepts better, we will first define stereotyping. Cuddy and Fiske (2002) define stereotypes as ‘cognitive structures that store our beliefs and expectations about the characteristics of members of social groups’ and stereotyping as ‘the process of applying stereotypic information’. Hereby stating that when we have established that the majority of a group possesses a certain characteristic, we tend to generalize this characteristic to the whole of the group.Buelens et al. (2002) give a similar but more elaborate explanation of
stereotyping. They state that stereotypes are formed through a four-stage process. It begins by categorizing people into groups according to various criteria, such as gender, age, race, and occupation. The second stage consists of assuming that all people in a certain category possess the same traits or characteristics. In the third stage expectations are formed of others, and their behaviour is interpreted according to the stereotypes that are found. The fourth stage is the maintenance-stage; stereotypes are maintained because people tend to overestimate the stereotypic behaviour and they often explain expected and unexpected behaviour incorrectly.
Both definitions of stereotyping are very similar, and shape the definition of stereotyping; stereotyping is a method of categorizing people around you; once a stereotype has been formed, it will be applied to all members of the categorized group.
Stereotyping is something everyone does, and which happens to everyone (Levy, 1996; Vorauer et al., 1998; Cuddy & Fiske, 2002; Lun et al. 2009) and it is not necessarily a bad thing. But research has shown (e.g. Nelson, 2002; Cuddy et al. 2005; Scholl & Sabat, 2008) that there is a general tendency to negatively stereotype the elderly. Nelson (2002) found that there are a lot of negative stereotypes concerning the elderly employees in the workplace. Often found stereotypes are that they are forgetful, unable to learn, and slow workers. Cuddy et al. (2005) concluded in their research that elderly are placed in the same mental category as the disabled and the retarded people; they score high on warmth, but low on competence. Even though people hold positive stereotypes of old age as well, negative stereotypes are more prevalent (Levy, 1996).
A question that rises is how does the fact that people tend to think so negative about elderly, affect the elderly themselves? Do they try to counter the image by proving everyone is wrong in their beliefs and expectations about them, do they reinforce the image by living up to the stereotypical beliefs or even worse; do they believe the
stereotypes, and adjust themselves to conform to the stereotype? Though this last option may sound a bit strange, research has shown (Steele & Aronson, 1995; Hess et al., 2003; Sinclair et al., 2006; Sinclair et al., 2009)that this is actually what happens. The process is better known as self-stereotyping.
Self-stereotyping is the process of applying stereotypes of the in-group to yourself (Hess et al., 2003). Though this is a correct definition of self-stereotyping, it doesn’t fit this research. As we are more interested in the negative side of self-stereotyping, we have to expand this definition. Self-stereotyping is a process which can originate from
With this information we can form a definition of self-stereotyping that will be used for this research: Self-stereotyping is the (implicit) process of applying negative
stereotypes of the in-group to yourself. These self-stereotypes are stereotypes members of the in-group (sub-)consciously apply to themselves, and believe them to be true. But not all elderly are the same, and not all believe the stereotypes that are readily available in the world. Even so, they have an expectancy of how other people stereotype them as a group. The stereotypes as members of the in-group expect out-group members to have are commonly known as meta-stereotypes (Vorauer et al., 1998).
Meta-stereotyping as a concrete variable was first introduced by Vorauer, Main and O’Connell (1998). They define meta-stereotyping as the beliefs a person has regarding the stereotype that out-group members hold about his or her own group. For example, an elderly employee may believe that his co-workers (the out-group members) see all elderly (the in-group) as forgetful and old-fashioned (stereotypic beliefs). Stated in the words of Lammers et al. (2008), meta-stereotypes are thoughts about “how I think that they think we are like”. Meta-stereotyping is different from self-stereotyping in the sense that meta-stereotypes are more externally focused. They have a necessary relational component; meta-stereotypes refer to an individual’s belief about how the group he or she belongs to is viewed by a particular out-group. Thus, without a social environment, meta-stereotypes cannot be present. Self-stereotypes focus on an individuals’ personal belief about the in-group, even without a social environment, self-stereotypes can be present. Another difference is that meta-stereotypes seem to be comparatively negative (Vorauer et al., 1998). This gives us some interesting information, where self-stereotypes can be both negative and positive, apparently the elderly tend to think that others are mostly negative about elderly. The next chapter will deal with the negative effects of this image.
Stereotyping and commitment and the intention to retire
a person ages, these stereotypes seem to become truer. This process leads someone to the acceptance of false stereotypes, i.e. negative self-stereotyping. Research by Crocker and Major (1989) explains that a generally seen coping response to social stigmatization is withdrawal. We can argue that social stigmatization can come from within, i.e. self-stereotyping, or is experienced from external parties, i.e. meta-stereotyping. Either way, when an employee feels himself unfit to cope with a situation (self-stereotyping), or thinks others perceive him as unfit to cope (meta-stereotyping), a certain discrepancy between self-identification and the situation occurs. A common-seen reaction to this phenomenon, is withdrawal from the in-group (e.g. Schlenker & Weigold, 1992; Chiu & Chan, 2001; Desmette & Gaillard, 2008). Unfortunately, even with our current options in plastic surgery, it’s practically impossible to withdraw yourself from the group ‘elderly’. The option that remains is to withdraw from the situations where the discrepancy occurs, i.e. stop working, or passively retract from work by losing commitment.
Commitment has always been a topic of discussion for many scientists. Some authors define it as a one-dimensional concept (e.g. Brown, 1996), others argue it is a multi-dimensional concept (O’Reilly & Chatman, 1986; Allen & Meyer, 1990; Meyer & Allen; 1991; Meyer et al., 1993). But over the years a consensus towards a general model seems to have evolved; Allen and Meyer (1990) developed a model that should grasp the entire spectrum of commitment. They concluded that the majority of different commitment theories can be merged into three different components; affective, continuance and normative commitment (Meyer & Allen, 1991). They state that affective commitment is a form of commitment in which an employee is committed because he or she feels an intrinsic desire to keep working at the organization. Continuance commitment is a form of commitment in which an employee maintains employment because he feels the cost of quitting the organization is higher than the cost of staying. Normative commitment causes employees to keep working because of extrinsic expectations. Their social surroundings expect them to be at work, and therefore they live up to this expectation (Allen & Meyer, 1990; Meyer et al., 1993).
Affective commitment, an intrinsic commitment to one’s organization, finds its origins in satisfaction with the work. An employee who feels appreciated, at place at his
self-image, whether it’s true or not, makes an employee feel out of place at the
organization. Negative stereotyping does exactly this; it creates a negative image inside the mind of the employee. Based upon this we expect to find that a negative self-image (self-stereotyping) and the suspicion that employees think negative about you (meta-stereotyping) lead to a diminishment in affective commitment.
Normative commitment seems to rely more on surroundings; it stems from what an employee expects of himself, or believes to be expected of him (Allen & Meyer, 1990). This resembles self- and meta-stereotypes; ideas on how we see ourselves, and what we think others expect of us. This resemblance in theory gives a good indication of a possible relationship between the constructs. Whereas Schlenker and Weigold (1992) explain that when more negative stereotypes are triggered withdrawal occurs, we hope to find that a higher amount of stereotypes causes an employee to expect less of himself, and to believe his co-workers expect less from him. The extrinsic expectations Allen and Meyer (1990) mention in their research will diminish, i.e. normative commitment lowers.
The last component of the commitment model, continuance commitment doesn’t fit the profile; it relies heavily on external factors. It’s not so much commitment towards a company, but a fear of not being able to find a suitable alternative. An employee commits himself to the organization, since he fears that any alternative would leave him worse off. This research focuses on factors that are internal to the employee, i.e. self- and meta-stereotyping. Therefore no relations between continuance commitment and self- and meta-stereotyping are expected and we decided to omit this component from the research model.
Where commitment is a good indicator of an employee’s willingness to work at an organization, the intention to continue working is much more straightforward. As one might expect from the name, it measures employees’ intent to continue working for the organization. The construct is very similar to Beehr’s (1986) construct of the intention to retire, which measures the intention to stop, or significantly reduce the time spent
retirement; the construct developed by Beehr does not take this into account. Research by Armstrong-Stassen et al. (2007) used a construct to measure the intention to continue working, which we will use in this research.
Previous research into (concepts similar to) stereotyping give an interesting insight into the results we can expect to find. Gaillard and Desmette (2008) investigated the identification with older workers, and the permeability of younger and older worker boundaries, respectively comparable to self-stereotyping and meta-stereotyping. They found strong relationships with the outcome variables early exit intention, affective organizational commitment and others. They call for future research into the effects of self-stereotyping, since identification with older workers combined with a negative attitude towards that group seemed to predict the outcome variables strongly. Turner and Hogg (1987) conclude that self-categorization as an older worker is a high predictor of early exit intentions. Self-categorization is very similar to self-stereotyping; an employee considers himself to be part of the group elderly, and thus ascribes himself all the
(negative) traits that belong to that group. This image-forming can deteriorate the well-being employees experience at work (Major et al., 1995; Witt et al., 1999). Based upon the theory and previous researches we expect to find a relationship between stereotypes and the constructs commitment and the intention to continue work. We formulate the following hypotheses:
Hypothesis 1: Negative self-stereotypes are negatively related to affective (H1a) and normative (H1b) commitment towards the company.
Hypothesis 2: Negative meta-stereotypes are negatively related to affective (H2a) and normative (H2b) commitment towards the company.
Hypothesis 3: Negative self-stereotypes are negatively related to the intention to continue working of the employee.
Co-Worker-Exchange (CWX) as a Mediator
Stereotypes are embedded in a social context. Self-stereotypes originate from media, friends, family, etc. and meta-stereotypes exist only because we interact with other people. In the previous paragraph we theorized stereotypes can have a negative effect on an employees’ intentions to continue working and his commitment towards the
organization. How the social context plays a role in this relationship will be the focus of this paragraph. Because we interact with others, stereotypes come to live. Without the possibility of others judging and evaluating, stereotypes would not exist (Vorauer et al., 1998). We shape expectations of the other, of how the other sees us, and how we respond to certain situations (Scholl & Sabat, 2008). We can conclude that stereotypes originate from social interactions. Meta-stereotypes are logically derived from social interactions, since these are thoughts about what you think they think we are like. But self-stereotypes have a similar social context. As Scholl and Sabat (2008) explain; self-stereotypes are subliminally formed by social media, friends and colleagues. As a person grows older, he finds these stereotypes become truer than before. This leads to acceptance of these stereotypes and with this process, unfortunately, false stereotypes are also accepted.
not valued. Whereas in a different setting without these stereotypes, you might have been fine with making these social exchanges. The CAPS-theory suggests that not only actual social exchanges influence a person’s character development; people activate their own dynamics by thinking about situations, by selective recall and reliving past occurrences and feelings (Mischel & Shoda, 1995). Where does this remind you of? Exactly, meta-stereotyping! Meta-stereotyping entails exactly that process, in which we ascribe others to have stereotypes about the in-group based on selective recall, reliving past occurrences and feelings. What the CAPS-theory advocates is that every living human being has a dynamic personality, not a static one.
The whole concept of a dynamic personality, based upon both internal and external processes is very well captured in the book ‘One, No One and One Hundred Thousand’ by Luigi Pirandello (1936). In this semi-autobiographic novel the author describes how the main character, Moscarda, becomes aware of all the different personalities he has in different social situations. These different personalities are given shape by the current situation, past encounters, the social partner, and the expectancies of the social partner. But Pirandello continues to describe that personality attributes even change based upon what the main character expects other people to expect of him. And by contemplating over all this, Moscarda starts to realize he is just not just a single person, i.e. himself, but that he is actually a hundred thousand different Moscarda’s, who all reside in his body. And because he is all these different persons, he is actually no one. As a result of his contemplations he tries to break with all the different attributes his surroundings have come to expect of him. Eventually he ends up in bedlam, since his direct environment believes he has gone insane. This being true or not, what Pirandello already described back in 1936, is a novelized version of the CAPS-theory. He realizes personality is not just a static whole of traits, but that it can and will adapt to context; a person’s character is shaped by who you are, by the situations you encounter, the expectancies you hold of other people and by social exchanges, both real and expected, or imagined for that matter.
1998; Vorauer et al. 2000; Koudenburg, 2011). Vorauer et al. (2000) state that
stereotypes are easily activated in a context where evaluation by an out-group member is possible. The possibility of evaluation alone triggers uncertainty, and heightened self-presentation efforts.Each participant feels the intrinsic need to behave congruently with the identity which is selected for the situation and acknowledge the identity selected by the social partner.In this situation people are looking to minimize the discrepancy between self-identification and the way the social partner identifies him. The larger the discrepancy, the higher his discontent will be (Schlenker & Weigold, 1992).
What the CAPS-theory can tell us, is that when an employee thinks negatively of himself, as is the case with self-stereotyping, or expects others to be negative, i.e. meta-stereotyping, his character attributes will adjust to the situation. Furthermore, the
character attributes of his social partner would adjust as well based upon the situation, as Pirandello (1936) encountered. An experienced negative attitude concerning the elderly would deteriorate the exchange relationship. It is proposed that the availability of
negative self- and meta-stereotypes, e.g. being old-fashioned, slow to learn, stubbornness, etc. would inhibit an older employee in proposing ideas, giving support and feedback and the sharing of information. And these are exactly the factors that determine the quality of the information exchanges between co-workers.
To capture the quality of exchange relationships at work, we use an adjusted version of Seers’ (1989) construct of Team-Member-Exchange (TMX). This construct entails the individual member’s perception of the exchange relationship between him and the colleagues of his team. The exchange relationship as described by Seers, exists at the team level and entails team members’ willingness to exchange information, support and assist ideas, and give help, feedback and recognition to one another. Where Seers studied at the team level, we are more interested in the personal side, since we focus on
interaction of the older employee with colleagues. Following previous research (Sherony & Green, 2002), we decided to replace the words ‘team member’ by ‘co-worker’ to form the Co-Worker-exchange construct (CWX).
Kozlowski, Chao and Gardner (1995) did research into the effect of newcomer expectations and role development on several outcome variables, such as quality of exchange relationships, commitment and the intention to retire. They found that the newcomer expectations had a relationship with commitment and intention to retire, as did quality of exchange relationships. Newcomer expectations are in a lot of ways very similar to meta-stereotyping. Expectancies about the job, co-workers and culture are made, just like with meta-stereotyping expectancies of stereotypes are made. A newcomer has certain expectations of the job, of his co-workers, future prospects, and expects to be evaluated upon this.
Research into the relationship between the quality of exchanges between employees and outcome variables of commitment and the intention to continue working has been done more often (Seers et al., 1995; Witt et al., 1999; Liden et al., 2000). In their research on empowerment, Liden, Wayne and Sparrow (2000) accidentally found a strong direct relationship between the quality of social exchanges and commitment. It is suggested that support and guidance that one receives from coworkers is very important to determine the level of commitment to the organization, but further research is recommended. Other researches into the effect of social exchanges between coworkers on commitment came up with similar results; the social relations with coworkers seem to predict the level of commitment (Witt et al., 1999; Armstrong-Stassen & Schlosser, 2007). Research by Witt et al. (1999) into the effects of CWX, not only found a relationship to commitment, but also a strong negative relationship with the intention to retire. Their conclusions were proof of a statement that Porter and Steers (1973) made in their research into the causes of the intention to retire. The research was concluded by stating: ‘One of the most potent forces in the socialization process within an organization is the interactive dynamics between the individual and his peers. … Because of the potential importance of such a factor, it should prove useful to investigate the relation of peer group interaction to the employee's decision to remain with or leave his employing organization.’
deterioration in the quality of information exchanges between co-workers, i.e. CWX, can diminish older employees’ commitment and intentions to continue work. We suspect that CWX functions as a mediator in this relationship, leading us to the final hypotheses of the research:
Hypothesis 5: CWX mediates the negative relationship between self-stereotyping and affective (H5a) and normative (H5b) commitment.
Hypothesis 6: CWX mediates the negative relationship between self-stereotyping and the intention to remain.
Hypothesis 7: CWX mediates the negative relationship between meta-stereotyping and affective (H7a) and normative (H7b) commitment.
Hypothesis 8: CWX mediates the negative relationship between meta-stereotyping and the intention to remain.
METHOD Respondents and procedure
This study is part of a large survey into the effects of stereotyping and image-forming concerning the elderly in the Netherlands. A questionnaire was sent out among
employees of a large Dutch transportation company. A division of the company with the highest percentage of elderly employees was chosen; it employed 47.5% employees aged 55 and up. This particular division had departments spread out through the Netherlands; in the cities Alkmaar, Apeldoorn, Goes, Gouda and Hilversum. Because of the nature of the work, most of the employees’ working hours were spent away from co-workers; contact occurred at breaks and during time spent at the office.
After several meetings with the HR-manager a division wide letter was sent out, informing employees about the nature of the research, its goal and a link to a website where the respondents could fill out the questionnaire. Respondents were assured that participation was voluntary, and their privacy would be respected. Unfortunately we saw only a small response rate of 14.6%, or 168 out of 1152 employees. This was most likely due to the length of the questionnaire; it took a lot of time to fill it out completely.
the nature of this research we chose to only include the respondents aged 55 and up, leaving a sample of 109 participants. The website did not allow for missing data, so no data had to be substituted by means. The majority of the respondents was male (85%), the average age was 60.1 and they rated their own health a 3.1 on a 5-point Likert scale. A quarter of the respondents finished university, the majority (59.6%) had a secondary education and 13.8% only primary education. All data were obtained from existing surveys which were translated from English to Dutch. Responses were on a 5-point Likert scale (i.e. 1=”strongly disagree” to 5=”strongly agree”), unless otherwise specified. All scale items were averaged; higher scores indicate a greater degree of the underlying construct.
Measures
Independent variables
Self-stereotyping. There were no established measures for a proper assessment of this construct. Prof. dr. E.H. Gordijn, a leading researcher in the field of stereotyping, agreed to develop a scale for the purpose of this study, which contained questions like ‘It’s difficult for elder employees to obtain new skills.’. The construct was measured by 8 questions, a higher score on the construct means a higher amount of negative self-stereotypes are prevalent.
Meta-stereotyping. A new scale was created by Prof. dr. E.H. Gordijn for assessing meta-stereotyping as well. The scale contained questions like ‘I think that the majority of my colleagues believes older employees are not interested in learning new skills’. The construct was measured by 10 questions, a higher score on the construct means a higher amount of negative meta-stereotypes are prevalent.
with factor loadings ranging between .47 and .78 for self-stereotyping, and for meta-stereotyping between .33 and .75.
Mediator
Co-worker exchange (CWX). We used a modified Dutch translation of Seers’ (1989) scale on Team Member exchange (TMX). We replaced the word ‘team’ by ‘co-worker’ to gain the Co-worker exchange scale. An example question is ‘My co-workers have faith in my capabilities’. The construct was measured by seven items (α = .93), on a 7-point Likert scale. A higher score represents a higher quality of exchange relationships.
Dependent variables
Commitment. We used the commitment scale established by Meyer and Allen (1991) to assess affective and normative commitment. An example item of the eight-item subscale (α = .79) for affective commitment is ‘This company personally means a lot to me’. An example item of the eight-item subscale (α = .73) for normative commitment is ‘I have always learned that you should be loyal to one company’.
Intention to continue working. The three-item scale on the intention to continue working of Armstrong-Stassen et al. (2007) was used to assess participants’ intentions to remain with the organization (α = .82). It measured the construct with questions like ‘I expect to work as long as possible for this company’.
Covariates
Demographic covariates. To discover any variances explained by other factors, we included age, sex and health as covariates. Age is measured as an actual value in years. Sex is measured by a single question, where 1 = male and 2 = female. Health is measured as perceived health, on a scale from 1 to 5, where 1 is ‘bad’ and 5 is ‘excellent’.
Analytical strategy
dependent variable, controlling for the independent variable (Baron & Kenny, 1986; Preacher & Hayes, 2004). But researchers have often criticized this model for its lack of statistical power over some other methods. Baron and Kenny (1986) describe another method used for mediation; the Sobel test, which is claimed to have higher statistical power than regression techniques (Preacher & Hayes, 2004, Hayes, 2009). Due to the different way of calculating the mediating effect with the Sobel test, only two requisitions for mediation remain; (1) there exists an effect to be mediated, and (2) the indirect effect is statistically significant in the direction predicted by the mediation hypothesis. Preacher and Hayes (2004) describe Sobel’s method as superior, but also give out a warning that it presumes a normal distribution of the underlying constructs (Hayes, 2009). Because of the size of our population (n = 108) we chose to bootstrap our data a 1000 times as described in Hayes (2009) to achieve a normal distribution. Bootstrapping is used to overcome the limitations of statistical methods that presume a normal distribution, such as the Sobel test. It involves repeatedly sampling from the data set with replacement, and estimating the indirect effect of the antecedents on the dependent variable in each data set. This is called the a*b coefficient product, which is the effect of the independent variable on the dependent variable through one (or multiple) mediators. In this process confidence intervals (CI) are built, when the lower and upper CI does not contain zero, mediation is present. Logically, when the CI does contain zero, no mediation is present (Hayes, 2009). For all analyses a probability level of .05, or a CI of 95% was employed to make sure the indirect effect is statistically unequal to zero.
RESULTS
Table 1
Means, Standard Deviations and Correlationsª
Variable Mean s.d 1 2 3 4 5 6 7 8 9 1. Age 60.12 3.34 2. Sex 1.15 .36 -.334** 3. Health 3.1 .68 .019 .015 4. Self-Stereotyping 2.53 .66 .004 .089 -.165 (.81) 5. Meta-Stereotyping 2.33 .50 -.050 .184 .068 .398** (.73) 6. Co-Worker exchange 4.69 1.06 -.036 -.048 .073 -.098 -.385** (.93) 7. Affective Commitment 3.08 .61 .026 -.095 -.046 -.051 -.303** .479** (.79) 8. Normative Commitment 2.8 .59 .068 .002 -.144 .065 -.295** .297** .496** (.73)
9. Intention to Continue Working 3.28 .92 -.031 .091 .009 -.020 -.195* .451** .500** .300** (.82)
ª n=109. Coefficient alpha reliabilities are in the parentheses along the diagonal. **. Correlation is significant at the 0.01 level (2-tailed).
HYPOTHESIS TESTING
In Table 2 the results of the regression analyses are shown for the three dependent variables; Affective Commitment, Normative Commitment and Intention to Continue Working. The number in the parentheses behind the standardized coefficient is the corresponding t-value, and the significance is indicated by the asterisks. The first column shows the regression analyses containing only the covariates (Age, Sex and Health), the second column adds the independent variables (Self- and Meta-stereotyping) to the equation, and the last column includes the mediator (CWX).
Table 2.
Results of Regression Analyses Testing Hypotheses
Dependent Variable Model: Affective Commitment
Predictor Standardized Coefficients ª
Age -.01 (-.06) -.01 (-.05) .02 (.22) Sex -.10 (-.94) -.04 (-.44) -.05 (-.49) Health -.04 (-.46) -.01 (-.10) -.06 (-.70) Self-stereotyping .08 (.78) .04 (.42) Meta-stereotyping -.33 (-3.13)** -.14 (-1.34) CWX .43 (4.62)**
Dependent Variable Model: Normative Commitment
Predictor Standardized Coefficients ª
Age .08 (.79) .08 (.80) .09 (.95) Sex .03 (.31) .08 (.82) .08 (.84) Health -.15 (-1.51) -.09 (-.95) -.12 (-1.25) Self-stereotyping .19 (1.88) .17 (1.71) Meta-stereotyping -.38 (-3.69)** -.28 (-2.58)* CWX .22 (2.25)*
ª. Corresponding t-values are in the parentheses behind the coefficients **. Correlation is significant at the 0.01 level (2-tailed).
Table 2 (continued).
Dependent Variable Model: Intention to Continue Working
Predictor Standardized Coefficients ª
Age -.00 (-.01) -.00 (-.01) .03 (.27) Sex .09 (.88) .13 (1.26) .13 (1.37) Health .01 (.08) .04 (.37) -.02 (-.17) Self-stereotyping .08 (.70) .03 (.34) Meta-stereotyping -.25 (-2.36)* -.06 (-.57) CWX .44 (4.58)**
ª. Corresponding t-values are in the parentheses behind the coefficients **. Correlation is significant at the 0.01 level (2-tailed).
*. Correlation is significant at the 0.05 level (2-tailed).
The regression analysis confirms the results we have seen at the correlations; self-stereotyping does not relate to any of the outcome variables. No support is found for hypothesis 1 and 3, i.e. no relation with affective (β = .08) and normative commitment (β = .19), and with intention to continue working (β = .08) is found, significance was below .05 for all three variables. This raises suspicion that hypothesis 5 and 6 lack support as well, since Baron and Kenny (1986) clearly state a direct effect needs to be present for mediation to occur.
Table 3 shows the additional results for the Sobel’s analysis. The table is split up for the two independent variables of the model, self-stereotyping and meta-stereotyping. The indirect effect is similar to the ab-coefficient; the coefficient of the effect of the
independent variable on the mediator and the effect of the mediator on the dependent variable. The standard error (SE) is calculated by bootstrapping the results, and the confidence intervals indicate the 95% range of the indirect effect on a normal distribution after bootstrapping. Sobel advocates that zero should not be in the confidence interval for mediation to occur (Baron & Kenny, 1986; Preacher & Hayes, 2004; Hayes, 2009).
Table 3.
Results of Sobel Mediation Analyses
Independent Variable Model: Self-stereotyping* Confidence Interval**
Dependent Variable Indirect Effect SE LL UL
Affective Commitment -.04 .06 -.16 .06
Normative Commitment -.02 .03 -.12 .06
Intention to Continue Working -.05 .07 -.22 .09
Independent Variable Model: Meta-stereotyping* Confidence Interval**
Dependent Variable Indirect Effect SE LL UL
Affective Commitment -.26 .09 -.50 -.12
Normative Commitment -.12 .07 -.31 -.02
Intention to Continue Working -.35 .13 -.70 -.16
* Number of Bootstrap Resamples: 1000
** Level of Confidence for Confidence Intervals: 95
GENERAL DISCUSSION
How does stereotyping relate to employees’ commitment and decisions to continue working? In this research we conclude that meta-stereotyping has a strong relationship with these constructs mediated by the quality of information exchanges at work.
Employees who think their co-workers are positive (or, to be precise less negative) about their age-group, will have a higher quality of information exchange with co-workers, and by that process experience a higher commitment towards the organization, and a higher intention to continue working. We demonstrated this mediating relationship by use of two different statistical methods, both confirmed the same outcome. Self-stereotyping, to the contrary, did not relate to any of the aforementioned variables, it did, however relate strongly to meta-stereotyping. These findings offer meaningful theoretical and practical contributions to dealing with the problem of an aging workforce.
Theoretical Contributions
The effect negative meta-stereotyping can have on an elderly employee is evident in this research. Both affective commitment, the intrinsic desire to work, and normative
commitment, the desire to work because you feel your surroundings expect you to, were predicted by meta-stereotyping. Furthermore, meta-stereotyping related strongly with employees’ intentions to continue working at the organization. This sheds some light on the importance of the perceived circumstances an elderly employee considers himself to be in; where he perceives to be negatively stereotyped, he tends to withdraw from the organization. But of course the primary contribution lies in opening and clarifying the black-box of the relationship between meta-stereotyping and the previously mentioned dependent variables. This research found the quality of information exchange between co-workers, as measured by CWX to be a mediating variable in the aforementioned relationships. Because an employee perceives his co-workers to view him negatively, the quality of information exchanges between him and his co-workers decreases which causes his commitment and intention to continue working to lower.
meta-stereotyping by Vorauer et al. (1998) only a relatively little amount of research has been conducted into the construct. A handful studies have investigated the effects of meta-stereotyping (e.g. Fiske, 2005) but only a few relate it to organization relevant outcomes like commitment and intentions to continue. A study by Desmette and Gaillard (2008) investigated a comparable construct to stereotyping, namely perceived social identity. Perceived social identity is how you see yourself, e.g. as an elderly, which closely resembles self-stereotyping. Curiously, they did find perceived social identity to be a predictor of commitment and intentions to continue working, where our research did not find the same for self-stereotyping. We did, however, find these results for
meta-stereotyping. Other studies (e.g. Armstrong & Stassen, 2007; Gaillard & Desmette, 2008) did similar research into (concepts similar to) stereotyping and all found it to be a
predictor of commitment and intentions to continue working. To the author’s knowledge, this research is the first to explain the effects of negative stereotyping by means of a mediating relationship for the quality of exchanges. Research into stereotyping is moving a step forward by identifying the social exchange relationship at work as an important factor at work, mediating the effects negative stereotypes have on commitment and the intention to continue working.
Limitations and Future Directions
self-stereotypes. Secondly, the validity of the construct self-stereotyping could be too low because, as Levy (1996) continues to explain, the presence of self-stereotypes is very dependent upon culture. Not just the culture of a country matters, that of a company as well, even the culture of a department of a company. It’s very possible that other negative self-stereotypes played a big role at the company, but that we asked for the wrong ones. Unfortunately it went beyond the scope of this study to investigate the presence of (self-) stereotypes at this particular company, and thus we had to make due with a general questionnaire.
Another somewhat surprising result we found was the relational differences between affective and normative commitment. The relationship between meta-stereotyping and affective commitment was stronger than for normative commitment, and results show that normative commitment was only partially mediated by the quality of information exchanges, whereas affective commitment was fully mediated. We can seek an
explanation for this in the theory of normative commitment. As explained by Allen and Meyer (1990) normative commitment stems from what you expect from yourself and what you think others expect of you. While this closely resembles meta-stereotyping, what has to be remembered is that we measured work-related meta-stereotypes, whereas, as Meyer and Allen (1991) explain, normative commitment stems from societal pressure such as colleagues, but also friends, family, the neighbourhood, etc. It goes beyond mere work-relationships, and includes the larger social circle of employees. So part of the effect can be explained by meta-stereotypes, yet another part has to be sought in the larger social circle of the employees. Lastly, we could argue that part of the effect is explained by self-stereotyping, since the expectancies of yourself are an important aspect of normative commitment. Yet, as we explained, it is very possible that the construct self-stereotyping contained some bias. Future research could possibly explore the relationship between stereotypes and normative commitment into more detail.
Finally, future research might be able to clarify the causality of the relations, since the direction of the relationships could not be proven. Especially the relation between
that because of negative self- and meta-stereotyping older employees will be less inclined to invest effort in high quality information exchanges. However, one could also contend that a low quality information exchanges would lead to a higher amount of negative stereotypes, since less information is available, leading to the activation of more
stereotypes (e.g. Maurer et al., 2008). Interestingly, a study by Kite et al. (2005) conclude that when more information on a person is given, less (negative) stereotypes are applied. This means that when less information is available, a relatively higher amount of
stereotypes will be applied. This would mean that a low quality of exchange relationships is a causing factor of an increase in negative stereotypes. Future research could perhaps explain how the reciprocity of stereotyping and social interactions can be influenced for the better.
Conclusion and Practical Contributions
The purpose of this study was to explain how the relationship between the personal aspect of stereotyping with outcome variables commitment and intention to continue working could be explained. It was hypothesized that the quality of exchanges an elderly employee has with co-workers (CWX) mediates this relationship. The personal basis of stereotyping consists of self-stereotyping and meta-stereotyping, commitment was divided into affective and normative commitment. Supporting the combined findings of the literature research (e.g. Vorauer et al., 1998; Desmette & Gaillard, 2008) and the CAPS-theory (Misschel & Shoda, 1995), this study found meta-stereotyping to be related to the outcome variables affective and normative commitment and the intention to
continue working. However, no relation between self-stereotyping and the outcome variables was found, furthermore no relationship between self-stereotyping and the proposed mediator was found. The results did show strong support for a mediating role for CWX in the relationship between meta-stereotyping and the outcome variables. These findings demonstrate the importance of a culture of acceptance and understanding,
REFERENCES
Armstrong-Stassen M., Schlosser F. (2007). Benefits of a supportive development climate for older workers. Journal of Managerial Psychology. Vol. 23, No. 4, pp419-437. Allen N.J., Meyer J.P. (1990). The measurement and antecedents of affective,
continuance and normative commitment to the organization. Journal of Occupational Psychology. No. 63, pp1-18
Baron R.M., Kenny D.A. (1986). The Moderator-Mediator Variable Distinction in Social Psychological Research: Conceptual, Strategic and Statistical Considerations. Journal of Personality and Social Psychology. Vol. 51, No. 6, pp1173-1182.
Beehr T.A. (1986). The process of retirement: A review and recommendations for future investigation. Personnel Psychology, Issue 39, pp31-55.
Buelens M., Kreitner R., Kinicki A. (2002) Organizational Behaviour. New York, McGraw-Hill Eduction, 2nd edition.
CBS, Centraal Bureau voor de Statistiek (2009). Bevolkingstrends. Statistisch kwartaalblad over de demografie van Nederland. CBS, Vol. 57, Issue 4. Chiu W.C.K., Chan A.W., Snape E., Redman T. (2001). Age Stereotypes and
Discriminatory Attitudes towards Older Workers: An East-West Comparison. Human Relations. Vol. 51, Issue 5, pp629-661.
Crocker J., Major B. (1989). Social Stigma and Self-Esteem: The Protective Properties of Stigma. Psychological Review. Vol. 96, No. 4, pp608-630
Cuddy A. J. C., Fiske S.T. (2002). Doddering but dear: Process, content, and function in stereotyping of older persons. In T. D. Nelson (Ed.) (2002), Ageism: Stereotyping and prejudice against older persons (pp. 3–26). Cambridge, MA: The MIT Press.
Desmette D., Gaillard M. (2008). When a “worker” becomes an “older worker”: The effects of age-related social identity on attitudes towards retirement and work. Career Development International. Vol. 13, No. 2, pp168-185.
EurActiv (2009). Sweden admits Lisbon Agenda 'failure'. URL:
http://www.euractiv.com/priorities/sweden-admits-lisbon-agenda-fail-news-221962 Gaillard M., Desmette D. (2008). Intergroup predictors of older workers' attitudes
towards work and early exit. European Journal of Work and Organizational Psychology. Vol. 17, No. 4, pp450-481.
Goldberg, L. R. (1993). The structure of phenotypic personality traits. American Psychologist, Vol. 48, Issue 1, pp26–34.
Hayes A.F. (2009). Beyond Baron and Kenny: Statistical Mediation in the New Millenium. Communication Monographs. Vol. 76, Issue 4, pp408-420.
Hess T.M., Auman C., Colcombe S.J., Rahhal R.A. (2003). The Impact of Stereotype Threat on Age Differences. Journal of Gerontology. Vol 58B, No. 1, pp3-11. Kite, M.E., Stockdale, G.D., Whitley B.E., Johnson B.T. (2005). Attitudes toward
younger and older adults: An updated meta-analytic review. Journal of Social Issues, Vol. 61, Issue 2, pp241-266.
Koudenberg N., Gordijn E.H. (2011). “My Date Can Call Me Sweet, but My Colleague Can’t” Meta-Stereotypic Behavioral Intentions as a Function of Context and Liking of the Outgroup. Social Cognition, Vol. 29, No. 2, pp221-230.
Lammers J., Gordijn E.H., Otten S. (2008). Looking through the eyes of the powerful. Journal of Experimental Social Psychology. Vol 44, pp1229-1238.
Levy B. (1996). Improving Memory in Old Age Through Implicit Self-Stereotyping. Journal of Personality and Social Psychology. Vol. 71, No. 6, pp1092-1107.
Liden, R.C., Wayne, S.J., Sparrowe, R.T. (2000). An examination of the mediating role of psychological empowerment on the relations betweeen the job, interpersonal
Lun J., Sinclair S., Cogburn, C. (2009). Cultural Stereotypes and the Self: A Closer Examination of Implicit Self-Stereotyping. Basic and Applied Social Psychology. Vol. 31, No. 2, pp117-127.
Major D.A., Kozlowski S.W., Chao G.T. (1995). A Longitudinal Investigation of
Newcomer Expectations, Early Socialization Outcomes, and the Moderating Effects of Role Development Factors. Journal of Applied Psychology. Vol. 80, No. 3, pp418-431. Meyer, J.P., Allen, N.J. (1991). A three component conceptualization of organizational
commitment. Human Resource Management Review, 1, 61-89.
Meyer J.P., Allen N.J., Smith C.A. (1993). Commitment to Organizations and
Occupations: Extension and test of a Three-Component Conceptualization. Journal of Applied Psychology. Vol. 78, No. 4, pp538-551.
Michaels, C. E., & Spector, P. E. (1982). Causes of employee turnover: A test of the Mobley, Griffeth, Hand, and Meglino model. Journal of Applied Psychology, Issue 67, pp53-59.
Mischel W., Shoda Y. (1995). A Cognitive-Affective System Theory of Personality: Reconceptualizing Situations, Dispositions, Dynamics, and Invariance in Personality Structure. Psychological Review. Vol. 102, No. 2, pp246-268.
Nelson, T.D. (2002). Ageism: stereotyping and prejudice against older persons. Cambridge, MA: The MIT Press.
Pirandello, L. (1936). One, No One and One Hundred Thousand. New York, Marsilio Publishers.
Porter L.W., Steers R.M. (1973). Organizational, Work, and Personal Factors in Employee Turnover and Absenteeism. Psychological Bulletin. Vol. 80. No. 2. pp151-176.
O’Reilly III C., Chatman J. (1986). Organizational Commitment and Psychological Attachment: The Effects of Compliance, Identification, and Internalization on Prosocial Behavior. Journal of Applied Psychology. Vol. 71, No. 3, pp492-499. Schlenker B.R., Weigold M.F. (1992). Interpersonal Process Involving Impression
Regulation and Management. Annual Review of Psychology. Vol. 43. Issue 1, pp33-68.
Scholl J.M., Sabat S.R. (2008). Stereotypes, stereotype threat and ageing: implications for the understanding and treatment of people with Alzheimer’s disease. Aging and society. Vol 28, pp103-130.
Seers, A. (1989). Team-member exchange quality: A new construct for role-making research. Organizational Behavior and Human Decision Processes, Issue 43, pp118-135.
Seers A., Petty M.M., Cashman J.F. (1995). Team-Member Exchange Under Team and Traditional Management. Group and Organization Management. Vol 20. No. 1, pp18-38.
Sherony K.M., Green S.G. (2002). Coworker Exchange: Relationships Between Coworkers, Leader–Member Exchange, and Work Attitudes. Journal of Applied Psychology. Vol. 87, No. 3, pp542-548
Simon B., Hamilton D.L. (1994). Self-Stereotyping and Social Context: The Effects of Relative In-Group Size and In-Group Status. Journal of Personality and Social Psychology. Vol. 66, No. 4, pp699-711.
Sinclair S., Lowery, B.S., Hardin, C.D. (2006). Self-Stereotyping in the Context of
Multiple Social Identities. Journal of Personality and Social Psychology. Vol. 90, Issue 4; pp529-542.
Sinclair S., Pappas J., Lun J.(2009). The Interpersonal Basis of Stereotype-Relevant Self-Views. Journal of Personality. Vol. 77. No. 5. pp1343-1364.
Turner J. C., Hogg M. A., Oakes P. J., Reicher S., Wetherell M. S. (1987). Rediscovering the social group: A self-categorization theory. Oxford: Basil Blackwell.
United Nations (2002). World Population Ageing: 1950–2050, New York: United Nations.
Vorauer J.D., Main K.J., O’Connell G.B. (1998). How do individuals expect to be viewed by members of lower status groups? Content and implications of meta-stereotypes. Journal of Personality and Social Psychology. Vol. 74, Issue 4. pp917-937.
Vorauer J.D., Hunter A.J., Main K.J., Roy S.A. (2000). Meta-Stereotype Activation: Evidence From Indirect Measures for Specific Evaluative Concerns Experienced by Members of Dominant Groups in Intergroup Interaction. Journal of Personality and Social Psychology. Vol. 78. No. 4. pp670-707.
