1
Does individuating information increase selection and development opportunities for older workers? The influence of age on employee selection and employee training, with
ageism and individuating information as moderating factors Athena Huppes
Master’s Thesis
Work and Organizational Psychology University of Amsterdam, The Netherlands Date: 16-02-2015
Student number: 6022804
Supervised by mw. Dr. Mr. Angelique Bakker- Pieper Second supervisor mw. Prof. Dr. Annelies van Vianen
2 Contents
Abstract ... 3
Introduction ... 4
Literature review... 5
Age and ageism defined ... 5
Three common stereotypes toward older people ... 7
Mechanisms that can explain the negative impact of ageism ... 8
The relation between age and selection ... 9
The relation between age and training ... 10
Individual differences in ageist beliefs ... 11
Impression forming and individuation ... 12
Methods ... 14 Participants ... 14 Ethical Considerations ... 15 Materials ... 15 A. Selection assessment. ... 15 B. Training assessment. ... 15 C. Manipulation check. ... 16
D. Ageist beliefs, personal need for structure and need for cognition. ... 16
E. Demographics and control variables. ... 17
Design ... 17
Procedure ... 18
Results ... 21
Discussion ... 29
Main findings and implications for the literature ... 29
Limitations and future research ... 31
Practical implications and conclusion ... 32
References ... 35
Appendices ... 43
Appendix A: Selection ... 43
Appendix B: Training ... 47
Appendix C: Pilot studie 1 ... 51
3 Abstract
Purpose
The purpose of this study is to investigate the role of age and age biases in the selection and training opportunities of older people and if individuating information can work to improve these opportunities.
Methodology
A survey is conducted among 168 managers of an international bank in the Netherlands. A 2x2 experimental between-subjects design was used in which age and individuation were
manipulated. T-tests, ANOVA’s and Andrew Hayes’ PROCESS tool for moderation were used to test the main hypotheses.
Findings
In line with expectations, this study shows that older people have a lower chance to be selected for jobs than younger people. Older people also have a lower chance than younger people to be selected for training, but this effect only occurs when no individuating information is provided. When individuating information is provided, younger and older people have the same chances to be selected for training. Furthermore, results showed that the more ageist beliefs people have the stronger the negative relationship between age and employee selection becomes.
Conclusion
Corroborating previous research, this study presents evidence that older workers are recruited and developed less than younger workers. This study adds to the existing literature by presenting new evidence that individuating information can contribute to reducing ageist beliefs. Results of this study provide relevant insights for combatting the negative outcomes of ageism and
improving the sustainable employability of older workers. Future research can determine other forms of individuation to improve the selection and development opportunities for older workers
Keywords: Ageist beliefs, ageism, employee selection, employee training, individuation,
4 Introduction
As a result of the inevitable event of getting older all of us could potentially encounter age prejudices, stereotyping and/or discrimination (ageism). In fact, it is our future selves that could be the ones denied jobs, withheld from promotions and being left with the notion that we are no longer worthy of development within our job and not valuable enough to be recruited. In
addition to the effects on our own future careers, there are also important broader consequences for western societies and organizations resulting from ageism.
Today, developed industrial nations face the fact that the workforce is aging and
shrinking at the same time (Kunze, Boehm & Bruch, 2011). This is the result of lower birth rates and increased longevity (ABS, 2008). Research by CBS (Centraal Planbureau voor de Statistiek) in the Netherlands showed that in the years from 2010 to 2040 the number of 65+ year olds will almost double from 2.4 million in 2010 to 4.6 million people by 2040. The number of individuals in the age group 0-19 years old will decrease simultaneously from 23 % of the population in 2013 to 21 % of the population in 2025 (Van Duin & Garssen, 2010). In these times of demographic change it is crucial for the success of organizations to prevent knowledge loss and to retain talent, of all ages (Rabl & del Carmen Triana, 2013). The increased longevity of the workforce has implications for sustaining the provision of societal support (e.g. health provisions and pensions). Older workers are therefore needed to stay in the workforce longer than the current retiring age (Billet, Dymock, Johnson & Martin, 2011). While there have been suggestions that older workers staying on the job may be clogging the labor market and indirectly ‘stealing’ jobs from younger workers, these have now been discredited. Considering that the number of available jobs is not fixed, the early retirements of older workers do not per se provide new jobs for younger workers (Munnell & Wu, 2012). For example, waves of women entering the job market in the last decades has not influenced the labor force participation of men. Similarly the amount of older workers does not seem to influence the (un)employment of younger workers (Gruber, Milligan & Wise, 2009).
However, although older workers need to stay in organizations longer, many studies have shown that they are increasingly absent from the workplace. Much of the literature reports that employer practices favor the recruitment and retention of younger workers over older workers and are reluctant to train those already in employment (Taylor & Walker, 1998; McVittie,
McKinlay & Widdicombe, 2003; Barnett, Spoehr & Parnis, 2008). A generic term for prejudices, stereotyping and discrimination toward older people is ageism (Posthuma & Campion, 2009). Ageist beliefs can have the effect of discouraging older people from remaining in the workforce
5
and can limit the professional development and employability of older workers (Rosen & Jerdee, 1976; Posthuma & Campion, 2009; Fritzsche & Marcus, 2013). Furthermore, acting upon these beliefs could result in a loss of skills, experience and corporate memory, a narrowed recruitment pool, poor returns on investment in human capital and an uneven balance between younger and older employees (Taylor & Walker, 1995). However, Fiske & Neuberg (1990) suggest that due to the provision of individuating information, people may rely less on these ageist beliefs.
Individuating information entails information that refers to the unique personality, behaviors, etc. of a particular individual that is stereotype incongruent (Jussim, 2009). It has been suggested that due to individuation people may rely less on category-based processing and stereotypes (Fiske & Neuberg, 1990).
The aim of this research is to investigate the role age and ageist beliefs play in selection and training opportunities for older people and to investigate whether individuating information may help to improve these opportunities.
Literature review Age and ageism defined
When can an employee be described as an older worker? The answer to that question depends heavily on whom you ask; a 20 year old may find an individual over 35 old, while a senior of 58 may only find individuals over 65 old. The meaning of old in organizations depends partly on the demographic profiles of an occupational group, the relative age. A model may be old at 35 while a professor at University may be seen as old when arriving at the conventional retirement age, depending on the degree to which the individual is older than the work group (Shore, Cleveland, & Goldberg, 2003). A dictionary definition of age would describe age as an effect of time. However, many studies agree that time only does not fully grasp the changes people experience when aging (Arking, 1998; De Lange, Taris, Jansen, Smulders, Houtman & Kompier, 2006). Aging can be seen as a multi-dimensional process that refers to changes in biological, psychological and social functioning across time that can affect an individual on a personal, organizational and societal level (De Lange et al., 2006; Kooij, De Lange, Jansen & Dikkers, 2008).
Most organizational studies have used chronological age (or calendar age) to assess age-related influences on work behavior (Ng & Feldman, 2008; Kooij, De Lange, Jansen, Kanfer & Dikkers, 2011). Furthermore, the distinction between older and younger workers in research also rests most frequently on a definition based on chronological age (Sterns & Miklos, 1995).
6
Although it would be interesting to include other conceptualizations of age in this research, in view of the scope of this study, the structure of the research methods used and the effectiveness of the use of chronological age, this research focuses on the relation between chronological age and work-related outcomes.
The majority of researchers define a person that is 55 or older as an older employee, because these employees are approaching the conventional retirement age (65 years old). In this study the therefore old is defined as an individual that is 55 or older (Finkelstein, Burke & Raju, 1995; Munnell, Sass & Soto, 2006; James, McKechnie, Swanberg & Besen, 2013). Younger workers were defined as workers of 35 or younger, because this is roughly the median of the age distribution from when employees start working around the age of 20 and when they are old employees of 55 and above (Van dalen, Henkesn & Schippers, 2010).
During the course of human aging, mental and physical decline can occur, but when and how that happens varies considerably from individual to individual (Ghosheh, 2008). Ageism can lead to age becoming representative for the capabilities of an individual, instead of assessing their value as an individual who happens to be a certain age (Macnicol, 2006). Early research defined ageism as containing three components: prejudice expressed by erroneous beliefs, stereotypes, and discriminatory behavior directed at older people (Butler, 1980; Rupp, Crede & Vodanovich, 2006, pp. 1339). Extensive research on the definition of the concept of ageism was done by Iversen, Larsen and Solem (2009), because earlier definitions of ageism used in research had hitherto been diffuse and ambiguous. They established the following definition, which will be used in this research due to its clarity and it’s comprehensiveness ‘Ageism is defined as negative or
positive stereotypes, prejudice and/or discrimination against (or to the advantage of ) elderly people on the basis of their chronological age or on the basis of a perception of them as being ‘old’ or ‘elderly’. Ageism can be implicit or explicit and can be expressed on a micro-, meso- or macro-level’ (Iversen, Larsen & Solem, 2009, pp. 15).
Several aspects of this definition warrant explanation. First, it includes three social psychological components; stereotypes (cognitive), prejudices (affective) and discrimination (behavioral). Additionally, ageism does not have to be negative; positive ageist beliefs also exist, although these are less common than negative ageist beliefs. In this research, only negative ageist beliefs will be discussed because these are more common and pose a greater problem in organizations than positive ageist beliefs do (Iversen, Larsen & Solem, 2009). Furthermore, implicit and explicit ageist beliefs exist, meaning that ageist beliefs can be conscious or unconscious. Finally, ageism can exist on an individual (micro), social networks (meso) and institutional and cultural level (macro).
7
There are overlaps but also subtle differences between the terms age stereotypes, age prejudices, and age discrimination that can lead to a confusing usage of the terms. Even more confusing is that each of these terms (age stereotyping, age prejudices and age discrimination) can separately or together be described as ageist beliefs. A brief introduction to these three terms further clarifies the concept of ageism. The concept of ‘stereotype’ was introduced by Lippmann (1922), who defined it as the typical pictures that come to mind when thinking about a particular social group (Dovidio, Hewstone, Glick & Esses, 2010). Stereotypes of older workers are specific stereotypes that can be defined as beliefs and expectations about workers based on their age (Hamilton & Sherman, 1994). When stereotypes about older people translate into negative feelings against older persons, one can speak of prejudice, which has an affective component (Nelson, 2011). Finally, age discrimination occurs when one age group receives a different treatment from another age group on the grounds of chronological age.
Three common stereotypes toward older people
A meta-analysis by Posthuma and Campion (2009) identified the most common age stereotypes. Older workers were thought of as poorer performers, more resistant to change, to have a lower ability to learn, to have a shorter remaining tenure, to be more costly and more dependent. Posthuma and Campion (2009) have refuted some of these stereotypes in their meta-analysis. The most relevant stereotypes in the context of this research are the poorer performance
stereotype, the lower ability to learn stereotype and the resistance to change stereotype.
First, the poorer performance stereotype entails that older workers are perceived to perform poorly compared to younger workers, which has been proven not to be true in practice. Older and younger workers perform differently in different aspects of performance according to a meta-analysis by Ng and Feldman (2008). Older workers for instance show good citizen behavior, can control their emotions better at work and are less likely to engage in
counterproductive behaviors. There seems to be no relationship between age and core task performance (Ng & Feldman, 2008). When multiple aspects of job performance are taken in to account, older workers are just as productive as younger workers (Prenda & Stahl, 2001).
Second, the lower ability to learn stereotype entails that older workers are often seen as having less potential for development than younger workers (Martin, Dymock, Billet & Johnson, 2013). Indeed, older workers may experience a slight decline in some areas of mental abilities. However, they seem to have compensating mechanisms such as work experience, life experience and
8
(organizational) knowledge, which ultimately equals the mental abilities of younger and older individuals. Overall, older workers seem to learn as well and as quickly as younger workers (Smith, Smith & Smith, 2007). In addition to the stereotype that older people are less suitable for development, people seem to think older people show a lower return on investments in terms of development, because they are approaching the retirement age and therefore there is less time for the organization to reap the benefits of the investment (Hedge, Borman & Lammlein, 2006). However, older workers do not seem to be less profitable to invest in, because younger workers make more career changes than older workers
Finally, the resistance to change stereotype is the stereotype that older people are more resistant to change, not flexible and more set in their ways. However, not much research has been done on the validity of this stereotype. It is therefore not clear if older people are indeed more
resistant to change and what the possible consequences of this resistance would be (Posthuma & Campion, 2009).
Mechanisms that can explain the negative impact of ageism
A mechanism that can explain how ageism toward older people can occur, is the social process of in-group bias. The in-group bias is a phenomenon where members of an in-group receive preferential treatment in comparison to members of the out-group (Lewis & Sherman, 2003). In the case of ageist beliefs, the in-group bias and social identity theory predict biases in favor of members in someone’s own age cohort (Tajfel & Turner 1979; Chiu, Chan, Snape & Redman, 2001), leading to personnel decisions that are biased by higher ratings for employees that fall in one’s own age-cohort. Remarkably however, younger workers evaluated their in-group more positively on overall job qualifications and potential for development, but older workers did not show differences in how they evaluated younger and older workers (Finkelstein, Burke & Raju, 1995). An explanation for this is that older workers experienced the different phases of working life at all age levels and thus understand the phase younger people are in. Younger people have not yet experienced the issues and concerns older workers are facing and may therefor empathize less with them and rely more on stereotypes (Finkelstein, Burke & Raju, 1995).
Another mechanism that can explain the negative impact of ageist beliefs is social role theory. Social role theory proposes that people show behavior appropriate for socially defined categories (for example teacher, father or student) (Eagly, 1987). One possible social role is the employee role. Because younger people in the last decades mostly filled the role of employee, an
9
association has developed between employment and younger workers rather than employment and older workers. This may be a reason why the role of employee is perceived as less suitable for older people and older people are perceived as less likely to be competent employees than younger people (Kite, 1996). Also in observing others in their social role, people have the
tendency to underestimate the role of situational factors in behavior and overestimate the role of personality. This effect is known as the fundamental attribution bias (Gilbert & Malone, 1995). For example imagine observing an old professor who gives a long lecture and forgets to describe a particular issue. One may think of this old professor as having a forgetful mind as expected from the social role of old professor. Giving a lecture however, entails memorizing a vast amount of information and can lead to forgetfulness in any age-group. The situation could have been underestimated and the personality attributed to the role of an older professor may have been overestimated.
The relation between age and selection
The nature of labor has changed over the last decades, due to changes such as those in computer technology and organizational restructurings. This is also reflected in the phenomenon that individuals are expected to have a diverse career with multiple jobs and positions. Indeed, according to Cascio (2003) new workers entering the labor market will hold about seven to ten jobs in their lifetime. But the job search after a career change or after discharge does not seem to be that easy for older workers. Research showed that ageism in the form of discrimination is a serious problem in the labor market. A study found that a 31-year-old applicant had received 223 % more callbacks in the first stage of the hiring process than a 46-year-old applicant (Ahmed, Andersson & Hammarstedt, 2012). Older workers were rated least favorable in the selection process when no information about job experience was given, after a career change, and in situations where age-based stereotypes were made salient.
Hiring discrimination against older workers is disadvantageous for various reasons. It is counterproductive because the best person for the job may not get the job due to the age of that person. This could result in suboptimal financial consequences and a failure to capitalize on the human capital potential of organizations (Gringhart, Helmes & Speelman, 2005). In addition, it provides younger people with a pessimistic view of the future and leaves older job seekers with an increasing frustrated, depressed and dispirited attitude towards finding a job (Aquino, Russell, Cutrona, & Altmaier, 1996; Rife, 1992).
10
Just like some jobs are sex-typed (for example the expectation that most nurses are women and most doctors are men), jobs can be age-typed. One explanation for the occurrence of age stereotyping in hiring decision making is the age-type of the job combined with the cognitive matching theory. People engage in a cognitive matching process when perceived features of the candidate for a job are matched with the age-type of the job (Perry, 1994, 1997; Perry & Finkelstein, 1999). When the age-type of the job was young (e.g. energetic, efficient and able to handle multiple tasks), younger workers were perceived as more suitable for the job than older workers (Perry & Bourhis, 1998). The more salient the applicant’s age is, the more likely it is that age discrimination exists, because age saliency can lead to the activation of age-associated job stereotypes (Cleveland & Landy, 1983, 1987). Consequently, applicants are not only chosen based on merits but also on the age-type of the job combined with cognitive matching.
To conclude, both in practice as in research a strong preference for younger applicants exists in job selection (Fritzsche & Marcus, 2013). Similar results were anticipated in this research leading to the following hypothesis
Hypothesis 1: Older workers are less likely to be selected for a job than younger workers
The relation between age and training
Many developed industrial nations are currently facing or will in the future face an aging workforce. One strategy to encourage older workers to continue to work, to maintain and expand their employability, increase promotional opportunities, improve job security and earn higher wages is to provide them with training and development opportunities (Cully, Van den Heuvel, Wooden & Curtain, 2000; Kapsalis, 1998; Žnidaršič, 2012). Employee training is an important investment for companies because it is not only linked to organizational performance through increasing skill, motivation and knowledge, but also leads to more satisfied and
committed employees (Bulut & Culha, 2010; Latif, Jan & Shaheen, 2013). However, research from studies across Europe and North America shows that employers are far more likely to fund training of young employees than older employees (Brunello 2001; Brunello & Medio 2001; Giraud, 2002). An explanation for this low interest in training older employees is the influence of ageist beliefs. Several studies have shown that due to ageist beliefs people generally view older employees as having less potential for development than younger workers. Many managers regard older workers as difficult to train, unable to adapt to new technologies and too cautious (Brooke & Taylor, 2005). Furthermore, research found that older adults who are exposed to
11
ageist attitudes, performed with lower scores on among other memory related tasks. Thus an ageist attitude toward older workers may lead to lower performance of that worker resulting in a self-fulfilling prophecy where the initial ageist beliefs are confirmed (Levy, 2001). However, another explanation for the lack of development opportunities for older workers exists; older workers also seem to be less motivated and willing to participate in training and development activities (van Vianen, Dalhoeven & de Pater, 2011). The question is, why would willingness and enthusiasm to develop oneself decline over time?
Support of supervisors is important in the learning attitudes and developmental activities of their employees. Research has shown that a lack of support from managers negatively
influences the willingness of older people to participate in trainings. In addition, older workers in general tend to be insecure about their own learning abilities and consider themselves as less able and capable to learn new things, which may reduce their training and development willingness (Noonan, 2007; Van Vianen et al 2011). This unwillingness is unfortunate, because studies show that older workers do not need special treatment in terms of training, but they do need to be encouraged in undertaking training (Smith, Smith & Smith, 2007).
In general the literature shows a reluctance to provide training to older employees (Martin et al. 2013). Congruent with the literature, it is predicted in this study that older workers receive less training than younger workers.
Hypothesis 2: Older workers are less likely to be selected for training than younger workers.
Individual differences in ageist beliefs
The extent to which people have ageist beliefs differs between people. The personality traits personal need for structure and need for cognition can influence the extent to which people have ageist beliefs and use these beliefs in forming their impression about others (Fiske, Lin & Neuberg, 1999; Swenicionis & Fiske, 2013). People with a high need for structure, may be more prone to form impressions based on categorizations, whereas people with a need for cognition may prefer to process more individuating information, take their time and like to explore and consider different options when making a decision even if it is not necessary (Whitley & Kite, 2010). Furthermore, it was found that people with high scores on Openness, Agreeableness and
Conscientiousness had less ageist attitudes, whereas a high score on Neuroticism indicated more ageist
beliefs in an indirect manner namely through aging anxiety (Allan, Johnson & Emerson, 2014). Not much research has been done in the area of individual differences in category-based or
12
attribute-based processing, and it is a possible area for future research (Swenicionis & Fiske, 2013).
In the broad literature ageism is considered to have a profound influence on managerial decision making and can result in discrimination against older employees (Rosen & Jerdee, 1976; Finkelstein, Burke & Raju, 1995; Martin et al, 2013). The extent to which ageist beliefs influence the relationship between age and employment decisions, depends on individual differences in ageist beliefs. It was predicted that the more ageist beliefs managers have, the stronger the relationship between age and employment decision becomes.
Hypothesis 3: Ageist beliefs negatively moderate the effect of age on the decisions managers make with regard to selecting and training older workers.
Impression forming and individuation
Jussim (2006) distinguished between the definition of individuation and the definition of individuating information. Individuation is defined as judging a person as a unique individual rather than as a member of a group. Individuating information is defined as information that accentuates the unique personality, behaviors, attitudes, accomplishments, etc. of a particular individual that is inconsistent with the category they were stereotypically placed in (Jussim, 2006). Due to individuation people may rely less on category-based processing and stereotypes (Fiske & Neuberg, 1990).
However, taking into account the overwhelming amount of information people have to process every day, human functioning requires that people categorize, to simplify the world around them to be able to make sense of it. Thus, to be successful in a complex social
environment with limited cognitive resources, cognitive short cuts are taken. In the dual process theory this kind of thinking is called system 1 thinking or category-based processing. This refers to an intuitive system, which is fast, automatic, effortless, implicit and emotional. System 1 thinking is used for top-down processing which entails that first a person is placed in a certain category then stereotypical information can be inferred from this category. Yet, people are also strategic in their use of cognitive capacity and when people are motivated to do so, they can process information more effortful and carefully, described as system 2 thinking. This refers to slower, conscious, effortful, explicit and logical thinking (Stanovich, 1999; Kahneman, 2003). System 2 thinking, or attribute-based processing, is used for bottom-up processing which
13
involves people paying attention to the details and looking carefully at personal attributes before forming an impression (individuating) (Fiske & Neuberg, 1990).
If category-based processing is the default, when do people use the more effortful attribute-based processing? According to the literature this is dependent on the accessibility of stereotype incongruent information and the motivation and cognitive ability to process this information. When people have access to individuating information, are motivated to process the information, and have the cognitive ‘space’ to process this information, then modification of stereotypes because of individuating information can occur (Nelson, 2002). The motivation to effortful process information about another individual is dependent on whether the perceiver finds this individual relevant (Fiske & Neuberg, 1990).
It is conceivable that certain individuating information may ensure that older individuals are not subjected to unfair discrimination based on membership of the category ‘older people’ (Finkelstein, Burke & Raju, 1995). Not much research has been done on whether individuation can be used to deactivate ageist beliefs and how it can improve the chances of older workers in recruitment and development processes (Bal, Reiss, Rudolph & Baltes, 2011).
Hypothesis 4: Ageist beliefs negatively moderate the effect of age on the decisions managers make with regard to selecting and training older workers, but only when no individuating information is provided.
14
Figure 1: Research model
Methods Participants
This research was conducted among managers in lower and middle management positions in a large bank in The Netherlands. It was decided to conduct this research among managers because they are responsible for decisions concerning the selection and training of employees. An invitation for this questionnaire was sent out to 1000 managers. In total 215 managers participated in this research of whom 168 managers completed the survey. In this research Qualtrics (the questionnaire system, Qualtrics Labs Inc, Provo, UT) was programmed to randomly assign the managers over the four conditions. Ultimately there were 42 participants in the first condition, 44 participants in the second condition, 43 participants in the third condition and 39 participants in the fourth condition. 111 managers (66.1%) were men and 57 (33.9%) were women. Ages of the participants ranged from 29 to 62 and the mean age was 43.76 (SD = 7.43). The work experience participants had in the position of manager ranged from 1 to 35 years and the average was 11.54 years (SD = 7.42). The majority of the sample has a higher education degree. 71 (42.3%) participants have University degree, 84 (50%) have an HBO degree (higher vocational education) and 13 participants (7.2%) have a MBO (intermediate vocational
15
education) or high school degree. A one-way between-groups analysis of variance showed that there were no significant differences for participant age (F = .57, p = .64) or the average years of work experience in the four conditions (F = .93, p = .43). A Chi-square test for independence indicated no significant difference for gender in the four conditions (χ² = .65, p = .89). Finally, a Chi-square test showed the amount of participants with an HBO or University degree was not significantly different per condition (χ² = 4.74, p = .19). Other educational degrees were not included in this analysis because of the low number of participants in these groups.
Ethical Considerations
The Faculty Ethics Review Board (ERB) of the University of Amsterdam evaluated this research as adhering to the ethical standards that are formalized in the Ethics Regulations. Approval was given for the execution of this research.
Materials
The online survey included measures on selection assessment, training assessment, ageist beliefs, personal need for structure and need for cognition. Furthermore, the survey involved manipulation checks and demographic questions. All the survey questions were translated to Dutch with the use of the back-translation technique. All survey questions can be found in Appendix D.
To test the reliability of the scales used in this study internal consistency was tested using the Cronbach’s alpha coefficient. The Cronbach’s alpha is considered as sufficient above the value of .7 and ‘good’ when the coefficient is .8 or higher (DeVellis, 2003).
A. Selection assessment.To measure the probability and suitability for selection, a scale was used created by Cable and Judge (1997). One item was deleted because it did not add to the purpose of this study. The items were measured on a 5-point Likert scale ranging from strongly disagree (1) to strongly agree (5). A sample item is ‘It is probable that I would recommend to
hire this candidate?’ For this 3-item scale the Cronbach’s alpha coefficients was good α = .87. The
last item was measured in the form of an open comment. This item read: ‘Taken everything into
consideration regarding the applicants resume, what is your overall evaluation of the candidate?’
B. Training assessment.To measure the benefits from training and the suitability for training the 4-item ‘training benefits’ scale from the one used by Averhart (2012) was used. One item was removed because it was not relevant for the goal of this research and one item was added which was used to measure the training probability also developed by Averhart (2012).
16
The items were measured on a 5-point Likert scale ranging from strongly disagree (1) to strongly agree (5). A Sample item was ‘I believe that this candidate will benefit from attending this training course’ Cronbach’s alpha for this 4- item scale was good α = .82. In addition to that an open comment was added which read ‘Taking everything in consideration, what is your overall evaluation of the candidate?’
C. Manipulation check.To check the manipulations 4 items were included in the questionnaire (see Appendix D). To hide the actual manipulations, 6 items were added about for instance the nationality, gender and education level of the candidate. To measure if the
manipulation of age was successful the following items were included: ‘What was the age of the
candidate in the selection situation?’ and ‘What was the age of the candidate in the training situation? Response
options were between the ages of 25-45 and between the ages of 45-65. To investigate whether the manipulation of individuation had worked the following manipulation check items were included: ‘The reference letter gave me specific information that was useful for the evaluation of the candidate’ and ‘The introduction to the candidate gave me specific information that was useful for the evaluation of the
candidate’. The individuation manipulation items were measured on a 5-point Likert scale.
D. Ageist beliefs, personal need for structure and need for cognition.A survey by Hassell and Perrewe (1995) was used to measure the extent to which participants had ageist beliefs. The survey included 36 items, however 9 items were removed because they measured personal experiences with ageism, which was not relevant for this research and more importantly could reveal the purpose of this study. The items were measured on a 5-point Likert scale that ranged from strongly disagree (1) to strongly agree (5). A sample item is ‘Older employees are harder
to train for jobs.’ Cronbach’s alpha for the adjusted scale was α = .78, which indicates a sufficient
internal consistency.
Items from scales measuring personal need for structure and need for cognition were included and combined with the items on ageist beliefs to control for these constructs and furthermore to serve as a way to prevent hypothesis guessing, as the ageist beliefs survey may reveal the true subject of this research. To measure personal need for structure the scale by Wolfradt Sommer and Rademacher (1999) was used. Each statement was measured on a 5-point scale. Answers varied from strongly disagree (1) to strongly agree (5). In current study
Cronbach’s alpha was sufficient, .73. To control for need for cognition, a survey by Cacioppo, Petty and Kao (1984) was used. The scale was “designed to assess people’s tendencies to engage in or to
enjoy effortful cognitive endeavors” (Cacioppo, Petty, Feinstein & Jarvis, 1996, p. 199; Nicoli, 2011).
17
uncharacteristic (1) to extremely characteristic (5). In current study Cronbach’s alpha was low but still acceptable, α = .67.
E. Demographics and control variables.The demographical variables included in this research were: Gender, age, education level, position within the bank and years of experience as a manager inside as well as outside the bank.
Design
In this research a 2x2 experimental between-subjects design was used. The age of the candidate was manipulated to study whether age impacted the likelihood to be selected or trained. This was manipulated by creating resumes and performance evaluations in which only the birth dates differed. Furthermore, individuation was manipulated by providing either control information or specific individuating information (see Appendix A and B).
The independent variable employee age was conceptualized as chronological age and was classified as 35 years old or younger for the younger worker and 55 years old or older for the older worker. The two dependent variables were employee selection and employee training and there were two moderator variables, ageist beliefs and individuating information. A between subject-design was chosen which lead to 4 conditions (see table 1). Each participant was asked to evaluate one candidate for selection and one candidate for training, these two scenarios were presented as concerning two different persons. However in each condition both candidates were either young or old and both candidates were either presented with control information or with individuating information. A between subject-design was chosen to disguise the goal of this research and to enable participants to make evaluations on the candidate without being influenced by the age of the other candidate which could serve as a reference point. The 2-2 experimental between-subject design is shown schematically in table 1.
Table 1
Four condition of the 2x2 Experimental Between-Subject Study
Control information Individuating information Young applicant Condition 1 Condition 2
18 Procedure
Managers in a large bank in the Netherlands were approached to participate in this research. They were sent an Internet link where they could log on into the survey in Qualtrics (Qualtrics Labs Inc, Provo, UT). Subsequently they indicated their informed consent and were randomly assigned to 1 of the 4 conditions. Several demographic questions were asked to determine the participant’s gender, job level, education level and age. To hide the actual purpose of the research it was announced that this study was done to research the influence of joint decision making on the effectiveness of internal selection processes. This approach of concealing the true subject of the research was aimed at decreasing the social desirability bias.
Participants then read two instructions both concerning either a young or an old candidate. In the first instruction, participants were asked to advise a colleague manager on the hiring of a specific candidate for the job of mortgage advisor (see pilot study 1) based on a job description, a CV and a reference letter (Appendix A). To make the scenario as realistic as possible, a job opening and job description from the bank were chosen and the resume was based on resumes sent by candidates that applied for this position. The resume was also inspected by an employee that was responsible for the development of selection criteria of mortgage advisors. The younger and older worker had an equal amount of work experience in their current field and applied for the same job. In the second instruction participants were asked if they would advise the involved manager to select a specific candidate for training on the basis of a performance evaluation and an introduction to the candidate provided by the manager (Appendix B). The evaluation form was created based on evaluation forms of mortgage advisors from the bank and examined by a manager in the field of mortgage advisors and an employee involved with performance management.
The reference or support letters were either standard (control group) or individuating for both advices (Appendix A & B). In the selection vignette a reference letter was created by assessing which qualities are necessary and recommended for the job of mortgage adviser. In the training vignette an introduction to the candidate was created by using research from Averhart (2012) that assessed which characteristics are considered positive to have for successful participation in a training course. This research lead to three areas of performance; leadership, respect for others and client focus, which were used to create the introduction to the candidate. Stereotype incongruent information was added by contradicting the three common stereotypes described in the literature (the poorer performance stereotype, the lower ability to learn
19
Manipulation checks were incorporated to check if participants took into account age and the specific individuating information when deciding on advising the manager on selecting and training the fictional candidate. Then participants were asked to complete surveys containing questions about ageist beliefs, personal need for control and need for cognition (Appendix D). Finally, they were debriefed on the true purpose of this research and were thanked for their participation.
Pilot study 1
This pilot study was conducted to make sure that the job used in the manipulation of this research was not age-typed. In order not to test the job matching people engage in when
deciding who is right for the job, but to test what role age plays in the selection decision, an age neutral job had to be chosen (Perry, 1994, 1997; Perry & Finkelstein, 1999). The procedure of this pilot study was inspired by a pilot study by Acker (2008).
Participants. To ensure that participants had relevant experience with recruitment and could therefore provide valuable opinions on the age-type of the job, 20 recruiters of the HR Recruitment department of the bank were invited to participate in this pilot study of whom 10 completed the survey. Of the participants 6 were women and 4 were men. Their ages ranged from 30 to 49 and the mean age was 37.80 (SD = 7.08). All recruiters had relevant work experience in their field ranging from 7 to 28 years and an average of 14.50 years (SD = 7.01).
Procedure. The true topic of this pilot study was concealed to make sure the topic of the main study would also stay concealed. This was done by stating that this pilot was an investigation of ways to improve the effectiveness of the selection and training decision making process. To test how inter-manager advice could improve these processes, 5 positions were evaluated. In addition to questions about age, also questions about gender were added to hide the actual purpose of the pilot study.
The participants received a list of 5 jobs they had to evaluate. These 5 jobs were selected from the open vacancies database of the bank and included a job description. Participants were asked for each job to rate how suitable persons in different age ranges would be for it on a 5-point Likert scale ranging from 1 (very unsuitable) to 5 (very suitable). The age groups were categorized in 25-35 years old, 35-45 years old, 45-55 years old and 55-65 years old. Thus for every job low scores indicated a low suitability and high scores indicated a high suitability level. Finally they were asked in an open question which jobs they thought in general were most suited for different ages. When different age ranges were considered suited for a job, the job was
20
considered age neutral. The following 5 jobs were included in the pilot study. (1) Commercial product manager (2) Private banker (3) Mortgage consultant (4) Black belt (consultant
operational excellence) (5) Program manager finance (see Appendix C for descriptions). Results.This pilot study created scores on the age-type of 5 jobs. The job with the highest mean score would be considered as the most suitable for all 4 age cohorts. The mean scores and the standard deviations are reported in table 2. Conclusion of this pilot study was that the job of mortgage adviser was the least age-typed job, meaning that the job is seen as suitable for younger people as well as older people.
Table 2
Mean scores and standard deviation on the 5 jobs, (N=10)
M SD
Commercial Product manager 3.88 .61
Private banker 3.95 .59
Mortgage adviser 4.35 .19
Black Belt 4.10 .47
Program finance manager 4.08 .64
Pilot study 2
This pilot study was conducted to investigate if the CV and the performance evaluation of the candidate were realistic, whether the control information and the individuating
information had the intended effect and whether the subject of this research was hidden as intended. In addition to that, questions were asked on the sense of realism in the experiment compared to how things work at the bank and what the impact was of being an advisor instead of a decision maker. Comments and feedback were taken in to account in finalizing the survey
Participants. Three participants were recruited for this pilot study. They were all male and their ages were 34, 47 and 49.
Procedure.Three participants received the survey and were asked to answer the full questionnaire for each of the 4 conditions. While completing the survey, feedback could be given to the research leader about the survey. Each participant took about an hour to complete the survey. Afterwards, questions were asked about different aspects of the survey such as ‘Which
21
influence did the control reference letter have on your opinion of the candidate?’ and ‘What influence did the fact that you are an advisor instead of a decider have on the answers you gave?’.
Results.After studying the answers resulting from the interview and after reviewing the general feedback, changes were made in the survey where necessary. The most important changes were the addition of open comments to the questions in the selection and training vignette where participants could explain their evaluation and adjustments in the training vignette in terms of specifying the purpose of the training. After this last check the survey was ready to be sent out.
Results
The raw data were edited by removing unfinished surveys. These were removed because the participants were told that they could stop the survey at any moment in time and that in that case the data would not be used.
The means, standard deviations, reliabilities and correlations of different variables are included in table 3. Individuation correlated positively with employee selection (r = .11, p = .15) and employee training (r = .16, p < .05) , although only the relation with employee training was significant. Furthermore, ageism correlated positively with need for structure (r = .24, p < .01) and was moderately negatively related to the age of the participant (r = -.39, p < .01) and the work experience participants had in the role of manager (r = -.34, p < .01).
For the age manipulation check, t-tests were performed. These tests indicated a
significant association between the age (Young: M = 1.02, SD = .15 and old M = 1.65, SD = .48) of the fictional candidate and the age manipulation check for the selection scenario t (166) = -11.43, p<.01. Also for the training scenario a significant relation was found between age (Young:
M = 1.25, SD = .43 and old M = 1.73, SD = .45) and the manipulation check t (166) = -7.04, p
<. 01. To check the manipulation of individuation, t-tests were conducted. In the selection scenario there was a significant difference for evaluations on the control reference letter (M = 3.39, SD = 1 .03) and the individuated reference letter in how individuating they were perceived
M = 3.78, SD = .81; t (166) = -2.76, p < .01. In the training scenario there was also a significant
difference for the evaluations on the control introduction to the candidate (M = 3.28, SD = 1.00) and the individuating introduction to the candidate in how individuating they were perceived M = 3.96, SD = .83; t (166) = -4.81, p < .01. In conclusion all three manipulations had the intended effect on the participants.
22 Table 3
Means, standard deviations, correlations and alpha reliabilities on the diagonal (N=168, except for age and work experience mngr., N=167)
Young Old
Control
Information Information Individ. Information Control Information Individ. Total
M SD M SD M SD M SD M SD 1 2 3 4 5 6 7 8 1 Employee selection 3.23 .88 3.63 .72 3.07 .85 3.02 .88 3.24 .86 .87 2 Employee training 3.36 .58 3.40 .77 3.02 .65 3.44 .84 3.30 .73 .10 .82 3 Ageist beliefs 2.64 .37 2.78 .30 2.71 .30 2.73 .27 2.71 .31 .04 .13 .78 4 Individ. - - - .11 .16* .12 - 5 Need for structure 2.33 .63 2.63 .55 2.51 .59 2.55 .57 2.50 .57 .07 .02 .24** .15* .73 6 Need for cognition 3.76 .32 3.66 .32 3.65 .31 3.71 30 3.70 .31 .00 -.03 -.07 -.03 -.35** .67 7 Participant Age 43.12 7.44 44.77 7.29 44.14 7.61 42.92 7.49 43.76 .57 .04 .03 -.39** .02 -.15 .06 - 8 Work experience mngr. 12.12 7.62 11.52 7.64 12.48 7.96 9.92 6.27 11.54 7.42 -.08 .00 -.34** -.10 -.20* .12 .75** - Note. ** p < .01, * p < .05.
23
To test if managers are less likely to hire older workers than younger workers (hypothesis 1) an independent sample t-test was conducted. Younger candidates were significantly more positively evaluated (M = 3.43, SD = .82) than older candidates on employee selection, M = 3.0,
SD = .86; t (166) = 3.00, p < .01. The magnitude of the differences in the means (mean
difference = .39, 95% Cl: .13 to .65) was small but close to moderate (eta squared = .05). This confirms hypothesis 1.
To examine if managers are less likely to train older workers than younger workers (hypothesis 2) an independent sample t-test was performed. There was no significant difference in the evaluations between younger (M = 3.38, SD = .68) and older people for employee training, M = 3.22, SD = .77; t (166) = 1.44, p = .15. The magnitude of the differences in the means (mean difference = .16, 95% Cl: -.06 to .38) was small (eta squared = .01). This result did not confirm hypothesis 2.
Ageism was examined as a moderator in the relation between age and employee selection and employee training (hypothesis 3). Given the finding that older age was significantly
correlated with fewer ageist beliefs (r = -.39, p < .01), the following analyses were controlled for participant age. The interaction effect of ageism and age was analyzed using the PROCESS macro for SPSS (Hayes, 2013) and was tested for employee selection ( b = -.64, 95% CI [-1.42, .14], t = -1.62, p = .11) as well as for employee training (b = -.10, 95% CI [-.91, .71], t = -.25, p = .80) both interaction effects were not significant therefore hypothesis 3 was not confirmed. For a more detailed inspection of the direction of the combined effect of age and ageist beliefs, simple slope analyses were performed using the PROCESS macro (Hayes, 2013). This analysis estimates different conditional effects of the predictor on the outcome variables at low (one SD below the mean), mean (sample mean) and high (one SD above the mean) values of the moderator. Results showed that at low levels of ageist beliefs, there was a non-significant negative relationship between age and employee selection, b = -.19, 95% CI [-.55, .16], t = -1.06, p = .29. At mean levels of ageist beliefs, there was a significant negative relationship between age and employee selection, b = -.39, 95% CI [-.65, -.13], t = -2.98, p < 01. At high levels of ageist beliefs, there was also a significant negative relationship between age and employee selection, b = -.59, 95% CI [-.94, -.23], t = -3.27, p < .01. The simple slope analyses showed that there was a significantly stronger negative relation between age and employee selection at mean or higher levels of ageist beliefs. The simple slopes equations of the regression of age on employee selection at different levels of ageist beliefs, can be seen in figure 2. Simple slope analysis with ageism as a moderator in the relation between age and employee training did not show any significant simple slope.
24
Figure 2. Simple slopes equations of the regression of age on employee selection at three levels of
ageist beliefs.
Finally, individuation was analyzed as a moderator of the effect of ageism on the relationship between age and employee selection and employee training (hypothesis 4). This hypothesis was tested by using Model 3 of the PROCESS macro for SPSS (Hayes, 2013; see figure 1). The analyses were controlled for participant age. See table 4 for results of the moderated moderation analysis. First the interaction of age of the candidate, ageism and
individuation was tested for employee selection. The model summary was significant (R² = .10, F (8, 158) = 3.56, p < .01) but the three-way interaction term was not a significant contributor to the explained variance (b = -.45, 95% CI [-2.04, 1.15], t = -1.98, p = .58). However, a significant interaction effect was found between individuation and ageist beliefs (b = -.81, p < .05). A simple slope analysis was performed using the PROCESS macro (Hayes, 2013) to better understand the direction of the moderation effects. Results showed that without individuation the simple slopes were not significant but had the expected direction (see Figure 3). With
individuating information, 2 simple slopes were significant however not in the expected direction (see figure 4, low levels of ageism: b = -.34, 95% CI [-.87, .19], t = -1.24, p = .22; mean levels of ageism: b = -.58, 95% CI .94, -.21], t = -3.11, p < .01; high levels of ageism: b = -.82, 95% CI [-1.29, -.34], t = -3.40, p < .01).
25
Figure 3. Simple slopes equations of the regression of age on employee selection at three levels of
ageist beliefs without individuating information.
Figure 4. Simple slopes equations of the regression of age on employee selection at three levels of
ageist beliefs with individuating information.
Then the interaction of age of the candidate, ageism and individuation was tested for employee training (see table 4). The model summary was close to significant (R² = .08, F (8, 158)
= 1.96, p = .056) but the three-way interaction term was not a significant contributor to the
explained variance (b = -.19, 95% CI [-1.91, 1.54], t = -.21, p = .83). Interestingly there was a significant interaction effect of age of the candidate and individuation. Candidate age and individuation had a significant positive effect on employee training (b = .48, p < .05). A simple slope analysis was performed using the PROCESS macro (Hayes, 2013) to better understand the direction of the moderation effects. Results showed that without individuation the simple slopes were close to significant or significant (see figure 5, low levels of ageism: b = -.31, 95% CI [-.76, .04], t = -1.79, p = .07; mean levels of ageism: b = -.37, 95% CI [-.65, -.10], t = -2.67, p < .01; high levels of ageism: b = -.39, 95% CI [-.82, .05], t = -1.76, p = .08). With individuation none of
26
the simple slopes were significant (see figure 6). This result reveals that without individuating information ageism negatively effects the employee training evaluations of older workers, but with individuating information this effect does not occur. In conclusion, both three-way interaction terms did not significantly contribute to the explained variance and therefore hypothesis 4 was not confirmed.
Figure 5. Simple slopes equations of the regression of age on employee training at three levels of
ageist beliefs without individuating information.
Figure 6. Simple slopes equations of the regression of age on employee training at three levels of
Running head: THE INFLUENCE OF AGE ON EMPLOYEE SELECTION AND EMPLOYEE TRANING
27 Table 4
Linear model of predictors of employee selection and employee training, (N = 167). Covariate: age of the participant. Confidence intervals are shown between brackets.
b SE B t p Employee Selection Constant 3.02 [2.18, 3.87] .43 84.14 p < .01 Age candidate -.39 [-.65, -.13] .13 -2.94 p < .01 Ageist beliefs -.00 [-.47, .46] .24 -.001 p = .99 Individuation .15 [-.11, .41] .13 1.16 p = .25
Age candidate x ageist beliefs -.54
[-1.34, .26] .40 -1.33 p = .18
Age candidate x individuation -.37 [-.90, .16]
.27 -1.38 p = .17
Ageist beliefs x individuation -.81
[-1.61, -.00] .41 -1.98 p < .05 Age candidate x ageist beliefs x
individuation -.45 [-2.04, 1.15] .81 -1.98 p = .58 Employee Training Constant 2.88 [2.14, 3.61] .37 7.70 p < .01 Age candidate -.14 [-.37, .09] .12 -1.20 p = .22 Ageist beliefs .40 [-.04, .84] .22 1.81 p = .07 Individuation .19 [-.04, .41] .12 1.62 p = .11
Age candidate x ageist beliefs -.13
[-.98, .72] .43 -.31 p = .76
Age candidate x individuation .48
[.02, .94] .23 2.08 p < .05 Ageist beliefs x individuation .20
[-.65, 1.05] .42 .46 p =.65
Age candidate x ageist beliefs x
individuation [-1.91, 1.54] -.19 .87 -.21 p = .83
28
Two one-way between-groups analyses of variance were conducted to explore if there was a significant difference between the four conditions and the likelihood for employee selection and employee training. The results of the test showed that there was a significant difference in the evaluations that were given in the four conditions for employee selection (F (3, 164) = 4.72, p < .01) with a medium effect (eta squared = .08). For employee training the test showed that there was a significant difference in the evaluation that were given in the four conditions (F (3, 164) = 3.14, p < .05) with a small but close to moderate effect (eta squared = .05). Post-hoc analysis using the Tukey HSD were conducted to determine which groups differed from each other. Results of the ANOVAs can be found in table 5.
Table 5
Means and standard deviations of age and individuation, F’s from ANOVA and R²
Without individuation With individuation
Young Old Young Old
M SD M SD M SD M SD F(3,164) R² η² Employee selection 3.23a .88 3.07a .85 3.63b .72 3.02a .88 4.72** 9.84** .08 Employee training 3.36a .58 3.02b .65 3.40a .77 3.44a .84 3.14* 4.83* .05
Note. Young without individuation N=42, Old without individuation N=43, Young with individuation N=44, Old with individuation N=39; For each row: same subscripts = no significant difference between means; ** p < .01, * p < .05.
The table shows that older candidates are selected less than younger candidates but only in a situation with individuating information. In contrast, older candidates are selected less for training than younger candidates but only in a situation without individuating information. The differences are depicted in figures 7 and figure 8.
29
Figure 7. Mean evaluation scores for employee selection for younger and older people, with and
without individuation.
Figure 8. Mean evaluation scores for employee training for younger and older people, with and
without individuation.
Discussion Main findings and implications for the literature
The goal of present research was twofold: first it was aimed at investigating the influence of ageist beliefs on the relation between age and both selection and training opportunities. Second it was aimed at examining if individuating information can deactivate ageist beliefs and through that improve selection and development opportunities for older workers. It was confirmed in this research that the extent to which people have ageist beliefs is also dependent on the age of that person, the results showed that older people have significantly
30
fewer ageist beliefs than younger people. For that reason analyses were controlled for participant age.
The results suggest that older people have a smaller chance to be selected for a job than younger workers with the exact same qualifications for the job. This is in congruence with earlier research by Bal, Reiss, Rudolph and Baltes (2011) that showed in a meta-analysis that older people have a structural lower chance to be selected than younger workers. The strong
preference for younger applicants in job selection is often caused by ageist beliefs (Fritzsche & Marcus, 2013). In this study, ageism did not have the expected effect as it did not significantly moderate the relationship between age and employee selection or age and employee training. However ageist beliefs did influence chances for older and younger people to be selected for a job at mean and high levels of ageist beliefs. When ageist beliefs were divided into low, mean and high levels of ageism results showed that the more ageist beliefs people have the stronger the negative relationship between age and employee selection. A reason for not finding the expected effect of ageism may be because biases are hard to measure through a survey, due to response biases. As ageism is a sensitive subject and no one wants to be seen as an ‘ageist’ person, there is a risk for the social desirability bias. The subject of the survey was hidden as much as possible by mixing the items about ageist beliefs with other scale items. But to measure ageist beliefs quite direct questions on prejudices, stereotypes and discriminations toward the elderly needed to be asked. Considering these possible biases the results of current research are interesting and the expectation can be raised that the actual level of ageist beliefs may be higher.
Surprisingly, individuating information increased the chances for younger people to be selected for a job but not the chances for older people. An explanation of this finding may be, that the older candidate is not part of the organization yet resulting in a lowermotivation to process the individuating information concerning the older candidate. According to earlier research individuation only works when people have the motivation to actually process the information and this is dependent on whether the perceiver finds the other personally and motivationally relevant (Fiske & Neuberg, 1990; Nelson, 2002). Since no investments have been made in the candidate yet, the motivation to process the individuating information on the older worker may be too low.
Another interesting finding was that older people have a smaller chance to be selected for training than younger workers, but only when no individuating information was given. When personal and stereotype incongruent information was given, the chances for older and younger workers were about equal to be selected for training. Also individuating information only
31
significantly increases the chances for older workers to be selected for training, for younger workers the chances for training were equally high with or without individuating information. It was in line with expectations that older employees are less likely to be selected for training than younger employees. Several studies have shown that employers are far more likely to fund training of the young and well educated employees than older employees (Brunello 2001; Brunello & Medio 2001; Giraud, 2002). It was also expected that individuation improves the chances for older employees to be selected for training. Because of the individuating information people rely less on category-based processing and stereotypes and therefore evaluate the
candidate more on individual value (Fiske & Neuberg, 1990). Lastly, the individuation did not work in in favor of the development opportunities of younger workers. A reason for this may be that individuating information did not add any value for the younger candidates, as stereotypes about younger persons are different from older persons. The individuating information in this study was stereotype-incongruent for the older worker but not for the younger worker. These results are a significant contribution to the literature. Not much research has been conducted on how to reduce age biases and only a small number of studies researched the effects of
individuating information on reducing age biases (Bal, Reiss, Rudolph and Baltes, 2011). Limitations and future research
The present study has in addition to its strengths also its limitations. Most importantly, this was not a real selection situation for employee selection and employee training. This research design does not test how managers will select when real people are involved and when choices have real consequences. Furthermore, older and younger applicants had the same resumes except for the birth dates. On the one hand one can say that older and younger workers had exactly the same qualifications. On the other hand people in general expect older people to be further in their careers than younger people. When old and young are in the same stage in their career it may be perceived as a better resume for the younger worker than for the older worker, even though resumes were evaluated separately and therefor the older and the younger resume were not compared. In a real situation older and younger workers are attractive for different reasons. They do not have the same qualifications, they have different qualifications in different areas (Ng & Feldman, 2013). Furthermore, a small inconsistency in one of the resumes may have
influenced evaluations of the participants. One date in one of the resumes was not adjusted properly.
32
Future research on this subject may benefit from integrating a longitudinal research design. Outcomes of longitudinal research may reveal additional information about the complex attitudes toward older people and the effects of these perceptions on the workplace in real situations with organizational consequences. Also, the role of individuation could be researched more thoroughly in a longitudinal study. In this study individuation is only measured in the form of individuating information. Longitudinal research could also indicate what other forms of individuation such as intergroup contact through diverse ages in teams could contribute to more fair selection processes (Pettigrew & Tropp, 2006).
Another limitation is that exactly at the moment of distributing the survey, it was announced that due to a reorganization the workforce of the bank would be reduced by 1700 people. This may have influenced the perceptions of managers on employee selection and employee training. Research shows that employers are more open to selection and training of older workers if the organization is expanding. In a situation where an organization is
reorganizing and shrinking, employers are less receptive to hire and train older workers
(Arrowsmith & McGoldrick, 1997). Thus the reorganization may have negatively influenced the results for older workers. In future research more than one organization should be involved in the research to filter out cultural and situational differences. Then results can be generalized more broadly to the population.
Finally, managers’ opinions on which employees to train differed. The open comments in the survey revealed that most managers thought training should be given instead to
underperformers. Interestingly, other managers had different opinions on when to train employees and for whom training is relevant. Some managers thought that training should be given to excellent employees as a form of appreciations, others thought training should be given to good employees who could become excellent. The view managers have on who should receive training influences how the candidate is evaluated. This could explain the divergent answers in the training evaluation. In future research these views should be controlled for. Also it would be interesting in future research to take candidate preferences for training in to account when researching development opportunities of older and younger people.
Practical implications and conclusion
The findings of the present research have important implications for practice. A deeper understanding of ageism and its effects on older workers can improve HR strategies for dealing with an older workforce and can contribute to sustainable employability. Organizations focus
