An employee perspective on the content of the psychological contract : exploring the effects of age, gender, sector and function on the psychological contract

(1)

1

An employee perspective on the content of the

psychological contract

Exploring the effects of age, gender, sector and function on the psychological

contract

Amsterdam, 29-06-2015

Kim van Zon (10443134)

Supervisor: Drs. R. van Hemert

Bachelor’s Thesis

BSc. Business administrations

Faculty of Economics and Business

(2)

2

Statement of originality

This document is written by Kim van Zon who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

Date: 29-06-2015

(3)

3

Abstract

Violations of the employee’s psychological contract can cause worse job performance and motivation. To help employers avoid contract breach, more detailed insights into the content of the psychological contract are necessary. Specifically, it is hypothesized that age, gender, sector and function influence which dimensions of the employee’s psychological contract are most important and most violated. Further, it is expected that, based on the notion of availability heuristics, the most important and the most breached dimensions are similar. The hypotheses are tested using two samples, consisting of 642 (Study 1) and 65 (Study 2) employees in the Netherlands. Based on the data, 14 dimensions of the psychological contract are identified, ranging from wage to physical work environment. Additionally, the results indicate that as employees age they experience more violations of relational dimensions of the psychological contract, but value transactional aspects more. Further, those in higher functions also experience more violations of relational dimensions. No other effects are supported, although an exploratory study found some small effects of age, gender, sector and function on some of the individual dimensions of the psychological contract.

(4)

4

Foreword

I could not have written this thesis without the help of some people, therefore I would like to use this section to show them my gratitude. First and foremost, I want to thank my supervisor, Drs. R. van Hemert, for providing me with useful feedback and many suggestions. This helped me a lot in generating new ideas and significantly improved the content and structure of my thesis. Additionally, I want to thank my family, friends, students and all other respondents who filled in the surveys of this study. Specifically, my special thanks to those who were so kind to spread the survey among their own friends and family to help me obtain the number of responses I needed. Finally, thank you to my friends who helped me to pilot test the survey, as well as those who read the first versions of this thesis, which helped me to improve it.

(7)

7

1. Introduction

The survival of traditional Dutch retailers is threatened by changing economic markets. One cause is the rise of e-commerce which, due to its dynamic nature, provides new opportunities, but also increases uncertainty (Amit and Zot, 2001). An organization that failed to benefit from e-commerce is the traditional Dutch retailer, Vroom and Dreesmann. Even though internet sales in the Netherlands increased from 2.8 billion euros in 2005 to 10.6 billion in 2013, the organization was late in launching their webshop (Rijlaarsdam, 2015) and lacks a unique e-business strategy (Boex and de Groot, 2015). In combination with the growth of strong low-cost competitors, Vroom and Dreesmann’s financial position has weakened (Roggeveen, 2015). Seeking for measures to maintain viable, they proposed a 5.8 percent wage cut (Verhaar, 2015). Even though a part of the workforce is willing to accept this proposition to retain their jobs (AD, 2015), it may have serious consequences on employees’ motivation, job performance and organizational citizenship behaviours (Matthijs-Bal, Chiaburu and Jansen, 2010; Robinson, 1996). This results from the violation of expectations regarding pay levels and other benefits that the company can no longer fulfil. In turn, these consequences may aggravate the problems Vroom and Dreesmann is already in. How can Dutch organizations avoid harming employees’ expectations to minimize worse job performance?

A first step to a solution is generating a proper understanding of the content of the psychological contract. The psychological contract can be defined as a belief in mutual obligations between an employee and employer, based on a perceived promise and a consideration in exchange for that promise (Rousseau and Tijoriwala, 1998). From an employee’s perspective, the psychological contract consists of aspects such as pay level, training and job security (Rousseau, 1990; Herriot, Manning and Kid, 1997; Turnley and Feldman, 1998). However, there are multiple contract dimensions for which the literature is inconsistent, including the importance of job excitement (Turnley and Feldman, 1998). A possible explanation for these disparities is that the content of the psychological contract in different studies is measured using different samples. Since this thesis will focus on the Dutch labour force, additional research within this specific country is necessary. Additionally, Sparrow (1996) suggests that the content of the psychological contract may change over time. Since most research has been conducted in the late nineties, a revisit of the dimensions may be useful.

Furthermore, the content of the psychological contract can be divided into relational and transactional categories. Whereas relational content focuses on socio-emotional and intrinsic aspects, transactional content is oriented towards economic and extrinsic aspects (Rousseau, 1990). This is relevant, since these orientations can predict employee behaviour (Coyle-Shapiro and Kessler, 2000) and provide a simple guideline on how to motivate employees. Herriot et al. (1997) suggest that employees are more likely to value the transactional content of the psychological contract, whereas employers ascribe more importance to relational aspects. However, it may be argued that differences

(8)

8 within these groups exist. When considering the employee group, Ng and Feldman (2009) state that younger workers place more value on transactional aspects, such as pay, whereas older employees are more likely to value relational aspects, such as social relationships. Similarly to older workers, females are also more likely to value relationships and thus the relational psychological contract. However, males place more value on achievement, which suggests a preference for transactional aspects such as promotion opportunities (Konrad, Ritchie, Lieb and Corrigall, 2000). These difference indicate a possible effect of gender on the psychological contract. Furthermore, Willem, de Vos and Buelens (2010) discuss the differences between public and private sector employees. They state that relational characteristics such as job content are valued equally by both sectors, but that civil servants care less about promises of transactional aspects such as financial rewards and career development. In contrast, Bellou (2007) states that civil servants find career developments more important than private sector employees. This suggests that there might be some differences in the content of the psychological contract for different sectors, but it is not clear in what direction. Finally, there are indications that function influences which dimensions of the psychological contract are found to be most important. It is suggested that the higher jobs are in the hierarchy, the more likely that employees value relational aspects of the psychological contract (Millward and Hopkins, 1998), particularly the extent to which the job is interesting (Centers and Burgental, 1966).

As opposed to these statements, Herriot et al. (1997) did not find relationships between age, gender, sector and function and the content of the psychological contract. However, this may be due to an error associated with combining positive and negative instances of the psychological contract (Herriot et al., 1997). Additional research in which this error is minimized and in which the inconsistencies in the literature are investigated is thus necessary. In turn, this leads to the following research question: How do age, gender, sector and function influence the extent to which the content of an employee’s psychological contract is relationally or transactionally oriented?

To provide an answer to this question three additional objectives are formulated. First, the content of the psychological contract will be revisited for both current times and the Dutch work force. Second, this content will be categorized as relational or transactional. Third, the importance of each dimension of the psychological contract will be determined. Therefore it is necessary to verify whether the method used to determine the dimensions is also appropriate for measuring their importance. To achieve these goals an empirical study is conducted, in which a mixed methods design is used. The data is collected through an open-ended questionnaire using the critical incident technique. This survey is held among 642 Dutch employees varying in age, gender, function and sector. Subsequently, the importance of the dimensions is verified using an additional closed-ended questionnaire, held among 65 Dutch employees.

This paper will continue with an overview of the literature on the psychological contract. This is followed by the methodology section in paragraph 3 and the results in paragraph 4. This thesis will end with a discussion and conclusion in paragraph 5.

(9)

9

2. Theoretical framework

The purpose of this section is to review the existing literature to provide a more complete overview of the research problem. First, the content of the psychological contract is discussed, followed by an examination of the literature on the transactional and relational orientation of contracts. Next, the use of the critical incident technique is examined to determine whether this is an appropriate method to measure the importance of the dimensions of the psychological contract. Additionally, the literature on the effects of age, gender, sector and function on the content of the psychological contract are explored. Finally, the section ends with a conclusion that provides an overview of the findings.

2.1. Content of the psychological contract

As described in the introduction, the psychological contract includes the perception of mutual agreements, based on a perceived promise and some activity or obligation in exchange for that promise (Rousseau and Tijoriwala, 1998). Even though promises create expectations, they are two different concepts. Whereas promises are contractual, expectations include a broader amount of drivers that are not necessarily contractual, including causal reasoning (Rousseau and Tijoriwala, 1998). This distinction is important, because when a promise is perceived as not met, this results in a greater negative effect on job satisfaction as well as stronger employee emotions in comparison to a breach of expectations (Robinson, 1996). This is, however, also dependent on trust levels of the employee (Robinson, 1996). Nevertheless, a breach is generally perceived to occur by more than half of the organization’s employees (Robinson and Rousseau, 1994). As a consequence, employees’ motivation, job performance and organizational citizenship behaviours are negatively affected (Matthijs-Bal, Chiaburu and Jansen, 2010; Robinson, 1996). In turn, this may harm organizational performance. A better understanding of the content of the psychological contract is thus necessary to help employers to avoid breaching it.

Rousseau (1990), one of the leading scholars on the psychological contract, identifies seven dimensions of which the contract consists. These include promotion opportunities, high pay, performance-based pay, training, job security, development and support (Rousseau, 1990). Some of these dimensions are consistent across different studies, particularly pay, training and job security (Herriot, Manning and Kid, 1997; Turnley and Feldman, 1998). However, Herriot, et al. (1997) and Turnley and Feldman (1998) also find dimensions that are inconsistent with the findings of Rousseau (1990), including responsibility and power, and job excitement. These diverging results may be explained by differences regarding the circumstances under which the dimensions are measured. For example, whereas Herriot et al.’s (1997) sample consists of UK employees and managers, Rousseau’s (1990) sample contains US students. This indicates that not all dimensions are generalizable across different countries and samples. Furthermore, research also suggests that contracts are not generalizable over time. Sparrow (1996) observed the differences between “old” (60s) and “new”

(10)

10 psychological contracts (90s). Two examples of his findings include a shift in motivational currency from promotions to job enrichment and skills enhancing opportunities, and that personal development moved from the organization’s to the individual’s responsibility (Sparrow, 1996; Hiltrop, 1995). A related trend is that in “new” contracts employees are more likely to value tangible rewards and economic factors in comparison to early psychological contracts (Sparrow and Cooper, 1998). Such developments are triggered by lower levels of job security and employee’s revisited expectations of employer obligations (Hiltrop, 1995; Smithson and Lewis, 2000). The latter is caused by the increasing pressure on businesses to enhance flexibility and efficiency (de Meuse, Bergmann and Lester 2001). These finding are supported by de Meuse et al. (2001), who suggest that psychological contracts consisting of socio-emotional factors are less valued in the year 2000 in comparison to contracts in the 50s and 70s. These generalizability issues indicate that a revisit of the content of the psychological contract in current times, as well as a focus on the Dutch employee sector, is useful.

2.1.1. Transactional versus relational content. In addition to identifying the content of the

psychological contract, the extent to which this content is relational or transactional is examined. Rousseau (1990) makes a distinction between the categories based on five factors, including focus, time frame, stability, scope and tangibility. Based on these characterizations, the transactional contract can be defined as having an economic and extrinsic focus within a specific time frame. The contract is static, has a narrow scope and its dimensions are observable (Rousseau, 1990). Dimensions included here are, for example, pay-related ones (Robinson, Kraatz and Rousseau, 1994). Contradictory, the relational contract has both an economic and social orientation with an intrinsic focus. The time frame is indefinite and the contract itself is dynamic, has a broad scope and its dimensions are more subjective (Rousseau, 1990). Examples of dimensions included here are training and development opportunities (Robinson et al., 1994).

The focal point of this study will be the focus of the dimensions of the psychological contract. First of all, the transactional psychological contract has an organizational-extrinsic orientation, which is consistent with its economic focus. This orientation emphasizes the importance of tangible rewards, which are often provided by the organization as a means of motivating performance. Examples are pay, promotion opportunities and working conditions (Mottaz, 1985). Secondly, the relational psychological contract has an intrinsic and social-extrinsic focus. Intrinsic job aspects are directly related to the task itself, rather than external pressures or drivers (Ryan and Deci, 2000). Examples are responsibility and getting feedback (Mottaz, 1985). Additionally, the social-extrinsic dimension is associated with the value of interacting with others, including supportive co-workers and supervisors (Mottaz, 1985). Even though it is an extrinsic orientation, which is associated with the transactional psychological contract, it befits the social aspects of the relational contract better.

Based on these definitions, there are some differences in the extent to which Rousseau (1990) would classify a dimension as transactional or relational in comparison to this paper. For example, Centers and Burgental (1966) view job security as organizational-extrinsic motivator and, according to

(11)

11 this study, would thus be classified as transactional. On the other hand, Rousseau (1990) suggests job security is relational, because it is related to a long-term career path within the firm. This indicates some complications when using all five characterizations of Rousseau to determine the orientation of the psychological contract. Even though job security has an extrinsic and thus transactional focus, it is still relational, because of the time frame of the dimension. Focusing on only one characteristic will minimize this problem.

Despite the difficulties, distinguishing between the different dimensions is valuable, because it provides a simple guide employers can use to identify which employee groups can be motivated by transactional or relational rewards. In turn, this can help to avoid contract breach. Additionally, it provides a better understanding of the employee-employer relationship, since it can help employers to predict employee behavior. Employees who value transactional content of the psychological contract are less likely to fulfil their obligations towards the organization as opposed to employees who value relational content. This division can thus also be useful in avoiding a psychological contract breach from the employer’s perspective. (Coyle-Shapiro and Kessler, 2000).

2.2. Verification of the critical incident technique

The third goal of this study is to verify the use of the critical incident technique (CIT). The CIT identifies contract dimensions by asking respondents for instances in which they experienced a violation of the psychological contract. Rather than measuring importance, it thus identifies the most breached dimensions. However, it is expected that this provides a relatively accurate appraisal of the most important dimensions, based on the notion of availability heuristics. This means that when people evaluate something, they rely on examples that are easily recalled from memory. According to the heuristic, the fact that something is coming to mind is an indicator that it is more important than the alternatives (Tversky and Kahneman, 1973). Additionally, it is suggested that the most violated dimensions are easily recalled, because the survey responses provided later in this study are very intense. In other words, when presenting respondents with a list of dimensions of the psychological contract to rank, the ones for which they experienced a breach are likely to come easy to mind and are ranked highest. It is thus expected that when comparing the most breached dimensions with the most important dimensions, results are similar.

2.3. Psychological contract and age, gender, sector and function

The third and final goal of this study is to determine the influence of age, gender, sector and function on the content of the psychological contract. The purpose is not only to determine whether these person and job characteristics influence the extent to which the psychological contract is transactional or relational, but also to explore the effects on its individual dimensions. The latter provides a better and more detailed insight into the content of the psychological contract.

(12)

12 Since the literature on the psychological contract does not provide detailed explanations of its relationship with person and job characteristics, the following sections are mainly build on the work values theory. Work values are based on the importance that employees attach to specific aspects of the job, including factors such as job opportunities, interesting work and relationships with co-workers (Kalleberg, 1977). Such type of work values are similar to the content of the psychological contract as described in previous sections. From logical thought, it thus follows that important job values are also the important aspects of the psychological contract.

2.3.1. Psychological contract and age. As people grow older and learn or experience new

things in life, they change. Not only in personality (McCrae, Costa, de Lima, Simões, Ostendorf, Angleitner et al., 1999), but also in life values (Ryff, 1989) and work values (Twenge, Campbell, Hoffman and Lance, 2010). Studies indicate that this process of aging increases the extent to which people associate themselves with intrinsic and social-extrinsic values (Ng and Feldman, 2009). For example, people above and around their 50s state that relationships with others are most important to their well-being in general life (Ryff, 1989). For job attributes, valuing intrinsic factors already starts at the age of 30 (Twenge et al., 2010; Kooij, de Lange, Jansen, Kanfer and Dikkers, 2011). At the same time, studies conclude that young workers (defined as younger than 36) are more likely to rate extrinsic factors higher, including prestige, pay and security (Kooij et al., 2011) as well as leisure time (Twenge et al., 2010).

However, the literature is inconsistent. Some studies do not find any age differences (Cennamo and Gardner, 2008; Lacy, Bokemeier and Shepard, 1983), whereas others state the opposite. Namely that older workers value transactional aspects and younger workers relational aspects. For example, Twenge et al. (2010) state that those born between 1965 and 1981, the middle-age group, are more likely to value transactional aspects in comparison to both the younger and older group. Similarly, Jurgensen (1978) concludes that younger workers take their jobs less serious and attach importance to factors that make work more pleasant. This includes not only the suggested transactional variables, but also relational ones. Last of all, there are indications that aging changes the dimensions that employees value, but not necessarily the transactional or relational focus of the contract. For example, Ryff (1989) suggests that as people grow older, the importance of income decreases, but the significance of work hours increases.

Concluding, the literature is inconsistent and additional research is necessary. So far, most support is provided for the statement that when people age, their psychological contracts shift from transactional to relational. However, even though this is expected, the contradicting literature needs to be taken into account. It is therefore suggested that age groups differ in the extent to which both the most violated [H1a] and the most important [H1b] dimensions of the psychological contract are transactionally or relationally oriented.

2.3.2. Psychological contract and gender. There is some general agreement on the thought

(13)

13 (Feingold, 1994) and job values (Konrad et al., 2000). However, the extent to and the way in which males and females differ is yet unclear due to contradictory results in the literature.

On the one hand, studies indicate that women are more relationally oriented in comparison to men. Whereas personality characteristics of males include assertiveness and self-esteem, females are more trusting, extravert and tender-minded. The latter fit with a social-extrinsic orientation, including building relationships and talking about personal issues (Feingold, 1994). Similarly, feminine stereotypes in job attributes, such as relationship-orientations and intrinsic factors (Konrad et al., 2000; Vallerand and Bissonnette, 1992) as well as meaningfulness of the job, are valued more by women (Lacy et al., 1983). Masculine stereotypes, such as power and promotion opportunities (Konrad et al., 2000) as well as organization-extrinsic work values such as job security (Lacy et al., 1983) are rated higher by men.

On the other hand, some studies indicate that women score higher on transactional aspects, whereas men ascribe importance to relational aspects of the psychological contract. Walker, Tausky and Oliver (1982) state that women appreciate convenience-related job attributes more than men. Even though this includes the social-extrinsic aspect of opportunities to air personal problems, the main part of the category consists of organization-extrinsic dimensions such as physical surroundings, travel time and work hours (Walker et al., 1982). This is also supported by Konrad et al. (2000), who indicate that work environments are part of the feminine stereotypes. Additionally, it is suggested that men rate intrinsic items of the psychological contract higher, because their job is more of a central life interest in comparison to the life values of women (Konrad et al., 2000; Ryff, 1989).

However, independent of the direction, the extent to which gender differences are present at all may be modest. Effect sizes found in the literature are small (Konrad et al., 2000) and job function controls for a part of the gender differences (Walker, et al., 1982). Finally, even though some studies find gender differences on several dimensions, they indicate that, overall, men and women have similar job attribute preferences (Ryff, 1989; Lacy et al., 1983).

In summary, most support is found for the suggestion that women are relationally oriented, whereas men are transactionally oriented. However, the possibility remains that it is the other way around. Therefore, it is expected that men and women differ in the extent to which both the most violated [H2a] and the most important [H2b] dimensions of the psychological contract are transactionally or relationally oriented.

2.3.3. Psychological contract and sector. Sectors can be defined and distinguished on

multiple bases. In this study, it encompasses the distinction between public and private. Whereas the public sector is concerned with governmental services, the private sector includes all corporations not fully controlled by the state.

The literature indicates that public servants are more relationally oriented in their psychological contracts, whereas employees in the private sector are more transactionally oriented.

(14)

14 This statement is derived from the public-service motivation theory, which assumes that public servants have a tendency to respond to motives grounded in public organizations (Moon, 2000). In other words, they are driven by a need to deliver a valuable service to society (Buelens and Van den Broeck, 2007). Consistent with this theory, public servants are found to attach importance to meaningfulness of the work (Houston, 2000) as well as challenging and stimulating tasks (Lyons, Duxbury and Higgins, 2006). At the same time, the opposite is found for private sector employees who value organizational-extrinsic aspects of the psychological contract, such as pay (Karl and Sutton, 1998; Buelens and Van den Broeck, 2007) and work hours (Houston, 2000).

However, some of the literature conflicts with the public-service motivation theory. Studies find that public servants are not motivated by challenging tasks and other intrinsic factors (Buelens and Van den Broeck, 2007), but rather value organizational-extrinsic dimensions such as career developments (Bellou, 2007). Others state that both sectors equally value organizational-extrinsic factors such as pay and benefits (Lyons et al., 2006) or equally value relational aspects such as job content and social atmosphere (Willem et al., 2010). Especially regarding the dimension job security the literature is inconsistent. Several studies indicate that this dimension is equally important to both sectors (Karl and Sutton, 1998; Lyons et al., 2006), whereas others suggest that job security is more important to public sector employees (Houston, 2000), because they are more risk-averse (Bellante and Link, 1981).

In summary, most scholars support the statement that public servants are more likely to value intrinsic, and private sector employees organizational-extrinsic job aspects. However, there may be some transactional dimensions that are valued more by public servants, in particular job security. Taking this into account, it is expected that public and private sector employees differ in the extent to which both the most violated [H3a] and the most important [H3b] dimensions of the psychological contract are transactionally or relationally oriented.

2.3.4. Psychological contract and function. Whether employees have a high- or low-level

job position appears to be one of the most important factors influencing diverging job values. More so than age, gender and sector (Buelens and Van den Broeck, 2007). The literature distinguishes between high- and low-level positions based on either the separation of white- and blue-collar workers (Centers and Burgental, 1966), supervisors and non-supervisors (Jurkiewicz and Massey, 1997) or prestige, income and educational levels (Lacy et al., 1983).

Even though the literature on function and job values is limited, it is consistent. Most authors indicate that the higher employees’ jobs in the hierarchy, the more likely they are to value relational aspects of the psychological contract (Millward and Hopkins, 1998). For example, they find opportunities for self-expression, the extent to which the job is interesting (Centers and Burgental, 1966) and meaningfulness of the job (Lacy et al., 1983) more important than employees lower in the hierarchy. Additionally, workers in higher functions are found to be more committed, which correlates with valuing intrinsic aspects of the job (Moon, 2000). At the same time, studies support that those

(15)

15 lower in the hierarchy are more motivated by extrinsic and transactional job aspects. For example, they value income, promotions and job security (Moon, 2000; Centers and Burgental, 1966) more than those in high functions. Only Jurkiewicz and Massey (1977) state that both groups show similarities in their values.

Concluding, most of the literature suggests that employees in higher functions rate intrinsic and relational aspects of their jobs higher. The literature on this subject is, however, limited, so it is important to keep the option that subordinates value relational aspects of their jobs open. It is therefore expected that that employees in high and low functions differ in the extent to which both the most violated [H4a] and the most important [H4b] dimensions of the psychological contract are transactionally or relationally oriented.

2.4. Conclusion

Previous research on the content of the psychological contract, as well as motivational and job attributes theory, show that age, gender, sector and function may be related to the psychological contract. Determining in what way these characteristics influence the extent to which the content of the contract is relational or transactional, can provide an alternative and easy guide to help employers in avoiding to breach the contract. However, detailed research that relates these dimensions specifically to the psychological contract, is limited or inconsistent. Therefore, this study attempts to provide additional insights on these effects.

(16)

16

3. Methodology

The methodology section describes how a mixed-methods design is used to test the hypotheses defined in the literature review. Since the four goals of this study require different approaches to gather and analyze information, the following paragraphs discuss the design, data collection methods, sample, measurements and analyses sections for each objective individually. First, the methodology of the content of the psychological contract is examined, which also includes whether the content is transactional or relational. This is followed by the verification of the critical incident technique and finally, the effects of age, gender, sector and function on the psychological contract.

3.1. The content of the psychological contract

3.1.1. Design. The content of the psychological contract in the Netherlands could be identified

by applying quantitative methods to determine to what extent previously determined dimensions apply to the Dutch workforce. However, since the literature review showed that these dimensions may not be generalizable over time or to different countries, this study risks that aspects of the psychological contract specific to current times or the Dutch labour market are neglected. Therefore, an exploratory and qualitative data collection method is more appropriate, since this allows for a richer and more complete understanding of the psychological contract in this specific setting (Saunders, Lewis and Thornhill, 2012. p. 163). In the remainder of this paper, this qualitative method is referred to as Study 1.

The data is obtained using an open-ended questionnaire consisting of one opinion variable to measure the content of the psychological contract and four attribute variables to measure respondent characteristics. This is preferred to semi-structured interviews, since it enables the collection of a larger amount of data. Information will be less detailed, but remains comprehensive enough to answer the research question. The main disadvantage of this method is that, when having many participants, it is time-consuming to code the responses (Saunders et al., 2012, p. 433). By using a limited amount of questions, this method still fits within the time-frame of this study.

In addition, to determine whether the content of the psychological contract is transactional or relational, qualitative methods are used in the form of an additional literature search. Based on the found definitions in this literature, dimensions are divided into the categories.

3.1.2. Data collection and sample. To select a relevant sample from the target population,

convenience sampling is employed. Limitations of this approach need to be recognized, because choosing easily accessible respondents may result in a non-generalizable sample (Saunders et al., 2012. p. 291). However, since participants are contacted by 642 students following the course Organization and Management at the University of Amsterdam, the sample remains diverse. To collect the data, students are provided with a printed copy of the questionnaire, which is then completed by one respondent of their own choice. Since the survey consists of open questions, it is important to take

(17)

17 layout into consideration. A large amount of available space enhances the completeness of the response, but if there is too much space it can demotivate respondents (Saunders et al., 2012, p. 432). Therefore, half a page is provided to answer the main question. Additionally, the survey is self-administered, meaning that it is completed by the respondents themselves. The questionnaire starts with an introduction regarding the purpose of the study, after which anonymity and privacy are guaranteed to ensure that respondents answer truthfully. The survey concludes with a message thanking the respondents for their time and effort, and provides an opportunity to contact the researcher in case they have questions.

The data collection resulted in a number of 642 responses, of which 18 outcomes are removed because of missing data or double entries in the database. Further, some participants mentioned two or three examples in the question measuring the content of the psychological contract, resulting in a total of 666 critical incidents to be analysed. The participants vary in age, ranging from 15 to 69 years old, with an average of 36 (SD=15.4). Age is further categorized in groups based on Herriot et al. (1997)[1] to allow for comparisons. This includes the age groups of 20 and younger (23.9%), 21 to 30 (24.3%), 31 to 45 (8.4%) and 46 and older (43.4%). Furthermore, the participants are equally represented in gender (49.5% male, 50.5% female), but less so regarding sector. The private sector is overrepresented (83.6%) in comparison to the public sector (16.4%). However, these numbers are acceptable since it is relatively representative of the Dutch labour market, in which 27.9 percent of the Dutch employees have a job in the public sector (CBS, 2014). Additionally, 73.2 percent of the respondents is classified as subordinate and 26.8 percent as manager. According to Visshers and te Riele (2008) 14 percent of the Dutch population had a leadership position in 2014, indicating that the amount of managers in the sample is slightly overrepresented.

3.1.3. Measurements. Study 1 includes the following five measures:

3.1.3.1. Dependent variable: content of the psychological contract. To measure the content of the psychological contract, the individual dimensions comprising a contract need to be identified. To do so, an open-ended questionnaire is constructed, in which the following is asked: “We are interested in your role as an employee of your organization. Can you describe a situation in which, according to you, you were treated badly by the employer? With badly is meant: not in a way how you expect an organization to treat its employees.” As described in the literature review, this method is known as the critical incident technique (CIT). The advantages include that it allows access to rich information and insights into real-life situations (Hughes, Williamson and Lloyd, 2007). Additionally, describing a concrete example, rather than an abstract concept, enhances accuracy and simplifies the answering of a question (Saunders et al., 2012, p. 390). On the other hand, limitations of this method need to be recognized. For example, to make sense of the results this type of data needs to be interpreted, which

[1]

Herriot et al. (1997) include an additional fifth group of 61 and older, but due to an underrepresentation of respondents in this category, assumptions in the statistical models used are violated, as is discussed in more detail in following sections. Therefore, this group is combined with the age group of 46 to 60.

(18)

18 may lead to a subjective analysis (Hughes et al., 2007). However, according to the research philosophy of interpretivism this is not necessarily a problem, but a way of doing research (Saunders et al., 2012, p. 140). An additional limitation is that the CIT has less strong theoretical bases in comparison to other qualitative data collection techniques. Yet this may not be relevant, because of the inductive nature of this part of the study (Hughes et al., 2007).

3.1.3.2. Independent variables: age, gender, sector and function. First of all, age is measured using an open-ended question. To facilitate analyses, the results are re-grouped in larger categories later on. It may be easier to ask respondents directly in which age categories they fall, but this will make it impossible to alter groups if necessary. Additionally, both gender and sector are identified using closed questions in which respondents have two choice options, male or female and private or public, respectively. Finally, function is measured through an open-ended question. The results are then coded and will form the basis for how function is defined in this study, because the literature review did not provide an univocal characterization.

3.1.4. Data analysis. To analyze the responses, the raw data is typed out and put together in

one database to enable coding. The first hundred responses are interpreted and labelled individually by three researchers. This process consists of open coding and then refinement through focused coding. The resulting codes are then discussed and integrated, leading to a new agreed-upon set of dimensions. These contain clear descriptions, so that the researchers know what type of incidents fall within which category when coding the remaining responses. This minimizes the chance that the researchers interpret similar incidents in a different manner. To adequately label the different categories, names are derived from both terms used by respondents and those emerging naturally from the data. Existing labels provided by the literature are ignored due to the exploratory nature of this study and to avoid focusing too much on trying to match the data with existing concepts (Saunders et al., 2012, p. 569). Based on the list of codes, both the initial hundred as well as the remaining responses are (re)coded. Due to time constraints of this study, the data is divided between the researchers. If new responses do not fall into one of the previously determined categories, the researcher comes up with a new code and, again, discusses this with the other researchers to ensure consistency.

3.2. Verification of the critical incident technique.

3.2.1. Design. To verify the use of the critical incident technique (CIT), both qualitative and

quantitative survey designs are adopted. In addition to Study 1, a closed-ended ranking survey (Study 2) is used in which respondents are asked to indicate which dimensions of the psychological contract are most important. The dimensions between which participants can choose result from the findings in the first study. The advantage of this method is that rankings allow for improved comparisons between the different dimensions (Saunders et al., 2012. p. 177). Additionally, in Study 1 respondents are limited to indicating only one or two aspects that are breached, whereas they can rate the importance of multiple dimensions in Study 2.

(19)

19

3.2.2. Data collection and sample. To be able to compare the results of Study 1 and 2, the

samples should be similar. Therefore, Study 2 combines convenience sampling with snowball sampling, in which respondents with particular characteristics, such as having a job in the public sector, are asked if they are willing to spread the survey to acquaintances with similar attributes. The initial respondents are contacted through personal e-mails or messages via social media platforms, for which the personal network is used. As opposed to the paper-and-pencil questionnaire used in Study 1, this survey will be internet-mediated. An advantage is that responses do not have to be entered manually into the database (Saunders et al., 2012. p. 421), which saves time and reduces errors and bias resulting from typing mistakes. Additionally, it gives access to a geographically dispersed area. Since this study is aimed at the entire Dutch workforce, this needs to be represented in the sample (Saunders et al., 2012. p. 421). The main limitation of the internet-based survey is that some respondents may be less familiar with the use of the internet, causing particular groups of people to exclude themselves. However, the extent of this problems is expected to be limited, since 96 percent of the Dutch population owns a computer, of which 84 percent is acquainted with basic computer skills (CBS, 2014b). Additionally, those falling outside this group are foremost people of 75 and older (Akkermans, 2013). This group is unlikely to be part of the Dutch workforce and may be excluded from the research. Furthermore, the questionnaire has a similar structure as the survey of Study 1, but also includes a pilot study to ensure clarity of the questions and the privacy guarantees.

In turn, this leads to a collection of 89 responses, of which 17 are removed because of fall out. This is possibly caused by the fact that the layout of one of the questions is not compatible with the interface of mobile phones. An additional 7 responses are also excluded, because these participants indicate that they never had a job. Therefore, they are not part of the target population. This results in a final sample size of 65, consisting of employees varying in age, gender, sector and function. Age ranges from 17 to 75, with an average of 30.3 (SD=15.046). The age categories are identical to Study 1, consisting of 20 and younger (21.5%), 21 to 30 (49.2%), 31 to 45 (4.6%) and 46 and older (24.6%), indicating an overrepresentation of the group 21 to 30 and an underrepresentation of those between 31 and 45. On the other hand, gender is relatively equally represented, consisting of 47.7 percent males and 52.3 percent females. Additionally, employees in the public sector (38.5%) are, in comparison to the amount of public servants in the Netherlands (27.9%) overrepresented. Similarly, managers (23.1%) are also slightly overrepresented in comparison to the amount of managers in the Netherlands (14%).

3.2.3 Measurements. Study 2 incorporates 5 types of measures:

3.2.3.1. Dependent variable: most important dimensions of the psychological contract. The results of Study 1 will lead to an indication of the content of the psychological contract. Based on the descriptions of these emerging dimensions, items are created that represent the dimensions of Study 1 in Study 2 (Appendix B). For example, an item for work environment would be: “[I expect of my employer that] the physical work environment is safe and healthy,” whereas the dimension input is

(20)

20 represented by “[I expect of my employer that] my opinion is asked and valued.” As the results section shows, there are 14 dimensions and thus 14 items that need to be ranked in order of importance. However, since 14 items is too many to rank accurately, respondents have to choose a top 7 (Saunders et al., 2012, p. 436). To be able to compare the rankings, items that are not important are given the score of a 1 (because of statistical purposes a 0 is not possible), and the most important dimension are given a score of 8. This method does therefore not provide an overview of how important a specific dimension is, but only whether it is more important than one of the other dimensions. An alternative option that has been considered is using Likert-scales, but this may result in a situation where every dimension is found to be important. Ranking forces respondents to choose and is therefore more appropriate.

3.2.3.2. Independent variables: age, gender, sector and function. Both age, gender and sector are measured in the same way as described in the previous section. To determine function, the responses on the open question regarding function in Study 1 are examined and coded. This leads to the identification of two types of positions, subordinates and leaders. In Study 2, these options are provided in a closed ended question. To enhance accuracy of the results, the following descriptions of a leadership position is provided in the survey: “a manager is defined here as someone who coordinates processes, people and means to achieve certain goals.”

3.2.4. Data analysis. For both studies, either the amount of times mentioned or the average

mean is calculated for each dimension. These are compared, so that the dimensions can be ranked according to importance. Because both studies consist of different type of data, it is difficult to statistically test whether rankings between Study 1 and 2 differ. Therefore it is assumed that a difference is evident when the rankings diverge more than 5 positions. This is one third of the total amount of rank positions, and thus seems reasonably different to assume that this is not caused by chance alone.

3.3. The effects of age, gender, sector and function on the content of the psychological contract 3.3.1. Design, collection, sample and measurements. To identify the effects of age, gender,

sector and function on the type and the individual dimensions of the psychological contract, analyses will be performed on both the qualitative and quantitative data as described in previous sections. In theory, the best option would be to focus on quantitative methods, since this is most useful in investigating relationships between variables (Saunders, et al., 2012. p. 162). However, due to time and cost constraints it is not possible to collect large amounts of quantitative data. Therefore it is useful to supplement the small quantitative data set with the large qualitative data set. The collection and measurements methods are described previously.

3.3.3. Data analysis. The different data sets are analyzed in different manners. First of all, the

data collected in Study 1 is analyzed using Pearson’s chi-squared test of independence (Field, 2012. p. 871). This method tests whether the count of respondents falling into a category, differs from the

(21)

21 expected count (assuming the variables are independent), thereby indicating whether they are related or not. Fisher’s exact test is used as supplement when the samples within specific categories are too small to perform a chi-squared analysis (Field, 2012. p. 875). This is the case when more than 20 percent of the cells have an expected count of 5 or less (Field, 2013. p. 723-724).

The second data set is measured through different methods and thus requires a different analysis. In addition to be able to test the level of dependence for ranks, the tests need to be non-parametric, because even after transformations the data is not normally distributed. Therefore, the Mann-Whitney U and its extension, the Kruskal-Wallis tests are used. This is supplemented with Spearman’s rho, because assumptions of both type of non-parametric tests are violated for specific cases. These assumptions include that the two variables that are being tested have similar distributional shapes and that the relationships between the variables are monotonic. Although using two methods increases the likelihood of making a type II error, it decreases the probability of making a more severe type I error (Saunders et al., p. 513).

(22)

22

4. Results

4.1. Reliability

To ensure that consistent findings are generated, several actions are undertaken. First of all, as stated in the methodology, multiple researchers collaborated to identify the content of the psychological contract. This avoids multiple problems, such as interpretation and subjectivity bias. In addition, a second study is performed, which verifies whether this paper can use the most violated dimensions as the most important ones. This also serves to confirm whether the identified content is present in a different sample’s psychological contract. Furthermore, two statistical methods are applied in Study 2 to ensure the effects are legitimate. Since the measures for the content of the psychological contract in Study 2 all consist of only one item, no Cronbach’s alpha can be calculated.

4.2. The content of the psychological contract

By coding the responses obtained through Study 1, 14 dimensions of the psychological contract have been identified. An overview of their descriptions is provided in Table 1 on the following page.

(23)

(24)

24

4.2.1. Transactional versus relational. Table 2 provides an overview of the frequencies with

which the dimensions are breached, as well as whether they are transactionally or relationally oriented. This distinction is based on the definitions provided in the literature review, supplemented with opinions of scholars.

(25)

25 Whereas some aspects of the psychological contract, such as pay and humanity, can be straightforwardly assigned to a category, others require some additional explanation. One of which is justice. As the following quote illustrates, justice, or equal treatment, is mainly oriented towards transactional aspects of the psychological contract:

‘A male colleague working at the same level as me, and who was employed later than I was, had a higher wage than me.’

In accordance, Mottaz (1985) suggests that pay-equity has an organizational-extrinsic orientation and is thus transactional in nature. However, not all justice-related incidents are associated with pay. Therefore, the content of the above passage needs to be interpreted in more general terms. The respondent indicates here that it is the treatment that is unfair, not necessarily the pay level on its own. This suggests that justice is associated with the employer-employee relationship, since no supervisory support is provided to realize equal treatment. In turn, this is associated with a social-extrinsic categorization (Mottaz, 1985) and thus the relational psychological contract. Further support for this statement is provided by the following quote, which indicates that unfair treatment may also be caused by interpersonal relationships at the work floor:

‘One of my colleagues is a friend of the manager. She does not want to work nightshifts, because […] she cannot sleep well during the day. The scheduled nightshifts are then changed with the help of the manager to day or evening shift [thereby disadvantage other employees]’

Finally, this study also identifies separate dimensions that focus on pay levels, work hours and other transactional variables towards which equal treatment may be oriented. In cases where the respondents indicate that the incident is, for example, related to the pay itself rather than equal treatment, it is already subdivided in transactional dimensions such as wage. Consequently, justice is only about cases in which the treatments are viewed as unfair and is therefore categorized as relational.

A similar problem is present for validity. Validity is associated with retaining a job or being chosen for a function, and could be categorized as transactional. However, like justice, the concept itself is about the underlying explanations and feedback given when losing a job or position. The latter are part of an intrinsic orientation and thus validity is categorized as relational (Mottaz, 1985).

Finally, there is some ambiguity regarding the dimension ethics and legality. Unlike most relational variables, this factor is neither related to the value of performing the job itself, nor to the relationship between the employee and employer. However, Vansteenkiste et al. (2007) argue that community contributions are also included in the intrinsic dimension. This is defined as acting in

(26)

26 someone else’s interest and helping others in need (Vansteenkiste et al., 2007). The following example shows that this definition fits with the ethics and legality dimension as described in this study:

‘They said [the employer]: ‘always charge them for the large menu’s if they do not clearly state their preferences and include extra ketchup and mayonnaise as well. Especially with tourists, because they do not know the real prices [in this country].’ I had to process this and said: ‘But then they pay more than they are supposed to, meaning that I am ripping them off.’ After which the employer answered: ‘Yes, that is the purpose’. […] This went against my principles, so I did not do it.”

The quote shows that the person in question resists the employer’s unethical behavior and thereby acts in the interest of someone else, eventually resulting in her resignation. This is thus similar to how Vansteenkiste et al. (2007) define intrinsic values, and thus legality and ethics are classified as relational.

Based on these categorizations, the results show that overall, transactional (49.6%) and relational (50.7%) aspects of the psychological contract are equally valued[2]. However, when determining the averages scores of the dimensions in Study 2, discussed in more detail in the next section, it shows that transactional dimensions (3.97) are valued more than relational dimensions (2.88), although it cannot be determined how large this difference is.

4.3. Verification of the critical incident technique

To verify the use of the critical incident technique to identify the most important dimensions of the psychological contract, the most breached dimensions of Study 1 and the most important dimensions of Study 2 are ranked and compared, as shown in Table 3 below. Table 3 indicates that out of 14 dimensions, seven have similar rankings within a range of 5 positions. These include wage, input, respect, justice, job security, legality and ethics, and atmosphere. However, the remaining seven dimensions have divergent rankings. Most notable is training, which is ranked at the second last place according to Study 1, while respondents in Study 2 believe this dimensions to be the most important. Similarly, whereas work hours is ranked at the third place in Study 1, it is one of the least important dimensions in Study 2.

Concluding, the ranks indicate that the dimensions of the psychological contract that are most often violated, are not necessarily the most important. Since the critical incident technique only measures breach, it may not be an appropriate method for determining which aspects of the content of the psychological contract are most important.

[2]

(27)

27

4.4. Psychological contract type in relation to age, gender, sector and function

To determine the relationship between age, gender, sector and function, and the psychological contract, two separate analyses on Study 1 and 2 are performed.

4.4.1. Study 1: Most violated dimensions. Results of the performed Chi-squared analyses are

shown in Table 4 below.

4.4.1.1. Age, gender, sector and function. Since control variables cannot be included in the chi-squared analyses, it is important to take into account whether the independent variables correlate. If so, this can artificially enhance the effects of the independent variables on the content of the psychological contract.

Four relationships are found. First of all, both the public sector (𝜒2_{(1, N=666) = 79.407, p <} .01; 𝑉 = .345) and leadership positions (𝜒2 (1, N=666) = 142.755, p < .01; 𝑉 = .463) contain a larger

(28)

28 amount of older age groups, in particular those of 46 and older. Additionally, there are more females active in the public sector (𝜒2 (1, N=666) = 12.695, p < .01; 𝜙 = .138), whereas there are less females in leadership positions 𝜒2 (1, N=666) = 19.874, p < .0; 𝜙 = -.122). This indicates that relationships of age, gender, sector and function with the psychological contract may be exaggerated due to combined effects.

4.4.1.2. Psychological contract type and age, gender, sector and function. The analysis confirms the expected age effect on the content of the psychological contract as shown in line 5 of Table 4 (𝜒2 (1, N=666) = 38.078, p < .01; 𝑉 = .239). When exploring this in more detail, it is apparent that respondents of 20 and younger experience more violations of transactional dimensions of the psychological contract than the age groups 21 to 30 (𝜒2 (1, N=666) = 7.389, p < .01; 𝜙 = .152), 31 to 45 (𝜒2_{(1, N=666) = 9.515, p < .01; 𝜙 = .210) and 46 and older (𝜒}2_{(1, N=666) = 36.613, p < .01; 𝜙 =} .286). Additionally, respondents between 21 and 30 also experience more violations of transactional aspects than those of 46 and older (𝜒2_{(1, N=666) = 9.536, p < .01; 𝜙 = .145). This indicates that as} people grow older, they experience more breaches of relational dimensions of the psychological contract. In other words, the type of violations differ for different age groups, which provides support for hypothesis 1a.

Additionally, line 5 of Table 4 shows that gender and sector do not influence the content of the psychological contract, indicating that hypotheses 2a and 3a are not supported.

Finally, an effect of function on the content of the psychological contract is found (𝜒2 (1, N=665) = 8.935, p < .01; 𝜙 = .116). The results indicate that managers are more likely to experience violations of relational aspects of the psychological contract, whereas subordinates experience more violations of transactional aspects, thereby supporting hypothesis 4a.

4.4.2. Study 2: Most important dimensions. The results of the Spearman’s Rho,

Mann-Whitney U and the latter’s extension, the Kruskal-Wallis test are displayed in Table 5 below.

4.4.2.1. Age, gender, sector and function. In comparison to Study 1, less of the independent variables correlate. However, (partially) significant relationships between function, age and gender as

(29)

29 described in Study 1 are also confirmed in Study 2. This includes that there are more older age groups (N=65. 𝑅𝑠 = .309, p < .05) and less females in managerial positions (N=65. 𝑅𝑠 = -.208, p < .1). 4.4.2.2. Psychological contract type and age, gender, sector and function. Similarly to Study 1, line 5 of Table 5 shows that age and the type of psychological contract are related. This is confirmed by both the Spearman’s Rho (𝑅𝑠 = .272, p < .05) and the Kruskal-Wallis (𝜒2 (3, N=65) = 10.032, p < .05) tests. Surprisingly, respondents in the age category of 46 and older value transactional psychological contracts more in comparison to those of 20 and younger (N = 30; 𝑅𝑠 = .474, p < .01; U = 51.0, p < .05), 21 to 30 (N = 48; 𝑅𝑠 = .381, U = 137, p < .01) and 31 to 45 (N = 19; 𝑅𝑠 = .479, U = 6.0, p < .05). Hypothesis 1b is thus supported, but not in the same direction as Study 1 confirms. No support is found for the effects of gender, sector and function on psychological contract type. This means that hypotheses 2b, 3b and 4b are not confirmed by Study 2.

4.5 Psychological contract dimensions in relation to age, gender, sector and function

To provide a more detailed insight into the content of the psychological contract, an additional exploratory study is performed to identify the effects of age, gender, sector and function on the individual aspects of the psychological contract for both the most violated and the most important dimensions. First, the results of Study 1 will be discussed, followed by those of Study 2.

4.5.1 Study 1: Most violated dimensions. An overview of the results of the Chi-squared and

Fisher’s exact tests is provided in Table 8 in Appendix C.

4.5.1.1. Dimensions and age. Of the 14 dimensions of the psychological contract, six are influenced by age. First of all, this includes work hours (𝜒2 (3, N=666) = 50.410, p < .01; V = .275). line 1 of Table 6 shows that younger age groups encounter more violations of this dimension, which is consistent with the expectation that younger employees are more likely to experience breaches of transactional dimensions of the psychological contract. Additionally, a partially significant effect for respect is found (𝜒2 (3, N=666) = 6.910, p < .1; 𝑉 = .102). The middle age group experiences most violations of respect, as shown in line 2 of Table 6. This is inconsistent with the expectation that the older age groups encounter more violations of relational aspects of the psychological contract. A third effect is found for input (𝜒2 (3, N=666) = 31.721, p < .01; 𝑉 = .218). Line 3 of Table 6 shows that, consistent with expectations, input is more often breached for older age groups. Similarly, older age groups also experience more violations of the relational aspect of validity (𝜒2 (3, N=666) = 17.959, p < .01; 𝑉 = .164), as shown in line 4 of Table 6. Fifth, a partially significant effect of age on the physical work environment is found for three of the age groups. An overall effect could not be determined, because of a violation of the chi-squared analyses. However, the individual effects shown in line 5 of Table 6 show that, consistent with expectations, the physical work environment is experienced to be breached mostly by younger age groups. Finally, for the same reasons as mentioned for the physical work environment, an overall effect for training could not be determined. However,

(30)

30 line 6 of Table 6 shows that, when comparing the individual age categories, training is less valued by older age groups, which is consistent with expectations.

4.5.1.2. Dimensions and gender. Differences in the extent to which males and females experience violations of the dimensions of the psychological contract are limited. Only two dimensions are affected by gender and they are only partially significant. The results indicate that work hours (𝜒2_{(1, N=666) = 3.796, p < .1) and atmosphere (𝜒}2_{(1, N=666) = 3.508, p < .1) are more} often breached in women’s psychological contracts. Only the latter is consistent with the expectation that women are more likely to experience breach of relational aspects of the psychological contract. However, the effect sizes for both dimensions are small (𝜙 = .075), indicating there is little to no association.

4.5.1.3. Dimensions and sector. Similarly to gender, differences in breach between private and public sector employees are limited, although some minor relationships are found. First of all, work hours (𝜒2 (1, N=666) = 5.608, p < .05; 𝜙 = -.092), as well as wage (𝜒2 (1, N=666) = 2.879, p < .1; 𝜙 = -.066) are more often breached in private sector employees’ contracts, although the latter is only partially supported. Finally, input is more often experienced to be breached for public servants (𝜒2 (1, N=666) = 14.475, p < .01; 𝜙 = .147). These results are consistent with the expectation that

An employee perspective on the content of the psychological contract : exploring the effects of age, gender, sector and function on the psychological contract