Malaysian Airline System Human Resource

Work Motivation
-Studies of its Determinants and Outcomes

Christina Bjorklund

AKADEMISK AVHANDLING

Som for avlaggande av filosofie doktorsexamen vid Handelshogskolan i Stockholm framlaggs for offentlig granskning fredagen den 27 april 2001, kIlO.I5 i Aulan, Handelshogskolan, Sveavagen 65.

Work Motivation
-Studies of its Determinants and Outcomes

tJl\ STOCKHOLM SCHOOL OF ECONOMICS EFI, THE ECONOMIC RESEARCH INSTITUTE
~~,- '

EFIMission EFI, the Economic Research Institute at the Stockholm School of Economics, is a scientific institution which works independently of economic, political and sectional interests. It conducts theoretical and empirical research in management and economic sciences, including selected related disciplines. The Institute encourages and assists in the publication and distribution of its research findings and is also involved in the doctoral education at the Stockholm School of Economics. EFI selects its projects based on the need for theoretical or practical developnlent of a research domain, on methodological interests, and on the generality of a problem. Research Organization The research activities are organized in nineteen Research Centers within eight Research Areas. Center Directors are professors at the Stockholm School of Economics.
ORGANIZATION AND MANAGEMENT Management and Organisation; (A) Center for Ethics and Economics; (CEE) Public Management; (F) Information Management; (I) Center for People and Organization (PMO) Center for Innovation and Operations Management; (T) ECONOMUCPSYCHOLOGY Center for Risk Research; (CFR) Economic Psychology; (P) MARKETING Center for Information and Communication Research; (CIC) Center for Consumer Marketing; (CCM) Marketing, Distribution and Industrial Dynamics; (D) ACCOUNTING, CONTROL AND CORPORATE FINANCE Accounting and Managerial Finance; (B) Managerial Economics; (C) FINANCE Finance; (FI) ECONOMICS Center for Health Economics; (CRE) International Economics and Geography; (lEG) Economics; (S) , ECONOMICS STATISTICS Economic Statistics; (ES) LAW Law; (RV)

Prof Sven-Erik Sjostrand Adj Prof Hans de Geer Prof Nils Brunsson Prof Mats Lundeberg Acting Prof Jan Lowstedt Prof Christer Karlsson Prof Lennart Sjoberg ProfLennart Sjoberg

Adj Prof Bertil Thorngren Associate Prof Magnus Soderlund Prof Lars-Gunnar Mattsson Prof Lars Ostman Prof Peter Jennergren Prof Clas Bergstrom Prof Bengt Jonsson Prof Mats Lundahl Prof Lars Bergman Prof Anders Westlund Prof Erik Nerep

Chairman ofthe Board: ProfSven-Erik Sjostrand. Director: Associate ProfBo Sellstedt,
A dress EFI, Box 6501,8-113 83 Stockholm, Sweden • Internet: www.hhs.se/eftl Telephone: [phone] • Fax: [phone] • E-mail [e-mail]

Work Motivation
- Studies of its Determinants and Outcomes

Christina Bjorklund

t,!t\ STOCKHOLM SCHOOL OF ECONOMICS
~i7 EFI, THE ECONOMIC RESEARCH INSTITUTE

~~ID 'S 'J
". '6 ',11 ,~

f~" Dissertation for the degree of Doctor of Philosophy, Ph.D.
Stockholm School of Economics, 2001.

© EFI and the author ISBNNR 91-7258-555-2

Keywords:

work motivation job performance work interest child care insurance company

Printed by:

Elanders Gotab, Stockholm 2001

Distributed by: EFI, The Economic Research Institute Stockholm School of Economics Box 6501, S-113 83 Stockholm, Sweden

To mymother

Acknowledgements
The years of working on my thesis at the Stockholm School of Economics have been exciting, interesting, and filled with many discussions and stimulating encounters with people. I would like to express my deepest gratitude to my advisors. First of all, I would like to thank my n1ain advisor professor Lennart Sjoberg, who aroused my interest in the topic of work n10tivation. His advise, ideas and critical comments have been of great value to me. I would like to thank my second advisor, professor Erland Svensson, who has given me valuable comments on my thesis. I also would like to express my gratitude to professor Bo Ekehammar for all help and encouragement he has given me.

I have had the privilege of being surrounded by interesting and supportive people of different academic backgrounds, which has increased my understanding of different research traditions, and, moreover, inspired the integration of different academic disciplines in my work. I would like to thank the past and the present people of the Economic Psychology Unit and CFR: Jana, Ylva, Arja, Jing Guan, Johan, Anders, Frederic, Pemilla, Asa, Helena, Patric, Joakim, Hakan, Per Henrik, and Mattias. In addition, special thanks to Caroline Nordlund for checking the language in the thesis and Gudrun for all support during difficult times. I am very grateful to all the people who kindly participated in my studies. Thanks to the Swedish Council for Work Life Research, TIle Economic Research Institute, and The School of Economics for financing my studies and to Rune Castenas and Lisa Tilert who have been very helpful with all financial matters.

For all support and encouragement, I would like to thank my wonderful father and sister, Anna-Lena and her family. A special gratitude to my mother, who showed me to never give up. Last but not least, I would like to thank Thierry for always being there for me and for always believing me. I would not know if I would have made it without you. Thank once again to you all, I would not have had the courage and the strength to complete my thesis without you.

Bromma, March 15,2001

Christina Bjorklund

Table of content
1. Introduction
1.1. Purposes of the thesis............................................................ 1.2 Organization of the thesis............ 2. Theoretical framework. . .. ... .... ....... .... . .. . ... ... .... .... ........ .. .... 2.1 Definition of work motivation.................................................... 2.2 Trends in work motivation theorizing 2.3 Work motivation theories 2.3.1 Introduction....................................................................... 2.3.2 Work motivation defined as willingness to work........................... 2.3.2.1 Will/volitional variables as predictors ofbehaviors 2.3.2.2 Theconceptofwill 2.3.3 Beliefs and attitudes 2.3.3.1 Job satisfaction 2.3.3.2 Organizational commitment 2.3.4 Incentives and rewards theory (extrinsic motivation) 2.3.5 Intrinsic motivation 2.3.6 Goal and goal setting............................................................ 2.3.7 Perceived Control 2.3.8 Factors related to work motivation 2.3.8.1 Occupational and organizational stress 2.3.8.2 Work interest '" 2.3.8.3 Creativity 2.3.8.4 Work Environment 2.3.8.5 Perceivedrisks 2.3.8.6 Background/actors............................................................ 2.3.8.7 Management/leadership 2.3.9 Summary 2.4 Work related behavior............................................................ 2.4.1 Direction (withdrawal behaviors) 2.4.1.1 Absenteeism 2.4.1.2 Turnover (Intention to quit/leave) 2.4.2 Persistence (Number of hours worked) 2.4.3 Intensity (Job performance) 2.4.3.1 Ratings ofjob performance 2.4.4 Summary.... . . .. .. . . .. . .. .. . .. . .. .. .. .. . . .. .. . . ... . . .. . .. ... .. . .. . . .. .. . . .. . . . 3. Research approach 3.1 Introduction 3.2 Study 1 (pre-school employees) 3.2.1 Participal1ts........................................................................ 3.2.2 Procedure

1
2 3 4 4 5 9 9 11 11 12 14 17 21 24 28 30 34 36 36
39 41

45 47 49 51 53 54 54 55 56 58 58
62 63 64 64 66 66 66

3.2.3 Questionnaire 3.2.3.1 Scales include 3.2.3.1.1. Indices 3.3 Study 2 (employees of an insurance company). .... ..... . .. .. .... .. . ... .. .. 3.3.1 Participants........................................................................ 3.3.2 Procedure 3.3.3 Questionnaire 3.3.3.1 Scales include 3.3.3.1.1 Indices 3.3.3.1.2 Conclusion 3.4 Study 3 (salespeople) 3.4.1 Participants 3.4.2 Procedure 3.4.3 Questionnaire 3.4.4 Scales included. . .... . ... .. . ... .. .... . .. .. ... ... ..... . .. .. .. ... .... . ...... . 4. Results 4.1 Descriptive results 4.1.1 Introduction 4.1.2 Study 1: Descriptive results 4.1.3 Study 2: Descriptive results 4.1.4 Study 3: Descriptive results 4.2 Determinants of willingness to work .. . .. ..... . ..... ..... .. .. . .. ... ..... 4.2.1 Introduction.. .. .. ... .. ... ..... . .. . .. 4.2.1.1 Study 1. ' Determinants of willingness to work 4.2.1.1.1 Conclusion 4.2.1.2 Study 2: Determinants ofwillingness to work 4.2.1.2.1 Conclusion 4.2.2 A structural model of willingness to work................. 4.2.2.1 Introduction.................................................................... 4.2.2.2 Study 1: Structural model ofwillingness to work 4.2.2.1.1 Model estimation and modification. . . .. ......... ... .......... .. .. .. ... ... 4.2.2.3 Study 2: Structural model ofwillingness to work 4.2.2.3.1 Model estimation 4.2.2.4 Conclusion 4.3 Results of construct validity studies 4.3.1 Introduction 4.3.1.1 The relationship between willingness to work and work-related behaviors. . . . .. ... ... . ..... ......... .. ....... ........... ......... ... ........ .............. 4.3.1.1.1 Conclusion 4.3.1.2 Willingness to work and other work motivation measures 4.3.1.3.1 Conclusion

67
68

72 82 82 82 82 83 84 91 92 92 92 92 93 95 95 95 96 98 99 100 100 101 104 106 109 111 111 113 114 117 117 119 120 120 121 124 126 128

11

4.3.1.3 Is willingness to work a more efficient measure ofwork motivation than the two traditional ones? . . .. ... . .... .. .. ... ... ... ... .. ... ... .. . ... .. .. . .. . ... 4.3.1.3.1 Conclusion 4.3.1.4 Does work motivation in general predict work-related behaviors? .. 4.3.1.4.1 Conclusion 5. General discussion 5.1 A summary of the findings 5.2 Limitations 5.3 Further researc11.......................................... 6. References 7. Appendix. .. .. . ... ... .. .. .. .... . .... .. .. .. . .. ...... ... ... . ...... . .. . .. .. 7.1 Appendix A (service industry). ........ ... ..... ... ... ...... .... . ... . .. ..... . . ... ... 7.2 Appendix B (scales) 7.3 Appendix C....
0 • 0 • 0 0 • 0 0 •• 0 0 0 ••• 0 •• 0 ••••••••••••••••• 0 •• 0 ••••••••••••••••• 0 •••• 0 ••• 0" 0 •••• 0 •••• 0 ••••• 0

129 132 133 134 135 135 138 140 142 181 181 194 209

iii

1. Introduction
How to motivate employees has been one of the central concerns for many researchers and practitioners for decades. During the years, many work motivation theories have been presented in the literature. The theories have provided different conceptualizations of the factors that drive the process by which behavior is energized, directed, and sustained in organizational settings (Leonard, Beauvais, & Scholl, 1999). Increased interest in work motivation is due to higher global competition, changes in work tasks, and the fact that today 's growing industry is the service industry where the employee is the main asset. The main assumption of the concept of work motivation is that it predicts different work-related behaviors such as absenteeism, turnover, and job performance. To have a high rate of absence as well as turnover is very costly for organizations, and having personnel that is highly motivated might prevent some of the withdrawal behaviors. In addition, to have employees who are hard working and perform well will have the opposite effect on the profitability of organizations. Moreover, traditional employment arrangements have been replaced by outsourcing, temporary work, and individual career paths (Hall, 1996). Further, work activities have been reshaped by new technology and have shifted from manual labor to knowledge and service work (Adler, 1992; Howard, 1995). This has also been the case in Sweden, where the economy has gone through a revolutionary change, where the population working in manufacturing industries has been reduced and the expanding sector is that of services (Nelson, 1994). Because working conditions and work forms are constantly changing, new studies are called for in order to increase the understanding of factors and processes that influence an enlployee 's motivational level. In the present thesis, work motivation was examined in different service industries. Work motivation was defined as willingness to work, a measure introduced by Sjoberg and Lind (1994). They found that factors explaining the

variance of willingness to work were work interest, creativity, perceived risks, and organizational commitment.

1.1. Purposes of the thesis
Work motivation, its determinants, and work-related behaviors as a result of motivation have been studied for decades. Many different theories trying to explain work motivation as well as related factors have been developed over the years. Although work motivation has bee11 studied for quite some time, measures are still used that have not shown to be valid assessments. Two wellestablished measures of work motivation are job satisfaction and organizational commitment (Maehr & Braskamp, 1986). These measures are often used in Sweden and internationally without scrutinizing their validity as measures of work motivation. Ge11erally, these measures of work motivation have shown to be quite weakly related to job performance (see for example, Iaffaldano & Muchinsky, 1985; Leong, Randoll, & Cote, 1994; Mathieu & Zajac, 1990). However, the validity of these measures of work motivation has been taken for granted by many researchers and practitioners. Consequently, Sjoberg and Lind (1994) suggested an alternative measure of work motivation, defined as willingness to work. This measure has the purpose of assessing how willing a person is to work. An argument for a volitional approach of measuring work motivation is that previous studies have shown that volitional approaches are the most efficient ones in predicting and explaining action (Ajzen, 1991; Sjoberg, 1998). Moreover, how willing a person is to work n1ay be reflected in voluntarily actions, how important the work is to the person etc, and these types of questions were included in the present work motivation measure. The first objective of the present thesis was to explore the determinants of work motivation, defined as willingness to work. The independent variables included were selected on the basis of the previous findings of Sjoberg and Lind (1994) as well as results presented in previous studies in the international literature.

2

Moreover, the factors that in the Sjoberg and Lind (1994) study were found to explain work motivation were primarily work interest, perceived risks, creativity, and organizational commitment. These independent variables accounted for around 60% of the variance in work motivation. The second objective was to investigate the relationship between work motivation, defined as willingness to work, and different work-related behaviors. The work-related behaviors included were absenteeism, intention to quit, number of hours worked (per week), and job performance. This makes it possible to evaluate the construct validity of work motivation. It was also examined whether the new measure of work motivation could be considered as a more valid assessment than two more established measures Gob satisfaction al1d organizational commitment). Three work groups were included in the studies. The groups were: employees of pre-schools, employees of an insurance company, and insurance sales personnel. The reasons for selecting those particular groups were, firstly, that they all belong to the service industry 1, and secondly, that employees in these work groups have experienced drastic changes over the years.

1.2 Organization of the thesis
In Chapter 2, a theoretical framework is presented. In this part, the definition of work motivation, trends in work motivation theorizing, and different work motivation theories including a presentation of the work motivation measure used in the thesis are discussed. Further, different work-related behaviors are introduced and some general information about the service industry is given. In Chapter 3, the research approacll enlployed is presented. The results of the studies are presented in Chapter 4. In Chapter 5, a general discussion, including linlitations of the studies, and suggestions of studies further in the subject are shown. Finally, Chapters 6 and 7 include references and appendices.
I

Read more about service industry in Appendix A.

3

2. Theoretical framework
2.1 Definition of work motivation
Motivation can be described as the need or drive that incites a person to some action or behavior. The verb motivate means to provide reasons for action. Motivation, then, provides a reason for exerting some sort of effort. This motivation springs forth from individual needs, wants, and drives (Timm & Peterson, 2000). Motivation is not directly observable. Because it is 110t observable, motivational processes can be inferred only from analysis of these continuing streams of behavior that are determined both by environment and heredity and are observed through their effects on personality, beliefs, knowledge, abilities, and skills (Kanfer, 1990). There is no generally accepted definition of work motivation, and there are many reasons for the apparent difficulty to define motivation. Firstly, there are many philosophical orientations toward the nature of hllman beings and about what can be known about people (Pinder, 1998). Secondly, it is a fact that historically motivatiol1 has been used to explain "too much with too little" (Ferguson, 1976, p. 6). In the light of these difficulties in obtaining a comprehensive yet more informative definition, it seems reasonable to focus on the adequacy of those definitions offered in different approaches. The definitions appear, however, generally to have three common denominators which may be said to characterize the phenomenon of motivation. That is, when we discuss motivation, we are primarily concerned with: (1) what energizes human behaviors; (2) what directs or channels such behavior; (3) how this behavior is maintained or sustained (Steers & Porter, 1983).

4

A definition that covers these denominators is presented by Pinder (1998) where work motivation is defined as "a set of energetic forces that originate both within as well as beyond an individual 's being, to initiate work-related behaviors, and to determine its form, direction, intensity, and duration" (p.11). Further, Pinder (1998) described intensity as the momentary magnitude of actual motivational arousal, regardless of potential available. Moreover, to fully understand work motivation, one must allow for the specific goals toward which motivated energy is directed (Katerberg & Blau, 1983).

2.2 Trends in work motivation theorizing
Theories of work motivation have passed through many stages, influencing and being influenced by the prevailing management ideologies and philosophies of each era. Although it is possible to trace a sequence to this development, it does not mean that the old theories have died. There are employers and managers today adhering vigorously to one or other of them, basing their beliefs not on research or empirical evidence but on an almost ideological framework of values and assumptions. These beliefs help then1 understand their own role and those of others around them. During the early part of this century, the predominant theory about management was the classical or "scientific" management approach. This theory portrayed working people as making rational economic calculations and following a consequent logical pattern of behavior at work (for review, see Taylor, 1947). Employers, who accepted this theory, believed that their workforce was driven by the desire to earn the most money possible. However, after testing the dominating theory the conclusion was that behavior at work could not fully be explained by reference to the desire to earn as much money as possible. A new theory was Pllt forward (see Mayo, 1949) proposing

5

that the reason why son1e workers slowed down their effort towards the end of their day n1ust result from some factor which was preventing these workers from keeping up their effort. The most likely factor was fatigue; workers were not strong enough or sufficiently well nourished to keep their effort up all day. This led to research studies by Elton Mayo and his team fro111 Harvard University (see Mayo, 1949). The research tean1 set up a major series of studies at the Hawthorne Works of the Western Electric Company which continued for ten years. Their aim was to study the effects of a range of fatigue-inducing factors such as levels of lighting, temperature, frequency of breaks, etc. in combination with an incentive payment by results systems (Landberger, 1958). Their findings were not expected. The variable that enhanced the productivity among the employees was not the level of lightning, temperature etc but rather the increased interest shown by the company in its employees, by regularly asking questions about their health, morale, personal lives, etc. This unintentional effect of observing people at work became known as "The Hawthorne Effect", and the result had an almost revolutionary effect on prevailing theories of motivation to work. Instead of focusing on money as the motivator, attention turned to the importance of "human relations" as a mean of motivating employees. Motivation theories were then developed which under-pinned or built upon the "human relations" findings. The new focus for motivation theory was on the search for satisfaction of human needs. The new approach swept through n1anagen1ent thinking in the 1950 's. Motivation theories emphasizing what it is that motivates people, and theories included in this paradigm, were Maslow 's need hierarchy and Hertzberg 's two factor theory. Moreover, Maslow (1970) offered his "need hierarchy" according to which human beings have their needs arranged in a hierarchy such that they are motivated to seek satisfaction of the lower levels of need first. Once that level of need is satisfied it is no longer a motivator, and the person is motivated by the next level up the hierarchy. Basic needs such as s11elter, food and warmth
6

are at the bottom level of Maslow 's hierarchy, which then progresses through physical well being, social acceptance, self-esteem, to "self-actualization" (realizing one 's own potential). A second well-known theory in this category is Hertzberg 's two-factor theory. According to Hertzberg, work motivation is dependent on "hygiene" factors (salary, prestige) and motivators (achievement, responsibility). A person is motivated if both kinds of needs are satisfied. Herzberg declared that "real" motivation is only reached when a person experiences self growth, which can only be satisfied through work enrichment and teamwork (Herzberg, 1968, 1987). There is a third need theory developed by Hackman and Oldham (1976), the Job Characteristic Model (JCM). They claimed that job enrichn1ent is based on five increasing core dimensions: skill variety (the extent to which a job entails different activities and involves a range of different skills and talents); task identity (the extent to which ajob involves completion ofa whole piece of work with a visible outcome); task significance (the extent to which a job has meaningful impact on other people, either inside or outside the organization); autonomy (the extent to which a job provides freedom, independence and discretion in planning the work and determing how to undertake it); feedback (the extent to which work activities result in direct and clear information on the effectiveness of job performance). These core characteristics, if presented in work tasks, create three psychological states: experienced meaningfulness, experienced responsibility, and knowledge of results. These factors, if present, are assumed to produce satisfaction and motivation to promote high quality work (Hackman & Oldham, 1976, 1980). All the need theories have been tested empirically, but none of them has received much scientific support. According to other researchers, Maslow 's theory is difficult to test and very few studies have demonstrated the validity of the theory (Salancik & Pfeffer, 1977; Kanfer, 1990; Neher, 1991). Only a weak relationship has been found between Herzberg 's factors and work motivation
7

(Sjoberg & Lind, 1994). Hackman and Oldham 's theory has been tested by others and those studies have found a weak relationship, approximately 0.15, between experienced task characteristics and performance. Thus, the Job Characteristic model only explains around 2% of the variance in performance (Fried & Ferris, 1987; Stone, 1986; Loher, Noe, Moeller, & Fitzgerald, 1985). Although the core dimension does appear to influence, very weakly, personal and work outcomes, there is some doubt about the validity of the causal relationship (Wall, Clegg, & Jackson, 1978). A major development in motivation theory in the 1970s was based on Vroom 's Expectancy Theory (Vroom, 1964). The theory argues that a person 's motivation to perform a given act will depend on valences, instrumentality, and expectancy. According to the theory, the act with the highest motivational force is the one the individual will choose to pursue (Locke & Henne, 1986). The theory is founded on the idea that people prefer certain outcomes from their behavior over others. They anticipate feelings of satisfaction, should the preferred outcome be achieved (Vroom, 1964). However, Van Eerde and Thierry (1996) reported a quite low correlation, around 0.20, between the variables in Vroom 's VIE theory and performance. Another well established theory, Equity theory, has experienced its ups and downs since Adams (1965) first proposed it as a way of understanding how employees respond to situations in which they are treated more or less favorably in conlparison to a referent "other". Equity theory focuses on people 's feelings of how fairly they have been treated in comparison with the treatment received by others. The equity model (Adams, 1965) is based on the assumption that humans want to be treated equally for their services. People are motivated by their need for fair treatment. Equity exists when output (e.g., ability, seniority) and input (e.g., money, promotion) ratios for the individual employee and the reference source (e.g. co-worker, profession) are equal (Katzell & Thompsson, 1990). Job dissatisfaction is assumed to occur whenever a person perceives that he or she is not being paid equally with others. Witt and Nye (1992) found in a meta-analysis that perceived fairness of payor promotion and job satisfaction
8

had an average correlation of 0.28 and 0.43. In a study by McFarlin and Sweeney (1992), the same conclusions were reached. Theories of goals and targets have become popular in recent decades, for instance Locke 's argument that people are motivated by relatively difficult goals that they have agreed to seek (Latham & Locke, 1979). This puts the source of motivation not on some "need" of the employee, but on the achievement of a goal with which he/she has been involved. Similar to goal theory are those theories of behavior modification, such as guidance, prompting, feedback, and reinforcement to bring about the desired changed (for review, see Guest, 1984). Goal setth1g theory will be further discussed below.

Kanfer (1992) stated that the direction and focus of recent work motivation theory and research correspond closely to new development in the broader field of motivational psychology. During the past ten years, major integrative theories of hun1an motivation have been proposed by clinical, instructional, social, and personality researchers, including Bandura (1986, 1988), Carver and Scheier (1981), Heckhausen and Kuhl (1985), and Weiner ( 1986). Many new research directions in work motivation are built upon earlier theories and studies ,and as Barbuto and School (1998, p.1 011) wrote, " an integrative taxonomy that may better account for various motivations is necessary to advance our understanding of individual behavior".

2.3 Work motivation theories
2.3.1 Introduction
If there is a cornerstone il1 the science of human behavior, it must be the field of motivation. In fact, motivation has been described as "one of the most pivotal
9

concerns in organizational research (Baron, 1991, p. 1). Motivational theories ask a fundamental question, namely: What moves a person? Thus, they are concerned with the prime forces at work in human nature and human culture (Ryan, 1998). The importance of motivation as an organizing psychological concept is in one sense remarkable given its status as a hypothetical construct that can be only indirectly inferred from observed behaviors. As such, it can take on many forms, and this is reflected in the variety of theories and taxonomies of motivation (Heckhausen, 1991). However, Locke (1991) claimed that the field of work motivation has become increasingly confused over the past several decades. The major cause of this confusion has been an overabundance of theories and paucity of frameworks for integrating them (Locke & Latham, 1990a) and the fact that various theories involve different levels of analysis and thus deal with different stages of the motivation process (Locke & Henne, 1986). Kanfer (1992) as well as Locke and Henne (1986) have suggested that motivation theories and their associated constructs can be organized in terms of their conceptual proximity to action. Distal constructs or as Locke and Henne named it, general constructs (e.g., needs, values, motives) have only an indirect impact on behavior and performance. Most distal theories of motivation have enjoyed their greatest success in predicting other distal constructs, such as predecision and decision processes and intentions, rather than behavior or performance. Proximal constructs or specific constructs (e.g., intentions, goals, selfregulation) on the other hand focus on motivational constructs at the level of purposive action. Analyses of motivational processes in these theories tend to begin with the individual 's goals rather than with the factors that have shaped the individual 's objectives. Such theories concentrate on the processes and variables that affect the goal-behavior/performance relation.

10

In conclusion, the distinction between distal and proximal constructs and theories is based upon the call for differentiation of the motivational processes underlying choice and volition (e.g., see Ajzen, 1985; Heckhausen & Kuhl, 1985; Gollwitzer, 1990). Moreover, work motivation in the present thesis is defined as willingness to work and the assumption is that actions or behaviors are directed, predicted and explained by volitional processes or will.

2.3.2 Work motivation defined as willingness to work

2.3.2.1 Will/volitional variables as predictors of behaviors
Different approaches to the study of work motivation have been suggested during the years. A work motivation measure was introduced by Sjoberg and Lind (1994), where work motivation was defined as willingness to work and the assumption behind the measure was primarily that will is the process by which behavior is energized, directed, and sustained in organizational setting. To clarify, will is a collective term denoting processes involved in the regulation of actions and regulation simply means controlling an action, that is, causing it to follow the course required to achieve a certain goal (Binswanger, 1991). In previous studies, volitional variables, for example, intentions, have been shown to be more efficient predictors of actions compared to variables like, for example, personality (e.g., Fumham, Jackson, & Miller, 1999; Tett, Jackson, & Rothstein, 1991), and life style (e.g., Sjoberg, 1993), which according to Sjoberg (1998) is a strong argument for a volitional approach to explain and predict action. There is, however, a personality dimension, conscientiousness, that has been shown to be quite strongly related to work-related behaviors, such as job performance (Barrick & Mount, 1991; Mount, Barrick, & Strauss, 1999; Stewart 1999). This personality dimension is associated with a person 's conscience, selfcontrol (Stewart, 1999), organization persistence and achievement orientation

11

(Goldberg, 1993). Moreover, this personality dimension is in conceptualization quite close to the concept of will and volition, which may strengthen the notion of a volitional approach to predict actions. In earlier studies, behavioral intentions (e.g., willingness to work) have been shown to be efficient predictors of behaviors. For example, in the domain of health, intentions have been very effective at predicting a variety of healthpromoting actions, such as dieting (Schifter & Ajzen, 1985) and exercise (Godin, Valois, & Lepage, 1993). Studies regarding participation in an union have shown that willingness to participate is a strong determinant of actual participation (Kuruvilla & Sverke, 1993; Kuruvilla & Fiorito, 1994). As Fishbein and Ajzen (1975) expressed it: "Since much human behavior is under volitional control, most behaviors can be accurately predicted from an appropriate measure of the individual 's intention to perform the behavior in question" (p. 380). But what is will? This question will be discussed in the next part more in depth.

2.3.2.2 The concept of will
Volition or will is an old psychological construct with strong ties to modem philosophy and a range of colloquial meanings (Pervin, 1992). In everyday use, the term denotes willfulness, or dogged perseverance in pursuit of difficult goals. Psychologists have defined volition more precisely as the tendency to maintain focus and effort toward goals despite potential distractions. Another definition was suggested in Reber 's (1985) dictionary of psychology: "the voluntary selection of particular action or choice from many potential actions or choices". This indicates that the definitions can differ somewhat but include a conscious choice of action. Will was a central notion in classical philosophy and psychology up to the early 20 th century. Then it almost disappeared from the late 1950s Psychological

12

Abstracts no longer carried a specific entry for 'will ' or 'volition ' (Nilsson, 1998). The Greek divided the soul into three parts - knowing, feeling, and willing. From the late Antiquity the concept of will became connected with free will, which might be one of the reasons for the disappearance of will within empirical psychology. Secondly, psychologists dissociated themselves from philosophy and stressed the strictly experimental character of psychology. There is now again some room for concepts similar to will and there is an increasing willingness among psychologists to explore phenomena that belong to will (Wameryd, 1998). There is also a growing realization that traditional models of motivation do not explain the diversity of behavior found in organizational settings. Instead, attention to the role of volitional processes in models of motivation would be emphasized (Kanfer, 1990). In industrial and organizational psychology, issues related to volition first appeared in the context of goal-setting research. In 1968, Locke proposed that an individual 's goals served as the immediate regulator of action. Attempts to understand the psychological processes underlying these phenomena, in turn, led to increasing interest in self-regulation, social cognitive, and volitional concepts (Kanfer, 1992). In addition, in attitude research there is a special interest in an attitude component called "intention". "Intentions are assumed to capture the motivational factors that influence a behavior; they are indicators of how hard people are willing to try, of how much of an effort they are planning to exert, in order to perform the behavior. As a general rule, the stronger the intention to engage in a behavior, the more likely should be its performance" (Ajzen, 1991, p. 181). So far it is assumed that individuals consciously and willfully regulate their own behaviors. There are other scholars who have a different view that an individual perceives him/herself to have far more control over his or her everyday behavior than he/she actually has (e.g., Wegner & Wheatley, 1999). The source of
13

behavioral control is said to come not from active awareness but from subtle cues in the environment and from thought processes and information not readily accessible to consciousness (Park, 1999). Further, Bargh and Chartrand (1999) stated that to consciously and willfully regulate one 's own behavior, evaluations, decisions, and emotional states require considerable effort and that it is slow. Regulation of behavior appears to require a limited resource that is quickly used up, so conscious self-regulatory acts can only occur sparingly and for a short time. On the other hand, the non-conscious or automatic processes are u11intended, effortless, very fast, and many of them can operate at any given time. To sum up, the definition of work motivation used here is willingness to work. Previous studies have shown that volitional processes predict and explain action well. The standpoint here is that behavior is primarily willfully regulated.

2.3.3 Beliefs and attitudes
People have attitudes and hold them toward many or n10st other people in their life spaces. They also tend to form attitudes in relation to tangible as well as intangible objects, causes, concepts, acts, and other phenomena with which they are familiar. There are no limits to the attitudes people hold. Attitudes are learned throughout life and are embodied within our socialization process. Some attitudes may be central/stronger to us -a core construct- and may be highly resistant to any change, whereas other, more peripheral attitudes may change with new information or personal experiences (Mullins, 1996). In recent literature on attitudes, Olson and Zanna (1993) claimed that there are a varity of definitions of attitudes and no single commonly accepted definition. Nevertheless, there appear to be three common themes or elements that run through the most common definitions, according to Olson and Zanna (1993). The first element is that attitudes generally involve an evaluative component. A second common component of definitions of attitudes is that they are
14

"represented in memory". The third common element of definitions and conceptualizations of attitudes is that they entail cognitive, affective, and behavior components. A straightforward definition of attitude that is consistent with the thoughts of Olson and Zanna (1993), has been offered by Eagly and Chaiken (1993) "Attitude is a psychological tendency that is expressed by evaluating a particular entity with some degree of favor or disfavor...psychological tendency refers to a state that is internal to the person, and evaluating refers to all classes of evaluating responding, whether overt or covert, cognitive, affective, or behavioral" (p. 1). The nature of the relationship between attitudes and behavior has been departing over the years. It seems to make intuitive sense to many people that attitudes are major causes of behavior. However, the connection between attitudes and behavior has shown to be weak or unpredictable (for a review, see Andrich & Styles, 1998). A theory that has discussed the connection between attitudes and behavior is the theory of reasoned action (Fishbein & Ajzen, 1975). Fishbein and Ajzen have suggested that one useful way to conceptualize the notion of job attitude is to subdivide it into three related parts: (1) beliefs about one 's job, (2) the attitude itself, and (3) the behavioral intentions that result from the attitude. The process is presented in Figure 1.

Beliefs about ones job
1. Job is dull 2. Job is dirty 3. Job provides little autonomy, responsibility, etc

Job attitudes 1. Job dissatisfaction 2. 'Low commitment

Behavior intentions
1. Intention to leave 2.Intention to reduce effort

Actual behavior 1. Absenteeism 2. Turnover 3. Poor perfonnance

Figure 1. A conceptual model of attitudes (after Fishbein, 1967). Adapted from Steers and Porter (1983, p. 330).

15

As shown in Figure 1. beliefs about one 's job (e.g., this job is dull) lead to negative job attitudes (e.g., job dissatisfaction) which, in tum, lead to the behavioral intentions to leave or reduce effort on the job. These behavioral intentions are then translated into actual behavior, assuming the individual is able to carry out his or her intentions. The distinction between beliefs and attitudes should be clarified. To begin with, beliefs, which are concerned with what is known about the world; they center on what 'is ', on reality, as it is understood. As suggested earlier: "An attitude represents a person 's general feeling of favorableness or unfavorableness toward some stimulus object... as a person forms beliefs about an object, he automatically and simultaneously acquires an attitude toward that object. Each belief links the object to some attribute; the person 's attitude toward the object is a function of his evaluations of these attributes" (Fishbein & Ajzen, 1975, p. 216). The general attitude a person has toward an object is seen as an aggregation of all the beliefs that she holds about it, each weighted by the positive or negative evaluation she places on the various beliefs. 'Therefore, two employees may have the same set of beliefs about a job (e.g., it 's of low status) but hold different attitudes toward it because one of them prefers routine work whereas the other desires more uncertainty (Pinder, 1998). It is often discussed how stable attitudes are and the general view about the stability of attitudes and beliefs is that they are stable over time, are resistant to change, and exert strong influence on information processing, overt behavior and formation (e.g., Prislin, 1996). As stated above, the strength of attitudes varies and the features of strong attitudes are related to a number of attitudinal qualities. Thus, the persistence of an attitude over time is positively associated with the amount of experience with the attitude object (Doll & Ajzen, 1992), the certainty with which the attitude is held (Pelham, 1991), the importance of the attitude (Kronsnick, 1988), the internal consistency of the attitude (Norman, 1975), and the affective extremity of the attitude (Schuman & Presser, 1981). Furthermore, it might be quite difficult to change existing work attitudes. However, Pinder (1998) wrote that one way to influence a person 's attitudes

16

might be to introduce new information about the job that links the job with attributes that the employee evaluates as positive (such as its variety and status). Alternatively, one could attempt to change the employee 's assessment of the desirability or undesirability of the attributes that the employee associates with a job (Pinder, 1998). This requires the manager to be well informed about the employee 's preferences, because attitudes and beliefs differ among individuals. Lastly, two of the most widely studied work attitudes are job satisfaction and organizational commitment. These constructs are frequently used as measures of work motivation (Maehr & Braskamp, 1986; Schou, 1991). These job-related attitudes are discussed further in next section.

2.3.3.1 Job satisfaction
Job satisfaction is, without any doubt, the job-related attitudinal construct that has received most attention in modem times (Pinder, 1984; O 'Conner, Peters, & Gordon, 1978). A literature review by Locke (1976) estimated that over 3,300 projects had been conducted and reported on job satisfaction during the previous 25 years. Further, Cranny, Smith, and Stone (1992) suggested that more than 5,000 studies of job satisfaction had been published. Although a large body of research on job satisfaction has been accumulated over the years, there are still major shortcomings in job satisfaction research. One of them is the loose coupling between its theories and its measurement. This is a paradoxical situation. While job satisfaction is one of the n10st frequently studied areas in industrial and organizational psychology, it is also one of the n10st theory-free concepts in the field (Bussing, 1992). As a result of many decades of effort by social scientists, there appears to be a high level of agreement among them on the meaning of the construct of job satisfaction. Most often, job satisfaction is conceptualized as a general attitude toward an object, the job. For example, Porter, Lawler, and Hackman (1975, p. 53-54) as well as Locke and Henne (1986, p. 21) have defined job satisfaction

17

similarly and these definitions are consistent with the general construct stated above. The definition given by these authors are similar to the one offered by Locke (1976, p. 1300) who defined job satisfaction as "a pleasurable or positive emotional state, resulting from the appraisal of one 's job experience". Several articles have discussed whether a single-iten1 measure is adequate or not to measure overall job satisfaction. The conclusion which has been drawn that in most studies is that a single-item measure of overall job satisfaction is acceptable (e.g., see, Wanous, Reichers, & Hudy, 1997) but a single-iten1 measure tends to paint a rosier picture of job satisfaction than the impression conveyed from the multiple-item measure would justify (e.g., see Oshagbemi, 1999). However, in more recent work on the concept job satisfaction, multidimensionality is emphasized (Buckley, Carraher, & Cote, 1992). Moreover, Taber, and Alliger (1995) found that task-level measurement assessed different psychological processes than those assessed by traditional global and facet measures. They suggested that by complementing traditional, global measurement procedures, task-level assessment might facilitate new research into the nature of job satisfaction. Empirical data indicate that global measures of job satisfaction are not equivalent to the sum of multi-dimensional job satisfaction (Scarpello & Campbell, 1983). Job satisfaction has regularly been used as a measurement of work motivation, but it has turned out to be a quite poor assessment of work motivation. Firstly, job satisfaction is by definition not the same as motivation. Job satisfaction can be described n10re as an attitude, al1 internal state, a measurement of how satisfied or dissatisfied one is with one 's job, if one likes it or dislikes it in different aspects. It is possible to like a job because it makes low demands on one 's effort, or because one is able to work hard to complete a task 011e dislikes (Mullins, 1996). Secondly, Locke and Latham (1990a) stated that motivation and satisfaction are independent outcomes and different theories. Finally, previous studies have only shown weak relationships to different work-related behaviors. Iaffaldano and Muchinsky (1985) found in their meta-analysis that the relationship between job performance and job satisfaction was weak.
18

Another work-related behavior that has been related to job satisfaction is, for example, absenteeism, and only a weak correlation, approximately -0.2, between job satisfaction and absenteeism has been reported (Hackett, 1989; Hackett & Guion, 1985). Moreover, Tharenou (1993) found that non-legitimate absenteeism was more strongly correlated to job satisfaction than legitimate absenteeism. However, Haccoun and Desgent (1993) reported that most absenteeism was legitimate and only 20% reported no legitimate reasons for absenteeism. Intention to quit has shown to be more strongly related to job satisfaction than the actual turnover behavior (e.g., Hellman, 1997; Tett & Meyer, 1993). Even though job satisfaction might not be considered the best measurement of work n10tivation, it is still interesting to measure job satisfaction in order to shed some light on its relations to different factors. Empirical research has investigated numerous variables in an attempt to determine how job satisfaction is created and how it affects other work outcomes. It is not only external components that affect job satisfaction, it has been shown that approximately 30% of the observed variance of job satisfaction was due to genetic factors (Arvey, Abraham, Bouchard, & Segal, 1989). Further, research findings have shown that job satisfaction is stable over time (Staw & Ross, 1985; Steel & Rentsch, 1997). Personality has also been shown to affect job satisfaction. Staw, Bell, and Clausen (1986) discovered a link between childhood personality and job satisfaction later in life, and there has been considerable interest in the relationship between individual dispositions and job satisfaction. Weiss and Cropanzano (1996) suggested that affective temperament may influence the experience of emotionally significant events at work, which in tum influence job satisfaction. Furthermore, in several studies it has been demonstrated that an individual 's core self-evaluation was linked to job satisfaction (Judge, Locke, Durham, & Klunger 1998; Judge, Bono, & Locke, 2000). Judge, Locke, and Durham (1997) defined core self-evaluation as fundamental assessments that individuals make about themselves and their self-

19

worth: self-esteem, generalized self-efficacy, locus of control, and low neuroticism. Watson and Keltner Slack (1993) indicated that job satisfaction can be viewed in the context of more general emotional lives of employees. A fairly high positive relation between overall job satisfaction and life satisfaction was reported by Adams, King, and King (1996). Judge and Watanabe (1993) found that the relationship between job satisfaction and overall life satisfaction was fairly strong in cross-sectional results, but longitudinal results suggested a weaker relationship over a 5-year period, particularly with respect to the effect of job satisfaction on life satisfaction. There are additional aspects of job satisfaction that have been studied. Jabri (1992) found a moderate correlation between job satisfaction and appropriate task allocation. Melamed, Ben-Avi, Luz, and Green (1995) showed that job satisfaction was mainly related to subjective monotony, whereas absence due to sickness was equally related to work conditions and subjective monotony. Communication with the manager was reported to be an important predictor of job satisfaction (Callan, 1993; Miles, Patrick, & King, 1996). Miles, Patrick, and King (1996) also found that supervisors reported a higher level of job satisfaction than other employees. Introduction of new technology can increase the stress level of employees, which, in tum, can decrease job satisfaction (Aiello & Svec, 1993). Job content and participation significantly influenced job satisfaction after the introduction of new technology (Korunka, Weiss, Huemer, & Karetta, 1995). Kahn and Robertson (1992) found that training and experience had no effect on job satisfaction and work motivation when using computers. Blau (1999) reported that task responsibilities and employee performance appraisal satisfaction significantly affected overall job satisfaction. Further, Blau found that the supervisor 's role in affecting employee job satisfaction is important.

20

Because job satisfaction is one of the most studied constructs, several factors have been associated to it and only a few of them have been reported in the review above. The second job attitude frequently examined is organizational commitment. Like job satisfaction, organizational commitment is often used as a work motivation measure. Organizational commitment is discussed more in detail in the next section.

2.3.3.2 Organizational commitment
Like job satisfaction, organizational commitment has received substantial attention in past research. Several organizational commitment definitions have been offered over the years. They appear, however, to reflect at least three general themes: affective attachment to the organization, perceived costs associated with leaving the organization, and the obligation to remain with the organization (Meyer & Allen, 1991). Organizational commitment is often employed as an assessment of work motivation (Maehr & Braskan1p, 1986; Schou, 1991). However, it should be notified that organizational commitment is not by definition the same as work motivation, in fact, it is possible to be quite committed to an organization without being highly motivated to work. Also, the relationship between organizational comn1itment and different workrelated behaviors has been reported to be weak or inco11sistent. Baugh and Roberts (1994), Mowday, Porter, and Dubin (1974), as well as Ward and Davis (1995), concluded that organizational commitment and job performance are positively correlated. Brett, Cron, and Slocum (1995) examined the relationship between organizational commitment and performance using economic dependency on work as a moderator. They reported that there were stronger relationships between organizational commitment and performance for those with low financial requirements than for those with high requirements. Kalleberg and Marsden (1995) discovered a modest relationship between
21

organizational commitment and job performance, while Leong, Randell, and Cote (1994), Mathieu and Zajac, (1990) and Tett and Meyer (1993) found only a weak correlation between these two variables. On the other hand, Wright (1997) found a negative association between measures of organizational commitment and job performance. Recently, Benkhoff (1997) investigated the link between employee commitment and organizational performance in terms of sales targets met and changes in profits. The result indicated a link between commitment and performance. Generally, it is believed that employees who are more strongly committed to the organization avoid withdrawal behaviors, such as being absent (Blau & Boal, 1989). However, little evidence shows that a meaningful and consistent attitudeabsenteeism relation does exist (Hackett, 1989; Hackett & Guion, 1985; Mayer & Schoorman, 1992; Randell, 1990). Sagie (1998) divided absenteeism into voluntary absence and involuntary absence and the result showed that organizational commitn1ent was strongly related to voluntary absence, but not to involuntary absence. Another withdrawal behavior is turnover, and a moderately strong relationship has been found between commitment and tun10ver (Hulin, 1991). A multidimensional scale to measure commitment has been suggested that involves the din1ension of affective commitment (affective reaction to an organization), continuance commitment (a commitment to continue one 's task), and finally a normative component (the employee 's feelings of obligation to remain with the organization) (Allen & Meyer, 1990). These three dimensions have been supported in several studies (Cohen, 1996; Durman, Grube, & Castaneda, 1994; Irving, Coleman, & Cooper, 1997; Ko, Price, & Mueller, 1997). Distinctions between different types of commitment have, however, only given an insignificant improvement in the prediction of performance (Mayer & Schoorman, 1992). Meyer, Paunonen, Gellatly, Goffin, and Jackson (1989) examined the relationship of job performance with affective commitment and continuance

22

commitment and concluded that affective commitment, which refers to identification with, involvement in, and emotional attachment to the organization, is correlated positively and continuance, which refers to commitment based on employees ' recognition of the costs associated with leaving the organization, is correlated negatively with all measures of performance. The dimension of affective commitment was found to be the only dimension of commitment that was related to turnover and to absenteeism (Somers, 1995). The results of a longitudinal study found that the correlation between job satisfaction and the dimension of organizational continual1ce commitment was quite high (Cramer, 1996). Jaros (1997) found that affective commitment had a significantly stronger correlation with turnover intentions than normative and continuance commitment. Eby, Freeman, Rush, and Lance (1999) reported that affective commitment was strongly related to turnover behavior as well as absenteeism. Organizational commitment is strongly affected by organizational factors including leadership, culture, values, and norms (Cohen, 1992; Gellatly, 1995; Markham & Mckee, 1995). Steers (1977) found that organizational commitment was influenced by need for achievement, group attitudes toward the organization, education (inversely), organizational dependability, perceived personal importance to the organization, and task identity. Steers stated that a common theme in these variables is the notion of exchange among individuals. Further, individuals come to organizations with certain needs, desires, skills, and so forth, and expect to find a work environment where they can utilize their abilities and satisfy many of their basic needs. When the organization provides such a vehicle, the likelihood of increasing commitment is apparently enhanced. When the organization is not dependable, however, or when it fails to provide employees which challenging and meaningful tasks, the commitment level tends to diminish. In a study by Nyhan (1999) it was found that to increase organizational commitment managers must be willil1g to build on their investments in the people in their organizations, by letting the employees participate in decision making, and allowing feedback from and to employees. The perception that one is competent also affects the degree of attachment to an
23

organization. Benkoff (1997) found that regular training at work influenced commitment to the organization and a person 's perception of hislher competence.

2.3.4 Incentives and rewards theory (extrinsic motivation)
Extrinsic motivators, such as incentives, have long been thought, both theoretically and practically, to affect employee motivation and performance (e.g., see Skinner, 1938; Hull, 1943; Taylor, 1947). Incentives are not only thought to enhance motivation and performance but also to shape the employment relationship. Simply, pay has been seen as the means for attracting, retaining, and motivating employees primarily because of its near universal appeal and the ease, directness, and certaintly with which it can be delivered (Greller & Parsons, 1995). Pay systems or compensation systems, philosophies, and practices act together with other elements of the organization environment influencing employees ' attitudes, behavior, and how they define their relationship with the organization (Rousseau & Greller, 1994). The relationship between employees and compensation plays different interrelated roles. First, pay systems make a statement regarding what is important. Some systems value technical work, other value customer service. Some recognize individual performance, some value performance in terms of group products (Murphy & Cleveland, 1991). Second, pay provides reinforcement. People like receiving pay. They work in directions that produce pay, if these directions are made clear to them (Lawler, 1971). Third, pay provides feedback on performance (Herold & Parsons, 1985; Greller & Parsons, 1995). It is quite clear in the literature that both financial and non-financial incentives can increase performance (Lawner, 1990; Milkovich & Newman, 1993), especially when the incentive system is properly designed (Guzzo, Jette, & Katzell, 1985). Financial incentives also convey symbolic meaning (e.g.,
24

recognition, status) beyond their monetary value: They meet multiple human needs and serve multiple functions (Steers, Porter, & Bigley, 1996). Financial incentives supplement intrinsic rewards. People need money (Steers et aI, 1996). Moreover, in general, financial incentives have a positive relationship to work motivation (Locke, Feren, McCaleb, Shaw, & Denny, 1980). However, a meta-analytic review presented by Jenkins, Mitra, Gupta, and Shaw (1998) found that financial incentives were not related to performance quality but had a correlation of .34 with performance quantity. Five primary theoretical frameworks address the relationship between money and performance. Expectancy theory suggests that tying financial incentives to performance increases extrinsic motivation to extend effort and consequently performance (Lawer, 1971, 1973; Vroom, 1964). Reinforcement theory, although less cognitive in focus, argues that tying money to perfornlance will reinforce performance (Komaki, Coombs, & Schepman, 1996). Goal setting theory indicates that financial incentives increase acceptance of difficult performance goals, and enhance performance (Locke, Latham, & Erez, 1988). By contrast, cognitive evaluation theory proposes that performance-contingent financial incentives erode intrinsic motivation and hereby diminish task performance (Deci & Ryan, 1985). Equity theory (e.g., Adams, 1963) argues that people are motivated to reduce inequity, but makes no specific predictions regarding the relationship between financial incentives and performance per se, although under certain conditions, deviations from fairness may erode the association of financial incentives to perfornlance (e.g., Kanfer, 1990). But money alone is not always enough to motivate high performance. Consequently, all researchers do not accept the role of money as one of the most salient motivational factors and overall, both theory and practice highlight the ambiguity ill how financial incentives affect performance. Kohn (1993) has published warnings in the popular literature about the harmful effects of rewards on employee performance. He states that "rewards typically undermine the very processes they are intended to enhance" (Kohn, 1993, p. 54). He claims that
25

rewards used in work organizations, such as stock options, pension plans, sale commissions, bonuses, and vacations generally result only in "temporary compliance". Perhaps the most significant argument against financial incentives concerns the detrimental effects of money on intrinsic motivation (Eisenberg & Cameron, 1996; Kohn, 1993). Opponents argue that financial incentives control employee behavior externally, reducing self-determination and intrinsic motivation (Deci
& Ryan, 1985). They also jeopardize the relationship between supervisors and

subordinates (Meyer, 1975). Opponents argue further that money is not a motivator (Kohn, 1993); financial incentives may reduce job dissatisfaction, but they do not motivate (Hetzberg, 1968). Further, Cavanagh (1992) found that the level of salary was not a significant factor to determine job satisfaction. Holzer (1990) discovered that experience and knowledge had more influence on productivity than earnings. Like Holzer, Guest (1990) could not find a strong relationship between productivity and salary in an analysis of British productivity in the 80 's. People 's attitudes toward money may have significant impacts on their perceptions of work-related tasks, the reward system, and their intrinsic motivation to a task. This in tum may influence their work-related behavior, task performance, job satisfaction and morale, and the effectiveness of the organization (Tang, 1995). People 's attitudes toward money are consistent with their inner values, their frame- of reference, their culture, and their experience in society. Tang (1995) reported that those people who value money do not necessarily have a higher income than those who do not, and they therefore tend to have a higher level of pay dissatisfaction. Pay satisfaction has also shown to influence work related behaviors. One of the most well-known models of pay satisfaction was proposed by Lawler (1981) and later modified by Heneman (1985). Both of these models are based on the concept that discrepancies in employee 's perception of amount that should be received and perception of amount that is received are major determinants of
26

pay satisfaction. One of the differences between the two models is that Heneman 's (1985) model conceptualizes pay satisfaction in multi-dimensional terms. Heneman (1985) argued that individuals develop a general attitude to their compensation, as well as more specific attitudes towards each of several component parts of their compensation, including (a) salary level (external competitiveness), (b) pay structure (internal consistency), (c) individual salary (employee 's contribution) and (d) the administration of the entire pay system. The four dimensions have been supported (Judge, 1993; Judge & Welbourne, 1994). Heneman (1985) also stated that pay satisfaction can be expected to be related to such behaviors as turnover, absenteeism, and union activity. Lum, Kervin, Clark, Reid, and Sirola (1998) supported the finding that pay satisfaction has an effect on turnover intention. In addition, Aryee (1999) showed that the relationship between pay satisfaction and life satisfaction was significant and positive. How and if incentives will affect motivation and job performance among employees might depend on, for example, cultures, position in the organization, and other factors. According to Persson (1994), Swedish companies do not consider payment as a motivational aspect, partly because the level of salary is often decided outside the organization. An employee 's level of payment is often due to position only, not to performance. Pfeffer and Langton (1993) found that the greater the degree of wage dispersion within academic departments, the lower individual faculty members ' satisfaction and research productivity are and the less likely it is that faculty n1errlbers will collaborate on research. In a study by Levine (1993) both employees in the USA and Japan were studied. The result showed that workers receiving high wages (compared with an average worker) were less likely to quit the job and are more satisfied with their pay. To conclude, there are few systematic examinations of the relationship of financial incentives to employee behaviors, and little effort is devoted to assessing this relationship in the cumulative scientific knowledge base. This is perhaps because studies of the impact of financial incentives are complex. They
27

must address the different meanings, literal and symbolic, of money, and other outcomes (e.g., promotions, coworkers resentment, turnover) that might deliberately or inadvertently be associated with financial incentives, differences in utility of money, social comparison that financial incentives evoke, group norms, organizational structure, and so on. An estimate of the overall association of financial incentives with performance is necessarily affected by a multitude of factors. Such factors are attitude toward money, cultural differences, and the tax system.

2.3.5 Intrinsic motivation
A more current trend in work motivation views behavior as being intrinsically motivated. Intrinsically motivated behavior can be defined, loosely, as behavior that is performed for its own sake rather than for the purpose of acquiring any material or social rewards. Intrinsic motivation, or engaging in a task for its enjoyment value, is one of the most powerful forms of motivation according to Deci and Ryan (1987). Further, it is associated with enhanced performance, improved conceptual and creative thinking, superior memory recall, positive affect, subsequent willingness to engage in other tasks, and better psychological and physical health compared with other forms of motivation (Deci & Ryan, 1992). One theoretical guide for much of that work has been self-determination theory. A central point of self-determination theory is that people work, not only for extrinsic rewards, but also to fulfill psychological needs such as autonomy, competence, and relatedness. Moreover, if those three psychological needs are fulfilled, for example in the workplace, this will lead to greater satisfaction, enhanced performance, and general well-being according to the theory (e.g., Deci & Ryan, 1985; 1987; Ryan & Deci, 2000). These findings have been supported in other studies as well (e.g., Ilardi, Leone, Kasser, & Ryan, 1993; Lu, 1999). Further, although feelings of competence and interest in the task are central to intrinsic motivation, a person must also feel free of pressure, for
28

example from rewards or potential punishment. The person must feel that her "locus of causality" is internal, meaning that she is responsible for the choice of the activity, that she is in command of how she is spending her time. Hence, the notion of choice is central to the concept of self-determination. The person must be in control of the alternatives for action and be able to choose among them (Pinder, 1998). One question extensively studied is the relation between intrinsic and extrinsic motivation or more specifically, how extrinsic rewards affect intrinsic motivation. According to Deci (1975) the detrimental effects of extrinsic events on intrinsic motivation depend on the perceived salience of "controlling" versus "informational" properties of the extrinsic event. According to Deci 's theory, events that are interpreted by the individual as informational should facilitate intrinsic task interest, particularly when the information provided conveys a sense of personal competence. Several studies testing this aspect of Deci 's theory provide further empirical evidence indicating that a variety of extrinsic organizational events, such as rewards, goal-setting, feedback, and modeling may also affect task interest, enjoyment, and behavior (Collar & Barrett, 1987; Harackiewicz, Sansome, & Manderlink, 1985). Moreover, Deci, Koestner, and Ryan (1999) found in their n1eta-analysis of 128 studies that engagement-contingent, completion-contingent, and performancecontingent rewards significantly undermined free-choice intrinsic 1110tivation, as did rewards, all tangible 'rewards, and all expected rewards. Positive feedback on the other hand enhanced both free-choice behavior and self-reported interest. To further support this, Wild, Enzle, Nix, and Deci (1997) found that participants in a study who read about an extrinsically motivated target expected that task engagement would be less enjoyable and associated with less positive affect and that there would be poorer quality of interpersonal relations, compared with participants reading about intrinsically motivated targets. Participants in a second study who were taught a skill by an extrinsically motivated (paid) target reported lower interest in learning and lower task enjoyment than those taught

29

by intrinsically motivated (volunteer) target, despite receiving identical lessons and learning to the same criterion level. However, other meta-analyses have not reached the same conclusion. One n1etaanalysis reported that, overall, a reward does not decrease intrinsic motivation. When interaction effects are examined, findings show that verbal praise produces an increase in intrinsic motivation. The only negative effect appears when expected tangible rewards are given to individuals simply for doing a task (Cameron & Pierce, 1994; Eisenberger, Pierce, & Cameron, 1999). According to Wiersma (1991, 1992) the effects of intrinsic and extrinsic rewards on motivation are additive. In a study by Reeve and Deci (1996) they showed that winning (relative to losing) increased intrinsic motivation by enhancing perceived competence and a pressured (relative to non pressured) interpersonal context decreased intrinsic motivation by diminishing perceived selfdetermination. To conclude, studies have shown that people who are intrinsically motivated, enhance their performance and increase job satisfaction. The view presented here is that opportunities for self-determination are regarded as a sign of the psychological well-being of individuals. This view is not shared by all researchers. Schwartz (2000) argues instead that freedom, autonomy, and selfdetermination can become excessive, and when that happens, freedom can be experienced as a kind of tyranny. Furthermore, in the literature it is argued either for extrinsic motivation or intrinsic motivation. Frey (1997) suggested a more balanced approach to understand the concept of work in firms. Frey stated that both intrinsic and extrinsic motivation must be taken into account.

2.3.6 Goal and goal setting
Goals are widely recognized as being central to the understanding of motivated behavior, with different research disciplines emphasizing different levels and types of goals and their consequences. Personality research has focused on the
30

nature of goals or higher level personal striving (e.g., Carver & Scheier, 1981; Emmons, 1989), whereas Locke and Latham (1990b) focused on middle-level task goals; cognitive researchers have emphasized even lower level goals such as script concepts (cf. Austin & Vancouver, 1996). Austin and Vancouver (1996) stated that few researchers integrate multiple levels into a single theoretical approach. Austin and Vancouver (1996) stated after reviewing goal constructs in psychology that "the use of goals to understand behaviors, ranging from the movement of a hand to life tasks to the way one sees the world, may leave some a bit queasy. Yet, in all cases, the goals serve as a standard with which perception of current or anticipated states are compared, which in tum affects son1e processes (p. 361). A definition of goal presented by Austin and Vancouver (1996, p. 338 ) said that a goal can be described "as internal representations of desired states, where states are broadly construed as outcomes, events, or processes. Internally represented desired states range from biological set points for internal processes (e.g., body temperatllre) to complex cognitive depictions of desired outcome (e.g., career success)". Goal setting processes have come to occupy a central role in current theories of applied motivation (Kanfer, 1990; Locke & Latham, 1984, 1990b. Goal attributes, such as specificity and difficulty, have consistently been linked to the effectiveness of goal setting based primarily on social cognitive theory (Bandura, 1986).

r

Goal difficulty

1

I

Values

I

Emotions and desires

Responses or actions Work behavior and performance

_ _ Levelof commitment

Figure 2. An illustration of Locke 's theory of goal-setting (adapted from
Mullins, 1996, p. 511).
31

The most fundamental belief of Locke and Latham 's (1984, 1990b) goal setting theory is that goals are responsible for human behavior. Locke and Latham accept the importance of perceived value, and suggest that these values create the experience of emotions and desires. People strive to achieve goals to satisfy their emotions and desires. Goals guide people 's responses and actions. Goals direct work behavior and performance, and lead to certain consequences or feedback (Locke & Latham, 1984; 1990b). Locke and Latham 's theory of goal setting is illustrated in Figure 2. Furthermore, a goal is something that a person tries to attain, achieve, or accomplish. In work settings, goals may take the form of constructs such as a level of job performance, a quota, a work norm, a deadline, or even a budgetary spending limit (Locke, Shaw, Saari, & Latham, 1981). Multiple reviews and meta-analyses of goal-setting literature have concluded that there is substantial support for the basic principle of goal setting theory (Locke & Latham, 1990a, 1990b; Wofford, Goodwin, & Premack, 1992). First, specific and difficult goals consistently lead to better performance than specific and easy goals, general goals such as "do your best", or no goals (Brown & Latham, 2000; Locke, 1996; Tubbs, 1986; Vance & Colella, 1990). The effectiveness of difficult goals is predicted on the assumption that commitment to such a goal is high (Locke, Latham, & Erez, 1988). As stated by Locke, Latham, and Erez (1988), "It is virtually axiomatic that if there is no commitment to goals, then goal setting does not work" (p. 23). Goal commitment is proposed to be a moderator of the relationship between performance and task performance. Higher levels of goal commitment lead to a stronger relationship between perfom1ance goals and subsequent performance. However, a meta-analytic review by Donovan and Radosevich (1998) found that the moderating effect of goal commitment on goal-performance accounted for only 3% of the variance in performance. Twenty years of research on the effects of goal-setting on performance have led several reviewers (Locke, et aI., 1981; Mento, Steel, & Karren, 1987; Tubbs, 1986) to conclude that the effects of goals on performance are an10ng the most robust of any theory to be found in the motivation literature. Pinder (1984)

32

observed in his own review of all the major motivation theories that "goal setting theory has demonstrated more scientific validity to date than any other theory or approach to work motivation ....." (p. 169). Goal setting increases performance motivationally through its effect on one 's intentions regarding effort and persistence; it increases performance cognitively by directing attention/intention to discover strategies that will lead to goal attainment (Earley, Conolly, & Ekegren, 1989; Latham, Winters, & Locke, 1994). However, Kanfer and Akerman (1989) concluded that goal setting is an effective motivational technique only when the first two of three stages of learning have taken place: declarative and knowledge compilation. It has been found that setting specific goals on tasks where the person has yet to acquire the requisite ability to perform well has a negative effect on performance.

There are different ways goals can be set. The goals may be self-set or set by others. However, concerning assigned goals, Wright, Hollenbeck, Wolf, and McMahan (1995) reported that how assigned goals are derived and how they are communicated to subjects affects goal setting outcomes. Additionally, Gollwitzer (1999) suggests that goals and resolutions stal1d a better chance of being realized when they are furnished with implementation intentions that link anticipated suitable opportunities to intended goal directed behavior. Implen1entation intentions delegate the control of goal-directed behaviors to specified anticipated environmental stimuli. People are frequently confronted with situations where they cannot rely on their habits and automatically activated goals. This is when action control through the formation of implementation intentions is most valuable (Gollwitzer, 1999). Sheldon a11d Elliot (1998) showed that autonomous goals, which are undertaken with a sense of full willingness and choice, are better attained than controlling goals, which are felt to be con1pelled by internal or external forces or pressure. In conclusion, although goal setting theory is known as valid and useful, there are some criticisms to it. Deci (1992) stated that goal setting theory is relatively mute on the issue of human nature. Deci claimed that the idea of goals, as efficient causes of behavior, represents a type of explanation that has a machine
33

metaphor flavor, and goal mechanisms are still mechanisms regardless of whether they are cognitions or associations. He argued that a comprehensive theory of motivation requires more than the concept of goals.

2.3.7 Perceived control
The concept of employee control has had a long tradition in organizational behavior theory and research (Ganster & Fusilier, 1989), being central to both participation in decision making literature (Locke & Schweiger, 1979) and job design literature (Hackman & Oldham, 1976). More recently, the construct has been the focus of research in the occupational stress area (Karasek, 1979; Karasek & Theorell, 1990). Moreover, personal control over one 's work environment is an important theme in many branches of social science. Personal control can be defined as an individual 's belief ill his or her ability to effect a change in a desired direction (Greenberger & Strasser, 1986), which in the work context reflects the extent to which employees perceive that they have the opportunity to adopt behavioral efforts to control the quality, occurrence, and duration of significant work-related events (Jimmieson & Terry, 1998). Typically, control at work is made available by providing employees with a range of different control options, including choice of work tasks, methods of work, work pacing, work scheduling, control over resources, and control over the physical environment (see Ganster, 1988). Baltes and Baltes (1992) claimed that, in work situations, personal control beliefs have consequences for job performance, largely because of their positive or negative effects on work motivation. Specifically, Baltes and Baltes argued that, because the attainment of work goals usually requires events to be changed to suit the goal seeker to some extent, low personal control situations at work upset and distract highly motivated employees much more than poorly motivated employees. Employees who are not highly motivated are more willing to accept the consequences of not being able to control events to their liking
34

because they are less concerned about reaching their work goals. Because being distracted and upset impairs performance in most jobs (Weiner, 1979), personal control should moderate the relations between work motivation and job performance. Orpen (1994) found that personal control moderated the effects of work motivation on job satisfaction and performance, with highly motivated employees being more adversely affected by low personal control. Dwyer and Ganster (1991) found that perceived workload and control interacted significantly to explain variance in satisfaction with work itself. Control over one 's work, including job autonomy and non-routine work, is positively associated with job satisfaction (Ross & Reskin, 1992). The results presented by Jimmiesson and Terry (1998) indicated that a high level of both objective and subjective work control has a positive impact on level of task satisfaction. Similar findings were presented by Sargent and Terry (1998) who presented a positive relationship between perceived control and job satisfaction as well as psychological well-being (Daniels & G-uppy, 1992). Finally, Ashforth and Saks (2000) found that there were two distinct responses to perceived personal control. The first implied a proactive orientation where control begets control and the second implied a reactive orientation where unmet expectation prompts a sense of futility and withdrawal. To conclude, it has been shown that it is important for employees, or for all people for that matter, to perceive themselves to have control over a situation. In fact, perceived control and generalized expectancies of control may be important cognitive mediators of actual control (Frese, 1989). Indeed, laboratory research suggests that a belief in control may be more important than actual control (Ganster & Fusilier, 1989).

35

2.3.8 Factors related to work motivation
2.3.8.1 Occupational and organizational stress (work stress)
Stress has become one of the most serious health issues of the twentieth centurya problem not just for individuals in terms of physical and mental disabilities, but also for employers and governments who have started to assess its financial damage. Matteson and Ivancevich (1987) estimate that stress causes half of absenteeism, 49% of turnover, and 5% of total lost productivity due to preventable occupational stress. Occupational stress has serious consequences for both individuals, employees and organizations. The term stress, meaning hardship or adversity, can be found - though without a programmatic focus - at least as early as in the 14 th century, (for further historical outlooks of stress see Khan & Byosier, 1992; Lazarus, 1993). Organizational stress or job stress has many definitions, but research under that label is usually concerned with the negative effects of the workplace ' environment, sometimes in conjunction with the employee 's own characteristics, on an employee 's health and well-being (Beehr, 1995). Much of the impetus for this interest in occupational stress has grown from the early work of Karasek and his colleagues (e.g., Karasek, 1979; Karasek, Marxer, Ahlborn & Theorell, 1981; Karasek & Theorell, 1990). Karasek (1979) hypothesized that job demand (e.g. high workload) was not in itself harmful, but when combined with low employee control, demand could lead to the development of cardiovascular diseases. When both job demands and control are high, Karasek describes the job as 'active ', that is, one in which the demands act as sources of challenge, rather than sources of mental and physical stress, which can also lead to the individual experiencing an increased motivation to perform. The model has been supported to some extent by some studies (Daniels & Guppy, 1992; Ross & Reskin, 1992), whereas others have shown no support for it (Fletcher & Jones, 1993). A prediction ofKaresek 's Job-demand model is that

36

motivation and job satisfaction will occur in situations where both job demands and worker 's control are high, which was supported by Jonge, Breukelen, Landeweerd, and Nijhuis (1999). In a Swedish study, it was found that both women and men having no control over their work situation, together with having monotonous work tasks, had a higher stress level. The highest perceived stress level, both in women and men, was found in female dominated professions (Knutsson & Nilsson, 1994). Many other theories explaining work stress have been suggested and contemporary approaches to occupational stress use a psychological model which views work stress as part of a dynamic il1teraction between the worker and the work environment. Two related theories can be found in this approach: interactional theories and transactional theories. Interactional theories emphasize the importance of the "fit" between a person and their work. According to French, Caplan, and Van Harrison (1982), work stress results primarily from an incompatible person-environment fit, producil1g psychological strain and stress-related physical disorders. Stress can best be looked at as a particular relationship between an individual and a specific environment, as a process taking place in a context. The other approach, transactional theories, focus on the cognitive processes and emotional reactions that underpil1 the interactions. A transactional approach to stress was suggested by Lazarus (1991), who argued that traditional approaches to work stress are not useful at the individual and group level. Stress can best be looked at as a particular relationship between an individual and a specific environment, and as a process taking place in a context. Stress also depends on how an individual appraises what is happening. However, stress has many effects. Generally, stress has been shown to lower job satisfaction (Kahn & Byosiere, 1992). Chaplain (1995) found a negative relation between job satisfaction and stress in a study of English primary school teachers. Further, high occupational stress reports were related to low level of

37

job satisfaction. Lu (1999) found in a study of employees in Taiwall that work n10tivation, especially intrinsic motivation, was positively associated with wellbeing. Fogarty Machin, Albion, Sutherland, Lalor, and Revitt (1999) examined the relationship between stress, strain and job satisfaction in two studies and found that stress and strain accounted for 29% of the variance in the first study and 35% in the second study. Begley and Czajka (1993), fOUIld, however, that stress increased job displeasure only when commitment was low. Spielberger and Reheiser (1994) suggested a work stress nleasurement called the Job Stress Survey (JSS), which was designed to assess the perceived intensity and frequency of occurrence of working conditions that are likely to adversely affect the psychological well-being of the employees. In a study of white-collar employees of a large n1anufacturing firm, the results showed that they considered lack of organizational support more stressful than experienced job pressure (Turnage & Spielberger, 1991). Results from a sample of MBA students and a sample of practicing managers showed that employees in highstress jobs were rated as more effective, committed, and burned out than employees in low-stress jobs when performance was measured by using subjective descriptive anchors (Rotondo Fernandez & Perrewe, 1995). Stressful situations may lead, not only to dissatisfaction, but also to somatic problems (coronary heart disease, hypertension, migraine headaches, insomnia), psychological difficulties (anxiety, depression, fear, etc.) and adverse behavioral reactions (for example drinking, smoking). While stress can thus manifest itself in many ways, one of the more significant symptoms from an organizational perspective is employee burnout. Burnout is characterized by deteriorating job performance and decreasing energy levels caused by the cumulative effect of continuing daily pressure (Etzion, 1988). One question that could be asked is if job satisfaction is an antecedent or a consequence of psychological burnout (Wolpin, Burke, & Greenglass, 1991).

38

To conclude, some researchers have pointed out the need of understanding stress not only from an individual perspective but also as an organizational phenomenon (Burk, 1993; Jaffe, 1995; James & Arroba, 1999). Further, James (1999) stated that a strategic initiative on stress can only be achieved by understanding stress as part of the fabric of organizational life.

2.3.8.2. Work interest
Interest is described as a pleasant emotional state that directs and sustains activity (Izard, 1977), and the presence or absence of interest in work tasks, and generally in life, colors the experience of existence and of what is worth our attention (Wiener, 1986). Izard (1977) classifies interest as an emotion others classify interest in attributional terms (Weiner, 1986). Interest is manifested behaviorally through voluntary participation in particular activities, sustained attention to and enjoyment of the activities, and subsequent persistence in them (Dawis, 1991). Bingham (1937) defined interest as "a tendency to become absorbed in an experience and to continue in it ... not only in terms of the objects and activities which get attention and yield satisfaction, but also in terms of the strength of the tendencies to give attention to and seek satisfaction in these competing objects of interest" (p. 62). Dawis (1991) stated that Bingham 's definition consists of several elements which are the following: An interest (a) is a dispositional variable (tendency), (b) has cognitive (attention), behavioral (experience), and affective (satisfaction) compol1ents, and (c) has dimensions of intensity (strength of tendency) and duration (continuance in experience). Moreover, Sjoberg (1997) claimed that interest is characterized by several states. Firstly, interest is a psychological condition (process) that is characterized by concentration and enjoyment. Secondly, it is easier for us to learn if we are interested. Thirdly, interest is strongly dependent on possibility of autonomy, like intrinsic motivation; we tend to lose interest if we perceive external control. Thus, it is possible that rewards associated with control might decrease work
39

interest. It is also possible that the way job tasks are construed has some influence on work interest. For example, optimal level of challenge is possibly one factor of importance, frequent feedback another. Finally, interest is a function of challenge and ability, which determines what is a moderately difficult challe11ge. It is important that a challenge can stimulate an activity where an individual has a good chance of succeeding. Similar findings as those of Sjoberg have been reported by Csikszentmihalyi (1992). Csikszentmihalyi has mainly studied expert performers ' experience of interest and fOllnd that interest is a psychological state characterized by strong concentration and sense of enjoyn1ent, which is described as "flow". The subjects described it as being strongly concentrated and that they felt joy and satisfaction. Some researchers argue that certain conditions must be fulfilled in order to develop interest: (a) the individual must have a certain specific ability and sensitivity, (b) there must be environmental possibilities and (c) social support (Deci, 1992; Lykke, Bouchard, McGue, & Tellegen, 1993). In addition, Cury, Biddle, Famose, Goudas, Sarrazin, and Durand (1996) found that perceived competence influenced interest positively, which was also found in the study of Harackiewicz, Barron, and Elliot (1997). Lykke, Bouchard, McGue, and Tellegen (1993) reported that about 50% of interest variance was associated with genetic variation. Further, in longitudinal studies, some stability over time was found for vocational interest (Austin & Hanisch, 1990; Dawis, 1991, Lubinski, Benbow, & Ryan, 1995; Swanson & Hansen, 1988). In fact, people often give interest as a reason for their vocational choice (Sjoberg, 1983). In prior studies, it has been reported that work interest has been associated with a host of positive consequences, such as greater perceived competence, freedom, and positive emotions (Deci, 1992) as well as a positive relation both to job satisfaction and commitment (Winer & Gati, 1986). A previous study showed that work interest is one of the most influential factors in explaining work motivation (Sjoberg & Lind, 1994). Furthermore, in a meta-analysis of predictors ofjob performance for salespeople, interest was a promising predictor
40

both of performance ratings (.50) and a sales criterion (.50) (Vinchur, Shippman, Switzer, & Roth, 1998). In conclusion, most people have fewer problems performing interesting activities at work than boring ones because they perceive interesting work tasks to be more enjoyable (Green-Demers, Pelletier, Stewart, & Gushue, 1998). That interesting work tasks was one of the most important factors when it comes to their job has been found in several studies (e.g., Giorgi, & Marsch, 1990; Kovach, 1995; Tollgerdt-Andersson, 1993; Quintanilla, 1990). Work interest has also been associated with meaningful work tasks, possibilities to make choices, self-determination, competence and producing results (Thomas & Velthouse, 1990).

2.3.8.3 Creativity
Creativity, which refers to employees ' generation of novel and useful ideas (Amabile, 1983) is a topic that is important at both the individual and societal level. Individuals, organizations, and societies must adapt existing resources to changing task demands in order to remain competitive. Therefore, creativity is important for organizations (Sternberg
&

Lubart,

1996).

Moreover,

organizations today operate in an increasingly uncertain world. Organizations constantly search for new marketing strategies, new products, services, new manufacturing processes, and new managerial practices to adapt to the rapidly changing environment. Almost all of these new ways of doing business start with new ideas. Thus, more than ever before, organizations strive for creativity and innovation in order to survive and grow (Kanter, 1983; Van Gundy, 1987). Numerous commentators have argued that enhancing the creative perfoffi1ance of employees is a necessary step if organizations are to achieve competitive advantage (Amabile, 1988; Shalley, 1995). When employees perform creatively, they suggest novel and useful products, ideas, or procedures that provide an orgal1ization with important raw material for subsequent development and
41

possible implementation (Amabile, 1988; Staw, 1990; Woodman, Sawyer, & Griffin, 1993). Scott (1996) has formulated a recently suggested definition of creativity. Scott stated that if creativity is defined as "the process of using imagination and skill to invent a unique product or thought, the employee is allowed more creative latitude" (p. 66). Scott claims that this proposal differs from standard definitions of creativity, because it incorporates artistic emphasis and the 'bottom-line concerns ' of management. Amabile (1985) suggested that the more skilled a person is in a specific area, the higher ability that person has to, for example, generate new ideas. Creative behavior is likely to be determined by a complex interaction between the attributes of the individual and the attributes of the environment (Mumford & Gustafson, 1988). A large body of literature has focused on determining a set of personal characteristics and attributes associated with creative achievement (Barron & Harrigton, 1981; Davis, 1989). This research has examined personal characteristics ranging from biological factors to measures of cognitive styles and intelligence (Amabile, 1983; Barron & Harrigton, 1981; Davis, 1989). In general, these studies have demonstrated that a stable set of core personal characteristics, including broad interest, attraction to complexity, intuition, aesthetic sensitivity, tolerance of ambiguity, and self-confidence, relate positively and consistently to measures of creative performance across a variety of domains (Barron & Harrigton, 1981; Martindale, 1989). Although the search for personal characteristics predictive of creative performance dominated creative research for several decades, recent research has begun to examine the effects of such contextual factors as goals, deadlines, and expected evaluations on the individual 's creative performance (Amabile, 1982; Amabile, Goldfarb, & Blackfield, 1990; Shalley, 1995). Most of this research, however, has been conducted in behavioral laboratories and has followed an "intrinsic motivation" perspective.

42

According to this perspective, the context in which an individual performs a task influences his or her intrinsic motivation, which in tum affects creative achievement (Amabile, 1988). Individuals are expected to be most creative when they experience a high level of intrinsic motivation, that is, when they are excited about their work activity and interested in engaging in it for the sake of the activity itself (Amabile, 1983; Shalley, 1991). Under these conditions, individuals are free of extraneous concerns and are likely to take risks, to explore new cognitive pathways, and to be playful with ideas and materials (Amabile, Goldfarb, & Brackfield, 1990) They are also likely to stay focused on the internal nature of the task and work longer on the idea or problem. Situations that encourage this exploration and persistence should increase the likelihood of creative performance. An intrinsically motivated individual tends to be cognitively more flexible (McGraw & Fiala, 1982), to prefer complexity and novelty (Pittman, Emergy, & Boggiano, 1982). Therefore, he or she is more likely to find many alternatives to solve a problem, to use nontraditional approaches, and to be persistent. All of these arguments suggest that an intrinsically motivated individual is more likely to exhibit high creativity. A number of studies have supported the arguments that intrinsic motivation leads to creativity (Amabile, 1985; Amabile, 1997; Koestner, Ryan, Bernieri, & Holt, 1984). Several organizational conditions or work environments have been suggested in order to enhance creativity. Ekvall (1996) identified ten organizational conditions that stimulate creativity and innovation: challenge, freedon1, idea support, trust/openness, dynamism/liveliness, playfulness/humor, debates, low conflicts, high risk taking, and idea time. Zhou (1998) reported that individuals who received positive feedback delivered in an informational style, and those who worked in a high task autonomy work environment, generated the most creative ideas. Amabile, Conti, Coon, Lazenby, and Herron (1996) stated that five work environment dimensions differed between high and low creative projects, and thus these dimensions may play an important role in influencing creative behavior in organizations: challenge, organizational encouragement, work group support, supervisory
43

encouragement,

and

organizational

impediments. Recent theory and research on organizational creatIvIty has emphasized the importance of creating favorable work environments to release employees ' creative energy (e.g., Amabile, 1988; Woodman, Sawyer, & Griffin, 1993). In addition, Oldham and Cummings (1996) reported that participants (171 employees from two manufacturing facilities) produced the most creative work when they had appropriate creativity-relevant characteristics, worked on complex, challenging jobs, and were supervised in a supportive, non controlling fashion. Creativity has been argued to be positively related to work motivation (Tollgerdt-Andersson & Sjoberg, 1992). Tollgerdt-Andersson and Sjoberg (1992) argued that creativity is a motivational factor in both general and specific work situations. Tierney, Farmer and Graen (1999) found that when employees enjoy creative-related tasks, their level of creative output is high. Kaufmann and Vosburg (1997) found that mood affects performance of creative problemsolving tasks and the worst performance was obtained in a positive mood condition. Furthermore, Amabile (1998, 1997) claimed that within every individual, creativity is a function of three components: expertise (knowledge-technical, procedural, and intellectual.), creative-thinking skills (determine how flexibly and imaginatively people approach problems) and intrinsic motivation (an inner passion to solve the problem at hand leads to solutions far more creative than do external rewards, such as money). Monetary rewards given for performance of an assignment over which the individual has no choice, can enhance creativity, but when the individual is offered a reward for consenting to perform the task, creativity n1ay be llndermined (Amabile, 1983). In addition, Amabile (1988) found that extrinsic rewards decrease creativity and the quality of what is produced. If a person is doing a work task for his or her own sake and not only because of external rewards the result is a higher creative level.

44

To encourage creativity, managers may consider giving positive feedback in an informal style. If feedback is going to be delivered, it is advisable to do so in an informal manner. In addition, when designing the work environment to facilitate creativity, managers need to consider multiple aspects of the organizational context, including social and task dimensions. The study suggests that by simultaneously giving positive feedback in an informal style, and allowing high autonomy at work, managers may substantially facilitate employees ' creativity (Zhou, 1998).

2.3.8.4 Work environment
The physical work environment includes everything from the design of the building to the location of public transportation or parking facilities. Characteristics of the physical work environment, such as the lightning, noise levels, temperature, air quality (Baron, 1994), and the availability of equipment needed to perform the job, may limit the degree to which employees can convert all of their well-intended efforts. Recognition that the work environment can significantly affect job satisfaction has stimulated research in all areas of work environment. For example, the general layout of working spaces in a building has been shown to affect such things as conununication (Nemcek & Grandjean, 1973). Lighting can be linked not only to feelings of fatigue but can also give impressions of clarity, spaciousness, and relaxation (Flynn, 1977). Altering the length or pattern of the work shift can frequently lead not only to happier but also to more productive workers as both boredom and fatigue are enemies of happiness and productivity (Evans, 1975). Four-day work weeks are becoming more and more popular. Management has found that the four-day week can facilitate recruitment, raise morale, lower absenteeism, and even increase production (Poor, 1973). In a study of nurses, varying shift work gave less job satisfaction than permanent shift work (lawaI & Baba, 1992).

45

Because noise can directly alter arousal levels, it can improve or retard performance depending on other task characteristics (Kryter, 1970). A number of offices allow music or play music to their employees. In general, most of these studies suggested that music has a small positive effect on performance (Sundstrom, 1986). Oldham, Cummings, Mischel, Schmidtke, and Zhou (1995) found that employees that used headset or could listen to music at work exhibited significant improvements in performance, turnover intentions, organizational satisfaction, mood states, and other responses. The mood state of relaxation explained the relation between stereo use and performance best. There is a second aspect of work environment which is the psychological one. G·umn1er (1998) emphasized tl1e enormous importance of the role played by social psychological factors in how people acted in organizations. Higgins (2000) reported that the more good relationships an individual has to colleagues the greater will his or her work satisfaction be. A study has shown that social support received from coworkers is an important factor for job satisfaction (Ducharme & Martin, 2000). Organizational support increases employees ' expectation and innovation behavior (Eisenberger, Fasolo, & Davis-LaMastro, 1990; Meyer & Allen, 1987). Eisenberger, Fasolo and Davis-LaMastro (1990) found that perceived organizational support and involvement in the employees ' work was quite strongly related to performance and absenteeism. Furthermore, a positive atmosphere and "good chemistry" between people at the workplace increases people 's productivity and well-being (Sjoberg & TollgerdtAndersson, 1985). Experimental studies of small groups have shown that there is a relationship between what people think about each other and performance (Lott & Lott, 1985). Furthermore, Giorgi and Marsch (1990) found that nice colleagues was one of the most important factors at work. Ravlin and Meglino (1987) showed that people valued that "one care about each other" at work. Eisenberger and his associates (Eisenberger, Huntington, Hutchison, & Sowa, 1986; Shore & Tetrick, 1991) proposed that employees in the workplace form a

46

general perception of the extent to which their organization values their contributions and cares about their well-being. In conclusion, most managers would certainly like to create a working environment in which people like working and in which people work well, a working environment which helps to enrich the life of those who work there. As shown here, the working environment is not only the physical one, like lightning and location, but also the coworkers ' support and social relationship in general at work is important for creating this kind of working environment.

2.3.8.5 Perceived risks
Risks and negative incidents can be perceived differently and have different effects depending on, for example, a person 's earlier experience. McLain (1995) claimed that different risks are associated with multifaceted subjective interpretations. The explorations of the ways in which individuals interpret societal, individual, and organizational risks suggest that different kinds of risks are associated with differing cognitive interpretations (March & Shapira, 1987; Slovic, 1987; Sitkin & Pablo, 1992). The aims of risk perception research are to understand what factors and what processes are involved in subjective judgment of risk (Drottz-Sjoberg, 1995). Risk is a common word in many languages, and its various ll1eanings were summarized by Drottz Sjoberg (1995): "risks cal1 be construed as either the mere

possibility of an adverse event, the cause of an event, the magnitude of the consequence, as someone or something judged as a hazard and as the

conceptualization of a procedure for the estimation of a quantity" (p. 123). This illustrates the variety of concepts included in the construct of risk. All those together constitute the risk concept (Drottz-Sjoberg, 1995). Perceived risks have been studied in relation to, for example, job satisfaction, where Leigh (1991) as well as Roberts (1993) found that dangerous working
47

conditions decrease job satisfaction. Moreover, Zaccaro a11d Stone (1988) reported that the perceived risk level and intellectual challenges are important aspects of job characteristics. Risk interpretation, such as an individual 's concerns about risk, the fairness of risk exposure, and the economic meaning of risk, has been found to be related to job satisfaction, just as concen1 with perceived work environment conditions, compensation fairness, and rewards for performance influence job satisfaction (Cook, Hepworth, Wall, & WaIT, 1981). McLain (1995) found that lower perceived risk and stronger belief that risk exposure was evenly distributed were positively related to satisfaction with both workplace conditions and with work in general. Hall and Spector (1991) reported that individuals working at the same job, and who experie11ced less satisfaction, perceived greater workload, dangerousness, and role conflict. Perceived risks have in a previous study by Sjoberg and Lind (1994) been related to work motivation. They found two kinds of perceived risks; one that is positively related to, and one negatively related to, work motivation. It is possible that some kind of risks and incidents decrease the motivation level, while others are accepted and do not affect motivation. Taking on a difficult task always involves a risk, and risk seems to have a dual effect on motivation. It gives rise to fear, of course, but in studies of fighter pilots flying difficult and somewhat risky missions under very high stress (Svensson Angelborg-Thanderz, Sjoberg 1993) a clear tendency was found that perceived risks gave challenge and a positive will to achieve (Svensson, Angelborg-Thanderz, Sjoberg, & Gillberg, 1988). People perceive risks or negative incidents differently, which makes it difficult to predict which factors are considered the most negative ones. Therefore, it is important to be aware of the influence that differently perceived risks have on people. It might be difficult to completely avoid some of the risks at work, but to be aware of them a11d maybe discuss them with the employees might help decreasing the negative influence of perceived risk on work motivation. A risk at a workplace could be dependent on the actual environment. It is important
48

that people feel safe at work and that it should be experienced as comfortable and that right facilities should be provided in order to do a good job.

2.3.8.6 Background factors
The age of the person, how long a person has worked in an organization or in a profession, marital status, and whether the employee is a man or a woman are only some of the background factors that have shown to affect an individual at work. Previous studies have shown that background data are among the best predictors of job performance (Hunter & Hunter, 1984; Mumford & Stokes, 1992). Several studies have examined the relationship between age and job satisfaction. In fact, observed age differences in overall job satisfaction are greater than those associated with gender, education, ethical background or income (Weaver, 1980). The relationship has shown to be inconsistent. Firstly, in an extensive review of the literature on age, Rhodes (1983) concluded that overall job satisfaction was positively associated with age with older employees tending to report higher satisfaction than younger ones (Doering, Rhodes, & Schuster, 1983; Warr, 1992). Secondly, the relationship between age and job satisfaction has also been shown to be u-shaped. Kacmar and Ferri (1989), as well as Oswald and Warr (1996), reported a U-shaped relationship between age and job satisfaction. Thirdly, Geyer and Daly (1998) examined the age-job satisfaction relationship in facility relocation settings and found a modest but significant negative correlation, which was not in line with the presented studies above. Finally, Rhodes (1983) found approximately equal support for each of four possible age-job satisfaction relations: negative, positive, curvilinear (inverted U), and non-significa11t. After reviewing 96 independent studies, McEvoy and Cascio (1989) concluded that age and job performance generally were unrelated across different occupations. However, for young employees the relationship between age and

49

performance was consistent and n10derately positive. In a meta analysis, Cohen (1993) reported that the relationship between age and organizational commitment as well as the relationship between tenure and organizational commitment was weak. The results of Wahn 's (1998) study showed a positive relationship between tenure and continuance commitment and a negative relationship between educational level and continuance commitment. Holtzer (1990) found that previous experience and tenure in the current job have significant positive effects on wages and productivity. Quinones, Ford, and Teachout (1995) found in their meta-analytic review a quite strong positive relationship between work experience and job performance. This was also the finding of Kolz, McFarland, and Silverman (1998). Meaningful work was more likely to be ranked as first preference by women, whereas promotion opportunities and security were more often ranked first by men (Tolbert & Moen, 1998). Such findings are consistent with a number of prior studies that have suggested that men are more oriented toward extrinsic rewards, whereas women set a higher value on intrinsic rewards (Beutel, 1995; Lueptow, 1996). Firebaugh and Harley (1995) found that 85% of U.S. workers were happy with their jobs and that men and women were equal in that regard. Aven, Parker, and McEvoy (1993) found in their meta-analysis no differences between men and women concerning affective commitment. Loscocco (1990) claimed that women may be less committed than men to employers because women 's first focus of loyalty is the home whereas career is second. Marsden, Kalleberg, and Cook (1993) reported that men tend to have slightly higher overall levels of general organizational commitment than women. Mathieu and Zajac (1990) found the opposite, that is, that women in general are more organizationally committed, although the difference was small. Wahn (1998) claimed that women reported a higher level of continuance commitment than men, but the sex difference was small.

50

2.3.8.7 Management/leadership
"Managers spend too much time in denial by insisting that they build a n10tivating workplace when they often sabotage it. A motivating work environn1ent is the responsibility of everyone" (En1ffierich, 1998, p. 20). Management is the process of getting productive work done through the efforts of other individuals; thus, understanding human motivation is essential for managerial success. To do this effectively, the manager must answer this question: "What will motivate people to willingly work toward organizational goals?" The manager n1ust predict, with reasonable accuracy, the kinds of behavior that result when different motivators are present. To make these predictions, the manager needs to understal1d what motives or needs are n10st likely to evoke productive behaviors in individuals at a particular time (Tin1m & Peterson, 2000). While understanding the nature of motivation and son1e of its effects on people is very useful, even more important is knOWil1g how to create situations where motivation can work for managers and for their organization (Timm & Peterson, 2000). Managers can influence motivation by doing the following: (1) Having realistic expectations, (2) Communicating about wants, needs, and goals, (3) Understanding the differences between motivators and maintenance factors, (4) creating a motivational climate (openness between managers and subordinates), and (5) using the reward system (Tin1m & Peterson, 2000). Supervisors and leaders have shown to have impact on subordinates ' motivation. Matching managers with teams that are not their natural fit will produce excessive stress and friction between both parties (Allender & Allender, 1998). Moreover, Lu (1999) stated that having a supportive supervisor at work could make all the difference in the world. Indeed, supervisors can supply necessary information and practical guidance, provide appreciation and recognition, decide promotion and increase in payment etc. Lu (1999) found that social support (from supervision) demonstrated consistent protective effects on well- being.
51

Tharenou (1993) investigated support received from supervisors, and found in a longitudinal study that such support reduced the level of uncertified absence. Supervisors ' consideration towards their subordinates was also shown by Zaccaro, Craig, and Quinn (1991) to be negatively associated with absenteeism in the previous six months. Miles, Patricks and King (1996) reported that communication (positive- relation communication, satisfaction. Furthermore, a leader 's behavior or leadership style has in several studies shown to be an important factor when it con1es to subordinates ' motivation as well as job satisfaction. Many different leadership/management styles have in the literature been suggested as successful. Holdnak, Harsh, and Bushardt (1993) found two correlations between a leader 's behavior or style and job satisfaction, suggesting that a leader who uses consideration in leadership has a positive impact on his or her subordinates ' job satisfaction. Holdnak et al. (1993) also found a negative relationship between initiating structure behavior (involves behaviors with which the leader organizes and defines the relationship in the group) and job satisfaction, suggesting that leaders who use an initiatingstructure style will see a decrease in job satisfaction and a negative relationship between initiating-structure leadership behavior and job satisfaction. Moreover, Bryman (1992) found that transformational leadership (by upward-openness communication and job-relevant communication) with one 's superior was a significant predictor of job

developing, inspiring and challenging the intellects of followers in order to go beyond their self-interest in the service of a higher collective purpose, mission or vision) behaviors were positively related to a number of important organizational outcomes, including perceived extra effort and job satisfaction.

52

2.3.9 Summary
To sum up, there are many ways of defining work motivation depending on how the human being is perceived. However, work motivation may be defined as the process by which behavior is energized, directed, and sustained in work settings. The definition of work motivation that was presented in this thesis is willingness to work, which is built on the assumption that it is the will of an individual that affects the work behavior. Moreover, many different factors and work n10tivation theories have been introduced in the literature over the years. The factors that were included in this thesis have in previous studies been shown to either enhance or decrease motivation. Some of the factors have traditionally been related to work motivation whereas others have more recently been presented. In addition, the presented predictors of work motivation emphasize different aspects. Some of the factors, such as payment, goal setting, work environment, leadership, and perceived risks are factors that can be found outside the individual, whereas intrinsic motivation, work interest, creativity, and perceived control are factors that are to a higher extent related to factors inside an individual. This shows the variety of factors that have been related to work motivation.

53

2.4 Work related behavior
Motivation is not directly observable. What we observe is a multidimensional stream of behavior and the products of those behaviors (Atkinson & Birch, 1970). There are three groups of behaviors or behavior outcomes associated with work motivation: directional, intensity and persistence behaviors (Kanfer, 1990). One way of looking at directional measures of work behavior is through withdrawal behaviors, for example, absenteeisn1 and turnover, which are most often used when investigating choice decisions between mutually exclusive courses of actions. In the category of intensity measures, job performance and task effort are common ways of measuring this. Nun1ber of hours worked is one way of assessing persistence. Sin1ilar criterion variables were used in a metaanalysis to evaluate the validity of Vroom 's Expectancy work motivation theory (Va11 Eerde & Thierry, 1996). Withdrawal, as well as job performance and nun1ber of hours worked, are presented in more detail below. A valid and useful measure of work motivation should be closely related to these behaviors, as work motivation is reflected in these work related behaviors. Sjoberg and Lind (1994) found that work motivation defined as willingness to work was strongly related to absenteeism as well as number of hours worked.

2.4.1 Direction (withdrawal behaviors)
Absenteeism and turnover are two of the most commonly acknowledged forms of withdrawal (Pinder, 1998). The voluntary turnover of desirable employees is in general considered detrimental to the organization, both in replacement costs and work disruption. Generally, withdrawal behaviors might not only mirror dissatisfaction but also low motivation level among employees.

54

2.4.1.1 Absenteeism
It is clear that absenteeism presents problems to both human resource practitioners and researchers. Discontinuous attendance at work can immediately lead to reduction in output, as a needed employee is not available. It has proven difficult to determine the factors influencing attendance or on nonattendance. One study has shown that absenteeism increased with age and won1en in general were more absent than men (WaIT & Yearta, 1995). Other studies have found that those employees who perceived social support at work as high reported fewer sickness absences (Cutrona & Russel, 1990; Dnden, 1994). Wooden (1995) listed organizational commitment as one of the most important factors that has impact on absence. He commented that higher levels of organizational commitment positively related to attendance, which suggested that increased organizational commitment could lead to substantially reduced levels of absenteeisn1. Significant, but weak negative relationships have previously been found to exist between organizational con1mitment and absenteeism in some studies (Blau, 1986; Mayer & Schoorman, 1992; Steers, 1977) but not in others (Ivancevich, 1985; Jamal, 1984). In fact, Ivancevich (1985) found that an individual 's past absenteeisn1 was a better predictor of subsequent absenteeism than the individual 's level of organizational commitment. Savery, Travaglione and Fims (1997) reported only a weak negative association between organizational commitment and absenteeism, and other researchers have cast doubt on the existence of a negative relation between organizational commitment and absenteeism (Angle & Perry, 1981; Mowday, Porter, & Steers, 1982). En1ployee absenteeism is a costly personnel problem attracting the attention of theoreticians and practitioners alike (Hackett, 1989). Considerable research on this topic has concerned the links between absence from work and work-related attitudes, such as job satisfaction. Hanisch and Hulin (1991) theorized that absenteeism reflects "invisible" attitudes, such as job dissatisfaction. However,

55

results have shown that no meaningful relationship between job satisfaction and absenteeism exists (Hackett, 1989; Hackett & Guion, 1985; Mayer & Schoorman, 1992). Two basic types of absences have been distinguished: involuntary (e.g., certified sickness) and voluntary (e.g., vacation, uncertified sickness). Voluntary absences are under the direct control of the employee and are frequently utilized for personal aims such as testing the market for alternative employment prospects. Conversely, involuntary absences are beyond the employee 's immediate control. Hence, voluntary rather than involuntary absence from work may reflect job dissatisfaction and lack of commitment to the organization. Consequently, one may expect that work attitudes will be more negatively related to voluntary absence than to involuntary absence (Sagie, 1998). Moreover, Sagie (1998) found that job satisfaction and organizational commitment were strongly related to aggregated duration of voluntary absence, but not of involuntary absence.

2.4.1.2 Turnover (Intention to quit/leave)
Employee turnover may result in dramatic financial costs to organizations. TUTI10ver not only has economic impact but also adversely affects training, organizational development, and other human resource-development interventions (Hatcher, 1999). Voluntary turnover has been the subject of extensive theorizing and research because of its potential costs to organizations in terms of loss of valuable human reSOtlrCes and disruption of ongoing activities (Cascio, 1991a). Over time, organizations invest substantial resources in their employees. The corresponding costs to the firm regarding employees quitting the organization and the subsequent hiring of replacement employees can be significant in terms of personal, work-unit, and organizational readjustments (Cascio, 1991b; Mobley, 1982).

56

In many models, turnover intention is the immediate precursor to actual turnover (e.g., Hom, Caranikas-Walker, Prussia, & Griffeth, 1992; Mobley, 1981; Mobley, Griffeth, Hand, & Meglino, 1979; Steers & Mowday, 1977; Sager, Griffeth, & Hom, 1998), and the relationship between intentions to quit and actual turnover has been well-documented in the literature (e.g., Hom, Griffeth,
& Sellaro, 1984; Steel & Ovalle, 1984). Fishbein and his colleagues (Fishbein,

1967; Fishbein & Ajzen, 1975) argued that behavioral il1tention is antecedent to actual behavior. Although the relationship between job satisfaction and intention to leave is generally thought to be negative (Carsten & Spector, 1987; Tett & Meyer, 1993), the magnitude of this relationship varies in the available literature. Spector and Jex (1991) found that the relationship was strong (r = -.59 for a sample of 322 U.S state civil-service employees. Hellman (1997) reported that the relationship between job satisfaction and intention to leave was negative
(r =-.47). A moderate relationship (r

= -.23, N = 190) was reported by Keller

(1984). Similarly, Newman (1974) used aspects of Fishbein 's behavioral intentions model to examine the turnover process among 108 U.S. nursing-home employees and found a quite weak relationship (r = -.10). Furthermore, employees from a large U.S. federal agency were less likely than employees within the private sector to leave the organization. Mathieu and Zajac (1990) as well as Randell (1990) have reported a weak relationship between organizational commitment and turnover. Werbel and Gould (1984) revealed an inverse relationship between organizational commitment and turnover for nurses employed more than one year. Previous work by Cohen (1991) has indicated that this relationship was stronger for employees in their early career stages (i.e., up to 30 years old) than for those in later career stages.

57

2.4.2 Persistence (Number of hours worked)
Behaviors that are referred to as persistence are behaviors related to, for example, how long a person works (Kanfer, 1990). It is important here to distinguish between voluntary and unvolutary hours worked, because a person who is forced to work overtime might not be motivated to do so, whereas a person who by free will stays at work might be motivated to do so. This behavior is rarely related to work motivation, because there might be other factors that affect the number of worked hours more than motivation itself. However, Wallence (1997) found that work motivation was positively associated with number of hours worked. Individuals who view their work as salient are highly committed to their work and/or career and therefore typically devote considerable time to their work role (Aryee, 1992; Frone & Rice, 1987; Greenhaus & Kopelman, 1981; Greenhaus, Parasuraman, Granrose, Rabinowitz, & Beutell, 1989; Wiley, 1987). Research that compares number of worked hours with gender of workers found that women work on an average fewer hours per week than men (Voydanoff, 1988). Wallence (1997) reported that internal factors Gob commitment) and external factors that reflect both work-related (work overload) and domestic (preschool children for won1en) pressures are related to number of hours worked.

2.4.3 Intensity (Job performance)
It is common to evaluate and compare theories of work motivation by the extent to which they successfully predict job performance (for a review, see Kanfer, 1992). Motivation and job perforrpance is not the same thing (Kanfer, 1990, 1992; Pinder, 1998). Kanfer (1990) stated that the construct of motivation subsumes the determinants and processes underlying the development of intentions, choice behaviors, and volitional activities.

58

The product of these motivational processes is the individual 's overt and/or covert behaviors. In contrast, performance typically refers to an evaluation of the individual 's behaviors. It is assumed that the performance occurs in the context of a job, position, or role in an orgal1ization, even if the organization consists of only two people. The organization is assumed to exist for accomplishing certain goals, even if they cannot be articulated. A full understanding of job performance depends on having son1e understanding of the organizational goals to which individual performance is supposed to contribute. Performance is something that people do and is reflected in the actions that people take. It includes only those actions or behaviors relevant to the organizatiol1al goals. Performance is not the consequence(s) or result(s) of action, it is the action itself. Admittedly, this distinction is troublesome in at least one major respect. That is, behavior is not always observable and can be known only by its effects (Campbell, 1991). Performance is to be distinguished from effectiveness and productivity. Effectiveness refers to the evaluation of the results of performance. By definition, the variance of a measure of effectiveness is controlled by more than the actions of the individual plus error (Campbell & Campbell, 1988). The definition of productivity, usually agreed upon, is the ratio of effectiveness to the cost of achieving that level of effectiveness (Mahoney, 1988). It is axiomatic that job performance is not one thing. A job is a very complex activity and there are a number of major performance con1ponents distinguishable in terms of their determinants and co-variation patterns with other variables (Campbell, 1991). Campbell (1991) suggested a performance model including three major determinants. The first is declarative knowledge, which refers to knowledge about the facts and things. Specifically, it represents an understanding of a given task 's requirements (e.g., general principles for equipment operation; Anderson, 1985; Kanfer & Ackerman, 1989). Examples of declarative knowledge include
59

knowledge of labels, facts, principles, goals, and self. The second determinant is procedural knowledge and skill, which is attained when declarative knowledge, or knowing what to do, has been successfully combined with knowing how to do it. Examples of procedural knowledge and skill include cognitive, psychon10tor, physical, perceptual, interpersonal, and self-management skills. The third determinant, motivation, could be seen as a direct detem1inant of performance. Often job performance can be divided into task performance and c011ceptual performance. Task performance consists of job-specific behaviors including core job responsibilities, for which the prin1ary antecedents are likely to be ability and experience. Conceptual performance consists of non-job-specific behaviors, such as cooperating with coworkers and showing dedication, for which the primary antecedents are likely to be volition a11d personality (Borman
& Motowidlo, 1993). The concept of conceptual performance has recently

attracted considerable research attention, most likely due to changes in the nature of organizations, such as team-based work (Borman & Motowidlo, 1997b). Motowidlo, Borman, and Schmit (1997) claimed that job performance is behavioral, episodic, evaluative, and multidimensional. Further, they suggested a definition of job performance as "the aggregated value to the organization of the discrete behavioral episodes that an individual performs over a standard interval of time". (p. 71). The theory that Motowidlo, Borman, and Schmit (1997) presented predicted that individual differences in personality and cognitive ability variables, in combination with learning experiences, lead to variability in knowledge, skills, and work habits that mediate effects of personality and cognitive ability on job perfoffi1ance. Moreover, the kinds of knowledge, skills, work habits, and traits that were associated with task performance are different from the kinds that were associated with conceptual performance.

60

As Campbell (1991) suggested, two of the major determinants of performance refer to different kinds of knowledge. It is important to let people in their own time acquire skills through practice (Ericsson, Krampe, & Tesch-Romer, 1993) to be able to perform well. To be ready to practice, the learner must, however, possess some initial competence. Exercising the initial competence by repeatedly attempting to perform a novel task engenders increased competence and, eventually, mastery, even in the absence of instruction. Kolz, McFarland, and Silverman (1998) found that both ability and experience were important predictors of work performance. The traditional view, supported by considerable empirical evidence, is that cognitive ability is the most important determinant of work motivation (Hunter & Hunter, 1984; Ree, Earles, & Teachout, 1994). The assumption is that motivated employees perform better, and that can depend on, for example, that they have a greater need for success, which can be met only by good performance (Staw, 1984). They are also more goal-oriented and compelled by their commitment to exert great effort in their work (Steers, 1977). It should, however, be noted that poor job performance is not always a result of low motivation (Pinder, 1998). In many cases it might be better attributed to external factors, such as characteristics of the work environment (Baron, 1994), and the availability of materials needed to perform ajob. Other factors than motivation and competence have been suggested to effect job performance. In pervious research it was found that autonomy (freedom to work independently) and discretion (participation in decision-making) enhance both motivation and job performance (Darden, Hampton, & Howell, 1989). Employees are also assumed to work better if their jobs are well defined and if they are recognized when they do their work well (Cheng & Kalleberg, 1996). Length of service with the current employer appears to enhance performance. Respondents who felt that they are noticed when they work well were more likely to report better performance. Workers who report that hard work leads to pay rises tend to report poorer quality performance, which is surprising because pay rises are presumably used by employers to stimulate better performance (Cheng & Kalleberg, 1996). Ree, Earles, and Teachout (1994) found that
61

general cognitive ability is the best predictor of all criteria and that specific abilities or knowledge added only a small amount to predict job performance (Ree, Earles & Teachout, 1994).

2.4.3.1 Ratings of job performance
Arvey and Murphy (1998) stated that the notion that job performance is more than just the execution of specific tasks and that it involves a wide variety of organizational activities has important implications for the understanding and measurement of job performance. Different approaches have been suggested in order to rate performance. One way of measuring performance is through subjective n1easurement and another is to use an objective measurement. In a study comparing self-reported job performance ratings and ratings given by their superiors, for example, Heneman (1974) found that self-report measures had less halo error, restriction of range, and leniency than the more objective measures, such as sales, a finding supported by Arvey and Murphy (1998) who stated that subjectivity does not automatically translate into rater error or bias and that selfratings are most likely valid reflections of true performance. Further, Jourden and Heath (1996) reported that after completing a task, most individuals typically rank their performance below the median (a negative performance illusion) and most group members rank their group performance above the median (a positive performance illusion). Group n1embers consistently assign their group a higher rank than individuals assign themselves. Moreover, Heneman concluded that self-reported measures are more appropriate for research purposes than for organizational evaluations. In addition, more objectives measures are useful only in specific settings and cannot be applied to the whole labor force (Steers, 1977; Judge & Ferris, 1993).

62

A number of recent studies have examined the relative value and interchangeability of different types of performance measures. Bommer, Johnson, Rich, Podsakoff, and MacKenzie (1995) assessed the relationships between relatively objective and subjective measures of employee performance. Using meta-analytic techniques to summarize the relationships for over 50 independent samples, the overall corrected mean correlation between the two types of measures was .39, suggesting that the two measures were significantly and moderately related but not totally substitutable. Heneman (1986) examined the relationship between supervisory ratings of performance and result-oriented measures of performance - essentially a comparison of subjective versus objective criterion measures. He reported a mean correlation of .27 between subjective and objective performance measures.

2.4.4 Summary
Work motivation itself is hard observe, instead it is reflected in different behaviors. Behaviors that often are studied to understand work motivation are absenteeism, turnover, number of hours worked, and job performance. For a work motivation measure to be valid and useful it should be able to predict differe11t work-related behaviors. Commonly used measures of work motivation (e.g., job satisfaction and organizational commitment) have not shown strong relationships to work-related behaviors, for example, job performance. However, Sjoberg and Lind (1994) found that work motivation defined as willingness to work was strongly related to both absenteeism and number of hours worked.

63

3. Research approach
This section of the thesis presents the research approach of the three empirical studies. The participants of the studies were, in Study 1, pre-school employees, in Study 2, employees of an insurance company and in Study 3, salespeople in an insurance company. In the introduction section, advantages and disadvantages using a qualitative approach is discussed. Further in this part of the thesis, the three work groups are presented with background data. The scales used in the studies are presented with descriptive statistics including reliability measures.

3.1 Introduction
The purposes of the thesis were to first explore the factors that explain the variance in willingness to work (work motivation) and secondly to investigate the relations between willingness to work and work-related variables. Three studies were included in the thesis, and two of these, Study 1 and 2, were very similar in research design and had the same research purpose. In fact, the questionnaires used were similar with only a few differences and included factors that were hypothesized to affect work motivation as well as self-rated work-related behaviors. In the third study, the purpose was to investigate the relationship between willingness to work and job performance, and in this study a somewhat different questionnaire was used. There are both advantages and disadvantages using questionnaires (a quantitative approach). First of all, using a questionnaire is a simple and effective method to collect data from a large sample. The researcher 's influence on the participants is minimized, and the respondents are able to choose whether they want to fill it in. Secondly, Chadwick, Bahr, and Albrecht (1984) argued that the questionnaire survey is a useful way to obtain information about sensitive topics. It is suggested that to report how often one exhibits deviant or

64

disapproved behavior it is easier for the respondent completing a questionnaire than it would be in face-to-face interaction with an interviewer. Finally, if the respondent is convinced that the questionnaire is anonymous, he or she can freely report attitudes and behaviors without embarrassment or fear of reprisal. Chadwick, Bahr, and Albrecht (1984) also brought up some disadvantages of using this kind of method. First, they indicated that valuable information might be lost if the researcher cannot follow up interesting leads. However, to minimize this problem, interviews were done prior to the survey in order to increase knowledge about the studied groups and to be certain that valuable information would be included in the final questionnaire. The interviews were done with 16 persons for Study 1 (employees in pre-schools) and 16 for Study 2 (employees in an inStlrance company) and 6 for Study 3 (insurance sales personnel). Second, another limitation might be that a questionnaire must be relatively brief, otherwise most respondents will not take the time to complete it (see Aaker & Day, 1990). Although the questionnaire in the samples was very long, the response rate in the three studies were quite high, 72% for the pre-school employees, 82% for employees in an insurance company, and 73% for salespeople in an insurance company. Although the response rates were quite high, there were employees who did not respond. The respondents ' average time for answering the questions was approximately 60 minutes in Study 1 and 2. Some of the respondents stated that the questionnaire included too many questions and that it took long time to answer and that might be one of the reasons for the non-responses. There might be other reasons as well. Rogelberg and Luong (1998) presented four classes of non-response, the individual (a) never actually received the survey, (b) is unable to complete it (e.g., ill, cannot read), (c) misplaced or forgot the survey out of carelessness, or (d) has made a conscious decision not to respond to the survey (i.e., noncompliance). Finally, one should also be aware of problems that could be related to using selfreports of, for example, behaviors and attitudes, because they are influenced by
65

features of the research instrument, including question wording, format, and context (Schwarz, 1999). In the next section, all of the studied groups are presented and the questionnaires used are shown in detail.

3.2 Study 1 (pre~school employees)

3.2.1 Participants
Twenty-three pre-schools in the county of Stockholm, including the city of Stockholm, participated in the study. The sample consisted of 179 employees. The number of women who participated was 165 while 14 respondel1ts were men. The age range of the participants was 25 to 64 years with an average age of 44 years. The number of years worked in the organization ranged from 1 year up to 37 years, with an average of 11.5 years. The group with the highest representation in this study was pre-school teachers, 48% of the sample, with a training of two-and-a-half years of post-secondary studies. The rest of the participants had a variety of other types of educations.

3.2.2 Procedure
The selection of pre-schools was based on the criterion of a sufficiently large, heterogeneous sample. There were 23 pre-schools randomly selected. The questionnaires and an envelope were distributed to each participant and also collected by the author. Instructions were both written and oral. The total number of distributed questionnaires was 250 of which 179 responded (response rate of approximately 72%).

66

3.2.3 Questionnaire
The questionnaire included 16 scales and background questions. The questions were to be rated on 5-, 6- or 7-point Likert-type scales. The midpoints were labeled "neither-nor". The extreme positive and negative anchors of the scales usually had the verbal labeling of "very" together with the positive or negative keyword. Regarding some background questions the respondents simply answered "yes" or "no", and supplied further information by using an openended format. Many of the questions included in the questionnaire were developed and used in the study of Sjoberg and Lind (1994). Other scales were either adapted from previous research or developed by the author. The instruments developed for the present study were based on 16 interviews with employees working at different pre-schools. The purpose of the interviews was to increase the knowledge of the studied group, mainly in the areas of work tasks, work environment, attitudes towards the job, different job risks and of stressful situations. In order to be certain that the questionnaire was relevant for the group, 16 employees other than the ones who were interviewed, were invited to participate in a focus group, where the questionnaire, with an emphasis on the developed questions for this particular study, was discussed and the face validity of the questionnaire was confirmed. The scales included in the study are presented in next section.

67

3.2.3.1 Scales included
The scales that were included in the study were either developed by Sjoberg and Lind (1994), developed by others or developed for this particular study. Work motivation**2 was the main variable in the study. Work motivation defined as willingness to work was measured using Sjoberg and Lind 's (1994) 12-iten1 scale. This scale has been used among people working in the service sectors as well as in the manufacturing industry (Sjoberg & Lind, 1994). The scale measured how willing a person was to work by, for example, asking questions about his or her general motivation, how often the respondent felt strong will to work, eto. The respondent was presented items with a response scale ranging from 5 (Positive) to 1 (Negative). Overall job satisfaction was measured by one single item. The questions asked how satisfied he or she was with the job. The 7-point scale was anchored by 7

(Very satisfied) and 1 (Not satisfied at all). Several studies have shown that a global rating of job satisfaction is an inclusive measure of overall job satisfaction (e.g., Scarpello & Campbell, 1983; Wanous, Reichers, & Hudy, 1997). Organizational commitment was measured by a 24-item scale developed by Allen and Meyer (1990). The scale assessed an individual 's degree of commitment to an organization. The respondent indicated how she or he felt about each iten1, using a 6-point scale ranging from 6 (Strongly agree) to 1

(Strongly disagree).

2

The scales that are presented in Appendix B are marked with

**.

The scales that are not

presented can be obtained from the author.

68

General work attitudes** was measured with 18 questions regarding an

employee 's attitude towards his or her job. This scale was developed by Sjoberg and Lind (1994) for this particular study. The respondent indicated how she or he felt about each item, using a 6-point scale ranging from 6 (Strongly agree) to
1 (Strongly disagree). Opinions of one 's job** was measured by 24-item scale. This measure was

developed for this particular study. The respondent indicated how she or 11e felt about each item, using a 6-point scale ranging from 6 (Strongly agree) to 1
(Strongly disagree). Questions included brought up questions concerning

general beliefs/opinions about one 's job.
Social relations was measured with 2 questions regarding relations to other

colleagues. This scale was developed by Sjoberg and Lind (1994). The 5-point scale was anchored by 5 (Very good) and 1 (Very bad).
Work environment included 15 questions about different aspect in the physical

environment. This scale was developed for this study. The 5-point scale was anchored by 5 (Very satisfying) and 1 (Very dissatisfying).
Work interest** was measllred by a scale that included 35 specific work tasks

for this particular group. The respondent rated how interesting he or she found different work tasks. The 7-point scale was anchored by 7 (Extremely interesting) and 1 (Completely uninteresting).

Leadership consisted of 4 question concerning the mangers knowledge and interest of one 's job. This scale was developed by Sjoberg and Lind (1994). The respondent indicated how she or he felt about each item, using a 5-point scale ral1ging from 5 (Yes, absolutely) to 1 (Absolutely not).

69

Creativity was measured by 17-item scale developed by Sjoberg and Lind (1994). The scale assessed the degree to which an employee use their creativity. The respondent indicated how she or he felt about each iten1, using as-point scale ranging from 5 (Strongly agree) to 1 (Strongly disagree). Perceived risks** was originally developed by Sjoberg and Lind (1994). In this study risks that were related specifically to pre-schools were added. The scale contained 27 different risks related to professional life (physical, personal or social). The respondent indicated the frequency of encountering different risks described by each item using a scale anchored by 4 (Never) and 0 (Very often). Intrinsic motivation was measllred by 21 questions and the scale was developed by Ryan (personal communication, 1994). The instrument assessed an individual 's perceived competence, autonomy, and relatedness. The respondent indicated how she or he felt about each item using a 6-point scale anchored by 6 (Strongly agree) and 1 (Strongly disagree). General intrinsic and extrinsic motivation was measured by Amabile, Hill, and Hennessey 's (1994) Work Preference Inventory including 15 questions about intrinsic motivation (e.g., the degree to which a person is motivated by activities that are spontaneously initiated and experienced as enjoyable) and 15 questions about extrinsic motivation (e.g., the degree to which a person is motivated by external h1ducements). The respondent indicated how she or he felt about each item using a 6-point scale anchored by 6 (Strongly agree) and 1
(Strongly disagree).

Occupational and work stress was assessed with Job Stress Survey (JSS) a 19item scale developed by Turnage and Spielberger (1991). The respondent rated how often he or she experience stress in different situations. The 7-point scale was anchored by 7 (Never) and 1 (Very often).

70

Pay satisfaction was measured with a 25-item scale. The instrument was developed by Carraher (1991). The questions asked how satisfied he or she was with different aspects of payment. The 5-point scale was anchored by 5 (Very

satisfied) and 1 (Not satisfied at alI).

Perceived control was assessed with a scale developed by Dwyer and Ganster
(1991). Questions about how an employee perceives hin1- or herself having control at work were asked. The respondent indicated how she or he felt about each item using a 5-point scale anchored by 6 (To a very high degree) and 1 (To

a very low degree).

Reported number of hours worked per week (effective time) was measured using one question. The participant was asked how many hours he/she spent each week at work. The response alternatives were 1 (Much less than 40 hours),

2 (Less than 40 hours), 3 (Approximately 40 hours), 4 (More than 40 hours), 5 (Much more than 40 hours). In Sweden, a full-time week is 40 hours.

Absenteeism was measured by only one question: how many days the respondent had been absent from work during a period of twelve months. The response alternatives were 1 (More than 20 days), 2 (16-20 days), 3 (11-15

days), 4 (6-10 days), and 5 (0-5 days).

Intention to quit was measured by two items, which brought up the questions about whether the respondent had thought about quitting his or her job and how often the participant had been looking for a new working position. The response alternatives ranged from 1 (Often) to 5 (Never).

Background factors included questions about age, income, sex, number of years working in the organization, and education.

71

3.2.3.1.1. Indices

A large number of measurements or questions were used. One way to reduce the multiplicity of test and ' measures to greater simplicity is to perform a factor analysis. The purpose of using factor analysis is to summarize the interrelationships among the variables in a concise but accurate manner as an aid in conceptualization (Gorsuch, 1983). Furthermore, most factor analytic methods initially produce results in a form that is difficult or impossible to interpret. A principal factor matrix and its loadings account for the common factor variance of the test scores, but they do not in general provide scientifically meaningful structures. It is the configurations of tests or variables in factor space that are of fundamental concern. In order to discover these configurations adequately, the arbitrary reference axes must be rotated (Kerlinger, 1986). The rotation that was used in the reduction process in the thesis was varimax

rotation.
Organizational commitment Organizational comn1itment was measured with Allen and Meyer 's (1990) scale. Allen and Meyer (1990) suggested that organizational commitment should be divided into three dimensions, affective, normative, and continuance organizational commitment. Questions related to the dimension affective organizational commitment deal with affective reactions to the organization, normative organizational commitment with the employee 's feelings of obligations to the organization, and continuance organizational conunitment with questions about commitment to continue one 's work tasks. In the factor analysis a number of factors were a prior determined due to earlier findings (c.f. Cohen, 1996; Irving, Coleman, & Cooper, 1997). A KMO measure (.811) and Bartlett 's test (p < .000) indicated that the data

3

were suitable for factor analysis (cf. Hair, Andersson, Tatham, & Black, 1998).
3

Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy is an index used to examine the appropriateness of factor analysis. High values (between .5 and 1.0) indicate that factor analysis is appropriate (Malhotra, 1999).
72

In Table 1, the result of a principal component analysis is presented with a varimax rotation. The result confirmed the three factors affective, continuance, and normative organizational commitment. There were no overlaps between the factors. However, two of the items had a factor loading smaller than .40, and therefore those two were excluded. Otherwise the factor loadings within the factors were rather high.

73

Table 1 Three-factor solution obtained in an analysis of Allen and Meyer 's organizational commitment scale (Study 1).
Factor loadings
Affective Continuance Normative

DC
I would be very happy to spend the rest of my career with this organization I enjoy discussing my organization with people outside it I really feel as if this organization 's problems are my own I think that I could easily become as attached to another organization as I am to this one I do not feel like ' a part of the fanlily ' at my organization (R) I do not feel 'emotionally attached ' to this organization (R) This organization has a great deal of personal meaning to me I do not feel a strong sense of belonging to my organization (R) It would be very hard for me to leave my organization right now, even if I wanted to Too nluch in my life would be disrupted if I decided I wanted to leave my organization now It wouldn 't be too costly for me to leave my organization now (R) Right now, staying with my organization is a matter of necessity as much as desire I feel that I have too few options to consider leaving this organization One of the few serious consequences of leaving this organization would be the scarcity of available alternatives One of the major reasons why I continue to work for this organization is that leaving would require considerable personal sacrifice - another organization may not match the overall benefits I have here I think that people these days move from conlpany to company too often I do not believe that a person must always be loyal to his or her organization (R) Jumping from organization to organization does not seem unethical at all to me (R) One of the major reasons why I continue to work for this organization is that I believe that loyalty is important and therefore feel a sense of moral obligation to stay If I got another offer of a better job elsewhere I would not feel it was right to leave my organization I was taught to believe in the value of remaining loyal to one organization Things were better in the days when people stayed with one organization for most of their careers .66 .49 .60 .44 .79 .59 .72 .42

DC

DC

.61 .62 .40 .48 .79 .75
.41

.58
.41 .70

.62

.55 .51 .65

Note. Loadings below .40 are not shown in the table. OC= organizational commitment. (R)= items with reversed coding.

74

General attitudes

General attitudes towards one 's job was measured with Sjoberg and Lind 's (1994) scale. In their study, two factors were found: positive evaluation of one 's job and negative evaluation of one 's job. A prior determined factor analysis with a two- factor solution was performed. The KMO (.885) measure as well as
Bartlett 's test (p < .000) showed that a factor analysis was suitable for the data.

The two factors were confirmed also in this study. The result is shown in Table 2. Furthermore, two of the items from the original scale were deleted because the factor loadings were under .40. Otherwise the factor loadings were high.

Table 2 Two-factor solution obtained in an analysis ofa general attitude scale (Study 1)
My job is nice My job is ideal for me My job is good My job is valuable My job is fantastic My job is better than most jobs I am satisfied with my job My job is excellent I am comfortable with my job My job is bad My job is a waste of time My job is not suited for me My job is worse than most jobs My job is not acceptable to me My job is very bad My job is worthless Positive evaluation of one 's job .59 .72 Negative evaluation of one 's job

.62 .51,
.88 .76

.86
.88

.70
.62
.76 .70 .55 .60 .70 .63

Note. Loadings below .40 are not shown in the table.

75

Work interest
Work interest for specific work tasks was measured with a scale that was developed for this specific work group. In order to n1inimize the number of questions a principal component analysis was performed. The KMO n1easure (.751) and Bartlett 's test (p < .000) indicated that a factor analysis was appropriate for the data. The number of factors chosen was based on the cumulative percentage of explained variance. Accordingly, three factors were selected due to the total variance explained (58%). Varimax rotation was then used to get a more interpretable solution. A satisfactory interpretation of the three-factor solution was found. The first factor was called interest in pedagogical work tasks (e.g., reading to the children), the second one interest in practical work tasks (e.g., preparing meals), and the third one interest in administrative work tasks (e.g., marketing the pre-school). Two of the items from the original scale were excluded because the factor loadings were smaller than .40. The result is illustrated in Table 3.

76

Table 3 Afactor analysis with a varimax rotation ofwork interest among pre-school employees (Study 1)
Factor loadings
Pedagogical work task To meet with colleagues To read to the children To lead creative playing (Le., painting) To play theatre with children To make excursions with children To work with a mixed age group of children Get further advance training To read literature about pedagogy To lead sports To tell the children about society To teach the children reading, counting and writing To let the children take initiative To meet with children To be outside playing with children To work with the oldest children To help the children to get dressed Parental n1eetings To supervise trainees To prepare meals To solve conflicts between the children To work with the youngest children To eat with the children To shop for the pre-school To cooperate with other pre-schools To clean To meet with parents To have contact with authorities Accounting Management .61 .63 .55 .43 .48 .67 .46 .60 .58 .48 .66 .73 .58 .60 .53 .74 .51 .62 .53 .55 .48 .52 .47 .58 .53 .66 .82 .79 .71 Practical work tasks Administrative work tasks

Note. Loadings below .40 are not shown in the table.

Intrinsic motivation
Intrinsic motivation was measured with Ryan 's (personal communication) scale. The developer of the instrument suggested that the scale contained three dimensions and therefore three factors were a priori determined in the factor analysis. The KMO measure (.816) and Bartlett 's test (p < .000) showed that a factor analysis was appropriate to use for the data. The suggested dimensions were perceived competence (e.g., the degree to which an employee perceives him/herself as competent), perceived relatedness (e.g., the degree to which an
77

employee feels related to the organization and colleagues), and perceived autonomy (e.g., the degree to which a person is not controlled). The three factors were confirmed. There were several items that were deleted, because the factor loadings were smaller than .40. The result is presented in Table 4.

Table 4 Three-factor analysis ofRyan 's intrinsic motivation scale (Study 1)
Factor loadings
Autonomy I feel like I can make a lot of input to deciding how my job gets done I am free to express my ideas and opinions on the job My feelings are taken into consideration at work I feel like I can pretty much be .myself at work I really like the people I work with I get along with people at work I pretty much keep to myself when I am at work (R) I consider the people I work with to be my friends People at work care about me The people I work with do not seem to like me much People at work are pretty friendly towards me I do not feel very competent when I am at work (R) People at work tell me I am good at what I do I have been able to learn interesting new skills on my job Most days I feel a sense of accomplishment from working On my job I do not get much of a chance to show how capable I am (R) When I am working I often do not feel very capable (R) .78 Relatedness Competence

.73 .47
.71 .85 .79

.56
.80 .78 .40 .78

.70 .40 .71

.63 .70 .63

Note. Loadings below .40 are not shown in the table.

Occupational stress
The developers of the Job Stress Survey (ISS; Turnage & Spielberger, 1991) suggested the scale to be divided into two factors, one measuring lack of support (e.g., degree to which lack of support affects an employee) and the other job pressure (e.g., the degree to which job pressure affects an employee). The KMO measure (.772) as well as Bartlett 's test (p < .000) showed that a factor analysis was suitable. A factor analysis with two factors was performed and the prior finding was confirmed. The result is shown in Table 5.

78

Table 5 Two-factor analysis ofjob stress (Study 1)
Factor loadings
Lack of support Lack of opportunity for advancement Poor or inadequate supervision Inadequate support by supervisor Personal insult from customer/consumer/colleague Fellow workers not doing their job Difficulty getting along with supervisor Inadequate salary Negative attitudes towards organization Lack of participation in policy-making decisions Poorly motivated co-workers Making critical on-the-spot decisions Inadequate or poor quality equipment Insufficient personnel to handle assignment Dealing with crisis situations Frequent interruptions Meeting deadlines Excessive paperwork Competition for advancement Noisy work area Assignment of new or unfamiliar duties Assignment of increased responsibility Covering work for another enlployee Frequent changes from boring to demanding activities Working overtime Performing tasks not in job description Insufficient personal time .53 .77 .74 .40 .59 Job pressure

.76
.48

.56
.42

.51
.54 .57 .58 .41 .60 .59

.43
.59

.52 .51
.57 .59 .47 .63 .54 .68

Note. Loadings below .40 are not shown in the table.

79

General intrinsic and extrinsic nl0tivation Amabile et al. (1994) developed a scale called Work Preference Inventory which assesses individual differences in intrinsic and extrinsic motivational orientation. The KMO measure (.683) and Bartlett 's test (p < .000) showed that a factor analysis was appropriate for the data. The two suggested factors were confirmed in this study. The result of a two-factor analysis is shown in Table 6. There were, however, 4 items in the intrinsic scale and 6 in the extrinsic scale with factor loadings smaller than .40. The result is illustrated in Table 6.

Table 6 Two- factor analysis intrinsic and extrinsic motivation (Study 1)
Factor loadings
Intrinsic motivation I enjoy tackling problems that are completely new to me I enjoy trying to solve complex problems The n10re difficult the problem is, the more I enjoy trying to solve it I want my work to provide me with opportunities for increasing my knowledge and skills Curiosity is the driving force b~hind much of what I do I enjoy relatively simple, straightforward tasks I enjoy doing work that is so absorbing that I forget about everything else What matters most to me is enjoying what I do It is important for me to have an outlet for self-expression I prefer work I know I can do well over work that stretches my abilities No matter what the outcome of a project is, I am satisfied if I feel I gained a new experience I want to find out how good I really can be at my work I am keenly aware of the (promotion) goal I have for n1yself I am strongly motivated by the (grades) (money) I can earn I am strongly motivated by the recognition I can earn from other people I want other people to find out how good I really can be at my work I seldom think about (grades and awards) (salary and promotion) . To me, success means doing better than other people I have to feel that I 'm earning something for what I do I prefer working on projects with clearly specified procedures I prefer having someone set clear goals for me .84 .81 .80 .52 .76 .41 .40 .30 .44 .55 .48 .60 .40 .46 .44 .60 .59 .58 .57 .62 .40 Extrinsic n10tivation

Note. Loadings below .40 are not shown in the table.

80

Perceived risks
Two dimensions of risks were suggested by the developers (Sjoberg & Lind, 1994). The risks that were positively related to work motivation were denoted accepted risks and the others that were negatively related to work motivation were denoted risk burden. The risks that were positively related to willingness to work were, for example, "That I could catch a less serious infection", "That I perceive myself as incapable of demands the job requires". Risks that instead were negatively related to willingness to work were, for example, "That I have to give l1egative information to others", "That a parent is critical about me".

Creativity
Creativity could according to the developers Sjoberg and Lind (1994) be divided into two indices, the degree to which an employee uses his or her creativity, spontaneous creativity, and the degree to which an employee does not use his or her creativity, creativity inhibition. The questions included in spontaneous creativity were, for example, "I often find solutions to problems when I least expect it" and "Intuition is most important in my job". Questions related to creativity inhibition were instead, for example, "My work tasks can best be solved with well-known routines" and "The most important thing in problem solving is not to make any mistakes".

One dinlensional scales
Pay satisfaction, perceived control, work environment, leadership, and opinion of one 's job were analyzed as a one-dimensional scale. Heneman and Schwab (1985) suggested, however, that the pay satisfaction scale could be divided into five dimensions which are: satisfaction with pay level, benefits, raises, structures, and administration. Although this might have had an impact on the result, the decision was made to treat the scale as one-dimensional. A reliability al1alysis was performed all of the indices and it is presented il1 the next chapter.

81

3.3 Study 2 (employees of an insurance company)
3.3.1 Participants
A total of 160 employees of an insurance company in Sweden participated in the study. The age range was from 25 to 64 years with a mean age of 44 years. There were 79 men and 81 women. The mean educational level of the respondents was 3-4 years of high school in the technical field. The length of employment in the organization was from 1 to 37 years with a mean of 11.5 years. The number of years working within the occupational area varied from 1 to 41 with 18.1 years as the mean length of experience.

3.3.2 Procedure
The questionnaires were distributed to each participant and also collected by the author. The participants were given a questionnaire together with an envelope, in order to ensure anonyn1ity. Instructions were written on the questionnaire. The author collected the questionnaires after about four weeks. The total response rate was 82% (160 out of 195).

3.3.3.Questionnaire
The majority of the of the instruments were similar to those used in Study 1 with a few exceptions. The scales that were included in Study 1, but not in this particular study, was general work attitudes (positive and negative evaluation of one 's job). The reason for not including this scale in this study was mainly the similarity of this scale to that of opinion about one 's job. Instead of the two excluded scales, two other instruments were added measuring goal setting and general work interest.

82

3.3.3.1 Scales included
A similar questionnaire was used in Study 1 and 2. Therefore, the only scales presented here are the ones not present ill Study 1.

Work interest** was measured by a scale that included 40 specific work tasks for this particular group. The respondent indicated how interesting he or she found different work tasks. The scale ranged from 7 (Extremely interesting) to 1 (Completely uninteresting). Perceived risks** were measured by 14 risk items concerning professional life (physical, personal, and social). The respondent indicated the frequency of encountering different risks described by each item using a scale anchored by 4 (Never) to 0 (Very often). Goal setting was assessed with 53-item scale developed by Locke and Latham (1984). The questions included were related to goals and how they affect an individual 's job. The scale ranged from 5 (Strongly agree) to 1 (Strongly disagree).

General work interest was assessed with 1 item. This question was developed by Sjoberg and Lind (1994) and the respondent was asked how their work interest developed over the years. The scale range from 5 (Increased very much) to 1 (Decreased very much).

83

3.3.3.1.1 Indices

A principal component analysis was used also in this study to obtain a more manageable number of variables. In several of the scales, factors were already suggested by the developer of the instrument. Principal component analyses were also in this case perfom1ed in order to confirm the factors found in Study 1.

Organizational commitment The results of a principal components analysis using a varimax rotation showed that the three factors su:ggested by Allen and Meyer (1990) were formed. The
KMO measure (.839) and Bartlett 's test (p < .000) indicated that the data were

suitable for factor analysis. Two of the items were deleted due to factor loadings smaller than .40. The factors that were obtained in Study 1 were confirmed also in Study 2, which validates that the instrument contain three factors. The factor loadings were quite high, and no overlaps between the factors were found. In Table 7 the result is presented.

84

Table 7 Three-factor solution obtained in an analysis of Allen and Meyer 's organizational commitment scale (Study 2)
Factor loadings
Affective OC Continuance OC Normative

DC

I would be very happy to spend the rest of my career with this organization I enjoy discussing my organization with people outside it I really feel as if this organization 's problems are my own I think that I could easily become as attached to another organization as I am to this one I do not feel like 'a part of the family ' at my organization (R) I do not feel 'emotionally attached ' to this organization (R) This organization has a great deal of personal meaning to me I do not feel a strong sense of belonging to my organization (R) It would be very hard for me to leave nlY organization right now, even if I wanted to Too much in my life would be disrupted if I decided I wanted to leave my organization now It wouldn 't be too costly for me to leave my organization now (R) Right now, staying with my organization is a matter of necessity as much as desire I feel that I have too few options to consider leaving this organization One of the few serious consequences of leaving this organization would be the scarcity of available alternatives One of the major reasons why I continue to work for this organization is that leaving would require considerable personal sacrifice - another organization may not match the overall benefits I have here I think that people these days move from company to company too often I do not believe that a person must always be loyal to his or her organization (R) Jumping from organization to organization does not seem at all unethical to me (R) One of the major reasons why I continue to work for this organization is that I believe that loyalty is important and therefore feel a sense of moral obligation to stay If I got another offer of a better job elsewhere I would not feel it was right to leave my organization I was taught to believe in the value of remaining loyal to one organization Things were better in the days when people stayed with one organization for most of their careers

.71

.73 .73 .53 .87 .87
.83 .90
.81

.82 .70 .69

.68 .51
.57

.41

.53 .67 .77

.69 .64
.65

Note. Loadings below .40 are not shown in the table

85

Work interest

Work interest for specific work tasks was measured with a scale that was developed for this specific work group. A similar work interest scale was developed for Study 1 with different work tasks. The KMO measure (.812) and
Bartlett 's test (p < .000) indicated that a factor analysis was appropriate for the

data. The unrotated principal components analysis showed that a three-factor solution accounted for 61 % of total explained variance, and therefore also in this study three factors were selected. A varimax rotation was performed to get a more interpretable solution. The result is shown in Table 8. Moreover, the first factor formed was called interest in administrative work tasks, the second one interest in client-related work tasks and the third one interest in business-related work tasks.

86

Table 8 A/actor analysis with a varimax rotation o/work interest among employees (Study 2)
Factor loadings
Administrative work tasks Client related work tasks Business related work tasks

To report To write letters, reports etc Accounting Administrative work tasks Management To solve conflicts To introduce new colleagues To collect information To work with changes within the organization To update handbooks and other materials To educate others To be responsible for payments to clients To market one 's company To give service and support within one 's department To solve problems in groups Meetings To teach others about products To be responsible for different work tasks To develop new ideas and products Further training To cooperate with others outside my own department Telephone contact with clients To use new techniques (e.g., computers) To travel for business purposes To present new ideas and products To participate in development projects To solve problems related to clients International contacts and cooperation To give service to clients To do business To negotiate To work with numbers To make money To create and increase networks To close a deal To make decisions

.83
.76 .79 .68 .66

.65
.64 .63

.62 .55
.51 .51 .49 .47

.47
.42 .72
.70

.69 .66 .63 .60 .52 .51 .50 .50 .49

.43
.42

.85
.80 .75 .74 .69 .68 .64

Note. Loadings below .40 are not shown in the table

87

Intrinsic motivation The result of a three-factor analysis of Ryan 's intrinsic motivation scale with the dimensions autonomy, relatedness, and competence is shown in Table 9. The KMO measure (.792) and Bartlett 's test (p < .000) showed that a factor analysis was appropriate to use. As in Study 1, several of the items were deleted, because the factor loadings were smaller than .40. However, the items deleted were the same as in Study 1, except for one item. This makes the scale quite stable over two different work groups. Table 9 Three-factor analysis ofRyan 's intrinsic motivation scale (Study 2)
Factor loadings
Autonomy Relatedness Competence

I feel like I can make a lot of input to deciding how my job gets done I an1 free to express my ideas and opinions 011 the job My feelings are taken into consideration at work There is not much opportunity for me to decide for myself how to go about my work (R) I really like the people I work with I get along with people at work I consider the people I work with to be my friends People at work care about me The people I work with do not seem to like me much There are not many people at work that I am close to (R) I do not feel very competent when I am at work (R) People at work tell me I am good at what I do I have been able to learn interesting new skills on n1Y job Most days I feel a sense of accomplishment from working On my job I do not get much of a chance to show how capable I am (R) When I am working I often do not feel very capable (R)

.55 .56 .40 .45

.67 .50 .63 57 .40 .41

.61
.58 .63 .69 .65
.65

Note. Loadings below .40 are not shown in the table.

88

Occupational stress Two factors were suggested by the developers of the Job stress survey (JSS; Turnage & Spielberger, 1991). The factors were confirmed in Study 1, which was also the case il1 this study. The KMO measure (.764) as well as Bartlett 's test (p < .000) were calculated and the result showed that a factor analysis was suitable for the data. The result is reported in Table 10.

Table 10 Two-factor analysis ofJob Stress Survey (Study 2)
Factor loadings
Lack of Support Lack of opportunity for advancement Poor or inadequate supervision Inadequate support by supervisor Personal insult from customer/consumer/colleague Fellow workers not doing their job Difficulty getting along with supervisor Inadequate salary Negative attitudes towards organization Lack of participation in policy-making decisions Poorly motivated co-workers Performing tasks not in job description Inadequate or poor quality equipment Insufficient personnel to handle assignment Dealing with crisis situations Frequent interruptions Meeting deadlines Excessive paperwork Assignment of new or unfamiliar duties Assignment of increased responsibility Covering work for another employee Working overtime Frequent changes from boring to demanding activities Making critical on-the-spot decisions Insufficient personal time .62 .64 .66 .54 .60 .44 .59 .62 .53 .54 .40 .72 .46 .48 Job pressure

.51
.56 .51 .57 .70 .52 .66 .72 .58 .42

Note. Loadings below .40 are not shown in the table.

89

General intrinsic and extrinsic motivation

Intrinsic and extrinsic motivation were measured with Amabile et aI 's. (1994) Work Preference Inventory. In Study 1, the two factors were confirmed. In this study, the two factors were once again confirmed with some of the items deleted due to factor loadings smaller than .40. A KMO measure (.792) and Bartlett 's test (p < .000) were also performed, which showed that the data were suitable for a factor analysis. The result is presented in Table 11.

Table 11 Two-factor analysis of intrinsic and extrinsic motivation (Study2)
Factor loadings Intrinsic Extrinsic motivation motivation .81 .74 73 .68
.73 .44 .48 .46 .54 .45 .50 .42 .47 .41

I enjoy tackling problems that are completely new to me I enjoy trying to solve complex problems The more difficult the problems, the more I enjoy trying to solve it I want my work to provide me with opportunities for increasing my knowledge and skills Curiosity is the driving force behind much of what I do I enjoy relatively simple, straightforward tasks I enjoy doing work that is so absorbing that I forget about everything else What matters most to me is enjoying what I do It is important for me to have an outlet for self-expression I prefer work I know I can do well over work that stretches my abilities It is important for me to be able to do what I most enjoy No matter what the outcome of a project is, I am satisfied if I feel I gained a new experience I want to find out how good I really can be at my work I 'm more comfortable when I can set my own goals I prefer to figure things out for myself I am strongly motivated by the (grades) (money) I can earn I am keenly aware of the goals I have for myself I am strongly motivated by the recognition I can earn from other people I want other people to find out how good I really can be at my work I seldom think about (grades and awards) (salary and promotion). I am keenly aware of the (goals I have for getting a good grade.) (income goals I have for myself.) To me, success means doing better than other peopIe I have to feel that I 'm earning something for what I do I am not that concerned about what other people think of my work I prefer having someone set clear goals for me in my work

.42
.50 .62 .67 .67 .59 .47 .68 .55
.61

.48

Note. Loadings below .40 are not shown in the table

90

Perceived risks and creativity
Perceived risks as well as creativity were divided, as in Study 1, into two indices. The indices that were forn1ed were accepted risks and risk burden. Creativity was divided into spontaneous creativity and creativity inhibition.

One dimensional scale
Pay satisfaction, perceived control, work environment, leadership, and opinion about one 's job were analyzed as one-dimensional as in Study 1. A factor not included in Study 1 was goal setting. In a study by Lee, Bobko, Earley, and Locke (1991), 10 factors were formed. However, in tl1is study goal setting is analyzed as if it was one-dimensional. A reliability analysis was performed for all the indices and is presented in the next chapter.

3.3.3.1.2 Conclusion

The KMO measure and Bartlett 's test showed that using a factor analysis was an appropriate way of reducing questions in all of the scales. Furthermore, the three dimensions of organizational commitment suggested by the developers were confirmed in both studies (Study 1 and 2). Work interest was measured with scales that were developed for the studies, and which measured the level of interest for different work tasks. Three factors were formed in Study 1 as well as in Study 2. Three dimensions of intrinsic motivation were suggested by the developer and also these dimensions were confirmed in both of the studies. The two-factor solution of organizational stress was found in Study 1 as well as in Study 2. Intrinsic motivation and extrinsic motivation could, as in previous studies, be separated and the results that were obtained here showed the same division. The findings in Study 1 and Study 2 confirmed the stability and validity of the instrument.

91

3.4 Study 3 (salespeople in an insurance company)
3.4.1 Participants
Data were collected in an insurance company 's sales department. The total number of salespeople working there was 62. The number of employees who participated in the study was 45, 12 women and 33 men, whose age ranged from 25 to 51 years (mean age = 38 years). The range of work experience within the profession was 1-26 years (average length of job experience
=

11 years). The

participants ' level of education ranged from high school to a university degree (modal educational level = three or four years of upper secondary high school).

3.4.2 Procedure
The questionnaires were distributed to each participant by the author, who also collected tl1em. Participation in the study was voluntary, and the respondents were informed of their rights to confidentiality and anonymity. There was no time limit on the participant 's con1pletion of the questionnaire. The total response rate was 73%.

3.4.3 Questionnaire
In this study, work motivation defined as willingness to work, job satisfaction and organizational commitment were also included in this study. However, organizational commitment was not divided into three dimension as in Study 1 and 2. The scales that were neither included in Study 1 nor 2 are presented further below.

92

3.4.4 Scales included
Job performance** was measured by two different scales. Objective job

performance was measured by the total sales volume (January 1 to May 31)4. Subjective job performance was measured by letting the respondents rate their performance in different work tasks on a scale developed by the author. The scale ranged from 5 (Very good) to 1 (Very poorly). The scales included in the study were all one-dimensional. Reliability analysis was performed for all for the indices and presented in Chapter 4.

4

The variable objective job performance was transformed into a logarithmic scale to obtain a less skewed variables.

93

DETERMINANTS
Overall job satisfaction Organizational commitment General work attitudes Opinions about one 's job Social relations Work environment Leadership Background factors Work interest
\0
~ lao. IJII '"

WORK MOTIVATION
(Willingness to work)

WORK-RELATED BEHAVIORS
--,...

..

Absenteeism Job performance Intention to quit Number of hours worked

Creativity Perceived risks Intrinsic motivation Intrinsic and extrinsic motivation Occupational and work stress Pay satisfaction Perceived control

Figure 3. An overview of the included factors in the thesis.

4. Results
In this section, results from all of the analyses performed are shown. The presentation of the analyses and results will follow the research questions in section 1.4. The first question was to explore the variables that contribute to explain the variance in work motivation defined as willingness to work. This was studied in Study 1 and 2. The second question was to investigate the validity of the measure willingness to work by examining the relationship between willingness to work and work-related behaviors. This was studied in Study 1, 2, and 3. In Figure 3, all of the indices included in the analyses are presented.

4.1 Descriptive results
4.1.1 Introduction
For all of the indices, the reliability test was estimated. Reliability refers to the consistency of examinees ' relative performances over repeated administrations of the same test or parallel fOffi1s of the test. One common way of estimating reliability is by calculating a Cronbach 's alpha (a), which is a function of internal consistency, that is, the interrelatedness of iten1s (Cortina, 1993). The level of Cronbach 's alpha has often been discussed and the most common level to be accepted is an alpha equal to or above .70 (Cortina, 1993; Peterson, 1994). It is often discussed in the literature how the number of items in a scale affect the reliability. The principle of aggregation holds that the use of multiple observations cancels out random error around an individual 's true score, thereby providing a n10re reliable measurement. Despite the well-known attenuating effects of low reliability on correlation size, researchers often use single-item measures so as to minimize questionnaire length (e.g., Tett & Meyer, 1993).

95

This practice has the obvious consequence of underestimating a relation of interest, and could lead to serious misjudgements in the relative contributions of two variables whose measures differ markedly in length (Cooper & Richardson, 1986). However, in an article by Embretson (1996) "the new rules of measurements" were presented and it was stated that short tests could be more reliable than longer tests. For a further discussion of reliability, see for example Crocker and Algina (1986).

4.1.2 Study 1: Descriptive results
111

Table 12, descriptive statistics including mean, standard deviation, and

reliability of index are reported. The majority of the indices showed a satisfactory reliability value (a) equal to or over .70. In two of the indices, risk burden and intention to quit, the magnitude of reliability (a) was lower than .70. However, Schmitt (1996) stated that there is no sacred level of acceptance level of alpha. In some cases, measures with (by conventional standards) low levels of alpha may still be quite useful. Thus, all of the indices were used in further analysis.

96

Table 12 Descriptive statistics and reliability ofindices (Study 1)
Variables Work motivation Job satisfaction Affective commitment Normative commitment Continuance commitment Social relation Work environment Work interest (pedagogical) Work interest (practical) Work interest (administrative) Management/leadership Evaluation of one 's job (positive) Evaluation of one 's job (negative) Spontaneous creativity Creativity inhibition Accepted risks Risk burden Job pressure Lack of support Intrinsic n10tivation Extrinsic motivation Pay satisfaction Opinion about one 's job Perceived competence Perceived autonomy Perceived relatedness Perceived control Number of hours worked Absenteeism Intention to quit
M SD

Cronbach alpha (a) .79 N/A a .85 .73 .72 .82 .82 .90 .85 .84 .75 .83 .89 .71 .75 .85 .65 .86 .83 .82 .71 .95 .73 .70 .74 .79 .88 N/A a N/A a .68

3.50 5.48 3.46 3.48 2.96 4.37 3.53 5.48 4.91 4.42 3.89 4.54 1.67 3.63 2.71 2.78 3,19 4.78 5.37 4.05 3.61 2.38 4.65 4.53 4.52 4.51 3.52 3.87 2.50 2.70

.49 .81 .99 .87 .96 .69 .53 .81 .83 1.39 .79 .95 .67 .45 .56 .43 .48 .97 .65 .59 .60 .73 .66 .80 .85 1.2 .42 1.4 .58 .92

a Alpha cannot be assessed for single-item.

97

4.1.3 Study 2: Descriptive results
In Table 13, descriptive statistics are shown for the indices included in Study 2. All of the indices except risks burden had an a value equal to or large than .70. Risk burden was, despite the value of a, included in further analysis.

Table 13 Descriptive statistics and reliability ofindices (Study 2)
Variables/ Indices Work motivation Job satisfaction Affective commitment Normative commitment Continuance commitment Social relation Work environment Work interest (Business) Work interest (Clients) Work interest (Administrative) General work interest Management/leadership Goal setting Spontaneous creativity Creativity inhibition Accepted risks d Risk burdene Job pressure Lack of support Intrinsic motivation Extrinsic motivation Pay satisfaction Opinion about one 's job Perceived competence Perceived autonomy Perceived relatedness Perceived control Absenteeism Number of hours worked Intention to quit a Alpha

M
3.26 5.03 3.04 2.64 3.48 4.20 3.83 5.13 5.20 4.67 3.26 3.90 3.00 3.40 2.80 2.90 2.75 5.15

SD
.64 1.14 1.01 .77 .85 .76 .64 .90 .77

Cronbach alpha (a) .80 N/Aa .91 .73 .72 .70 .86 .84 .85 .90 N/A a .78 .87 .74 .58 .79 .50 .88 .83 .83 .72 .97 .92 .71 .71 .83

.81
1.13 .87

.42
.49 .79

.45
.62

.81
.62

5.27
4.18

.56
.74 .80 .86 .72

3.47
3.12 4.17 4.33 4.13 4.66 3.50 4.41 2.69 2.96

.75
.68

.47
1.1 .94

.88
N/A a N/A a .71

1.2

cannot be assessed for single-item..

98

4.1.4 Study 3: Descriptive results
In Study 3, the a values of all the included indices were satisfactory. The result is presented in Table 14.

Table 14 Descriptive statistics and reliability ofindices (Study 3).
Variable Willingness to work Job satisfaction Organizational commitment Work interest Subjective job performance Objective job performance

M
3.95 5.67 2.90 5.30 4.04 15.87

SD
.47 .60 .59 .72 .45 1.2

Cronbach alpha (a) .77 N/A a .83 .85 .90

a Alpha

cannot be assessed for a single item

99

4.2 Determinants of willingness to work
4.2.1 Introduction
In Chapter 2, an overview of factors that in prior studies have shown to be related to work motivation was presented. Willingness to work was introduced by Sjoberg and Lind (1994) and in their study, the factors that explained the variance in willingness to work were work interest, perceived risks, creativity, and organizational commitment. Therefore, these factors were expected to contribute to explain the variance in willingness to work. In the present studies, a large nun1ber of indices were included. In order to explore which of the independent variables that accounted for explaining the variance in willingness to work, a stepwise multiple regression was performed. The specific procedure employed was backward elimination or a decremental approach of stepwise regression, as Bobko (1990) named it. The backward solution starts with calculating the total multiple regression equation, then drops variables from the equation if the variables do not contribute significantly to the prediction of the criterion, here willingness to work, based on the significance tests of their beta weights. In this procedure, the multiple correlation coefficient is reduced as each variable is removed. When enough variables are removed, a significant drop occurs in the magnitude of R2 . At this point, the stepwise procedure - dropping a variable at each step and recomputing R2 - is terminated. Only variables that are theoretically meaningful and significant at the 10%-level or lower were included in the final model. The results are presented below. It should be mentioned that in order to use a multiple regression analysis several assumptions have to be met. Therefore, the present data (in Study 1 and 2) were tested for outliers, n1ulti-collinearity, normality and heteroscedasticity (Berry, 1993). No severe problem was found in either of the studies 5 •
The result can be obtained from the author. 100

5

4.2.1.1 Study 1: Determinants of willingness to work
In this study, 179 pre-school employees participated. The factors or indices included from the start were job satisfaction, organizational commitment (affective, normative, and continuance), general work attitudes (positive and negative evaluation of one 's job), opinion about one 's job, social relations, physical work environment, work interest (pedagogical, practical, and administrative work tasks), leadership, creativity (spontaneous creativity and creativity inhibition), perceived risks (accepted risks and risk burden), intrinsic motivation (perceived competence, perceived relatedness, and perceived autonomy), general intrinsic motivation, extrinsic motivation, work stress Gob pressure and lack of support), pay satisfaction, and perceived control. The intercorrelations among these variables are shown in Table 15.

101

Table 15. Correlation matrix of all included variables (Study 1)
2 1.Willingness to work 2.Job satisfaction 3.Education 4.Age 5.Tenure 6.Accepted risks 7.Risk burden 8.Spontaneous creativity 9.Creativity inhibition 10.Social relations II.Perceived control 12.Work environment 13.Pay satisfaction 14.Affective OC 15.Continuance OC 16.NormativeOC 17.Perceivedautonomy 18.Perceived relatedness 19.Perceived competence 20.Jobperssure 21.Lack of support 22.Interest in pedagogical tasks 23.Interest in practical tasks 24.Interest in administrative tasks 25.Negative evaluation 26.Beliefs about one 's job 27.Extrinsicmotivation 28.Positive evaluation 29.Intrinsic motivation 30.Management/leadership 048 .06 -.08 .09 042 -.09 .25 -.19 .30 042 .34 .22 .60 .15 .36 043 Al .52 .22 .23 046 .58 .22 -043 .56 .02 .61 Al .49 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

o

........

tv

.09 .03 -.03 046 .03 .11 -.16 Al 040 046 .27 046 .06 .23 .38 .36 AS Al 048 .20 Al -.02 -.33 048 .17 .51 .12 .28

.03 .05 .09 -.12 -.18 -.04 -.05 .09 -.02 .07 .05 -.11 .02 .05 .09 .13 .16 .12 -.16 -.11 .04 -.09 .08 -.01 .01 .04 -.17

.27 -.02 .18 -.11 .19 -.06 .03 .04 .10 .06 .03 .16 -.12 -.16 -.14 .09 .00 -.25 -.09 -.19 .06 -.06 -.05 -.02 .31 -.08

-.04 .27 -.25 .09 -.02 -.18 -.04 .04 -.07 -.13 .07 -.07 .02 -.04 .03 .02 -.08 -.21 -.15 .04 -.10 .05 -.03 .16 -.14

.35 -.12 -.14 043 .52 042 .36 .38 .03 .19 040 Al AS .56 .62 .17 Al .05 -.33 045 .22 043 -.02 .22

-.28 .13 .16 .04 -.02 .06 .01 -.03 .09 .03 .08 -.05 .20 .21 .06 .00 -.18 -.03 .02 .11 .10 -.23 -.06

-.18 .08 .07 -.05 -.18 .11 -.06 .10 .18 .08 .16 -.14 -.16 .35 .33 .32 -.06 .19 -.05 .16 .49 .32

-.08 -.25 -.18 -.11 -.06 .22 .19 -.23 -.27 -.33 -.07 -.05 -.23 -.21 -.16 .26 -.26 -.34 -.12 -.31 -.13

.23 .22 .10 .28 -.13 .10 .38 .59 .30 .18 046 .13 .22 .03 -.19 .21 .15 .28 .13 -.15

042 .29 Al .06 .13 049 .29 .39 .35 .33 .19 Al .16 -.28 049 .05 045 .21 -.30

.32 .29 .03 .13 .15 .19 .29 047 .38 .13 040 .03 -.22 .38 -.10 .37 -.09 .19

.31 .06 .13 .15 .15 .20 .34 042 -.06 .13 .03 -.18 .2 .14 .22 -.20 .03

Al .57 AS .34 .35 .23 .34 .21 Al .06 -.22 .54 .07 048 .20 .34

.51 .04 -.09 -.05 -.07 -.05 -.07 .08 -.04 .08 .11 -.28 .04 -.12 .11

.16 -.01 .01 .07 .07 .08 .24 .13 -.04 .29 -.30 .31 .09 .20

.59 AS .10 .28 .32 040 .08 -.31 .50 .09 .57 .39 040

.57 .12 .30 .27 .26 .04 -049 .39 .22 Al .21 .27

.25 .32 .31 043 .09 -046 .58 .16 047 .28 .40

.73 -.04 .25 -.01 -.13 .25 .16 .27 -.08 -.03

.00 .21 -.01 -.20 .30 .27 .29 -.09 .03

.59 .37 -.37 042 .01 047 .46 .65

.39 -.35 049 .02 .57 .34 .80

-.07 .16 -.17 .14 .39 .23

-043 -.18 -044 -.17 -.39

.09 .60 .00 .30 .18 .34 .43 .02 .51 -.36

Note. DC

=

organizational commitment

After perfoffi1ing a backward elimination procedure 6 the variance in willingness to work was explained to 59%. The factors that to the highest extent contributed to explaining the variance in willingness to work were, affective organizational commitment, positive evaluation of one 's job, interest in pedagogical work tasks, accepted risks, risk burden, al1d perceived competence. General intrinsic motivation and normative organizational commitment were also factors contributing to explain the variance in willingness to work. The result is presented in Table 16.

Table 16 The multiple regression model with willingness to work as the dependent variable and its explanatory variables. N== 179
Explanatory variables b
13

Affective organizational commitment Positive evaluation of the job Interest in pedagogical work tasks Accepted risks Risk burden Perceived competence General intrinsic motivation Normative organizational commitment
2

.15 .17 .14
.23

.26** .20** .18** .17** -.16** .16** .12* .11 *

-.19
.12

.12
.01

R

ad .f

0.59

* p ::; .05.

**p ~ .01.

6

Other stepwise procedures were performed and similar results were obtained.
103

4.2.1.1.1 Conclusion

The main point of this study was to explore the explanatory factors of work motivation, defined as willingness to work, for pre-school employees. The variance in willingness to work was explained to 59%, which is a quite high percentage of explained variance. In fact, there is only limited room for further improvement here, because the dependent variable is also affected by reliability deficiencies. The explanatory factors were the following: Affective commitment, positive evaluation of one 's job, interest in pedagogical work tasks, risk burden, accepted risk, perceived competence, general intrinsic motivation, and normative organizational commitment. Sjoberg and Lind (1994), who introduced the work motivation measure used, found that similar factors explained willingness to work. An explanatory factor that explained quite a large portion of the variance in willingness to work was affective organizational commitment. According to the result, the more an employee feels affectionately attached to the organization, the higher the motivational level is. In a previous study, affective organizational commitment was shown to be quite strongly associated with different work related behaviors, like job performance (Meyer & Allen, 1997), absenteeism, and turnover (Somers, 1995). Normative organizational commitment also contributed to explaining the variance in willingness to work. The results showed that the more an employee feels obligated to the organization, the more motivated he or she is. The explanatory variable, POSitive evaluation of one 's job, which reflects general attitudes towards OI1e 's job, was also a relatively strong explanatory factor to explain willingness to work. This indicates that the more positive an individual is towards his or her job, the higher the motivational level is. This is in line with what Katzell and Thompson (1990) concluded. However, in general, the relationship between work attitudes and behavioral variables has been found to be weak (for a review, see Andrich & Styles, 1998).

104

Interest in pedagogical work tasks was another explanatory variable of willingness to work. The conclusion to be drawn here is that when people work with tasks they find interesting, it affects their motivation in a positive way. In this case, many of the employees in pre-schools probably might choose this particular job because they are interested in children and pedagogical work tasks, such as reading to the children and teaching children different things. In the study of Sjoberg and Lind (1994), work interest was the strongest explanatory variable of work motivation. Both groups of risks, accepted risks and risk burden, contributed to explaining willingness to work. The fact that risks can both increase and decrease the motivation level might be a little bit confusing. Normally, risks are associated with something negative at work, such as giviI1g colleagues negative critique, and when an individual is exposed to that, the motivational level will decrease. It has, however, been found that some high risk situations might instead be perceived as a challenge (Svensson, Angelborg-Thanderz, & Sjoberg, 1993). Another variable that contributed to explaining willingness to vv 'ork \-vas perceived competence (for a review, see Deci & Ryan, 1985). This study shows that willingness to work increased with improved perceived competence. Competence refers to knowledge and skills which people have at their disposal and which they can use efficiently to reach certain goals in a wide variety of contexts or situations (Kirschner & Van Vilsteren, 1997). Not only the dimension of intrinsic motivation suggested by Deci and Ryan, (1985), perceived competence contributed to explaining the variance in willingness to work but also intrinsic n10tivation introduced by Amabile, Hill, and Hennessey (1994), which measured intrinsic motivation more in general. The result indicates that the more intrinsically motivated in whole a person is the n10re willing to work he or she is. In a previous study, it was found that if the intrinsic needs are met in a workplace, this will lead to increased job satisfaction and improved job performance (Ilardi, Leone,& Ryan, 1993).

105

To sum up, the result of the study showed that in order to explain willingness to work among pre-school employees and most likely in other professions and organizations as well, a variety of factors must be considered. I believe that to increase the understanding and prediction of work motivation, it might be helpful to take into account a variety of factors, which has been demonstrated in this explorative study.

4.2.1.2 Study 2: Determinants of willingness to work
In this study, 160 employees of an insurance company participated. The variables included in the analysis from the beginning were, job satisfaction, organizational commitment (affective, normative, and continuance), goal setting, opinion about one 's job, social relations, physical work environment, work interest (client, business, and administrative work tasks), leadership, creativity (spontaneous creativity and creativity inhibition), perceived risks (accepted risks and risk burden), intrinsic motivation (perceived competence, perceived relatedness, and perceived autonomy), gel1eral intrinsic n10tivation, extrinsic motivation, work stress Gob pressure and lack of support), pay satisfaction, and perceived control. The intercorrelations among these variables are shown in Table 17.

106

Table 17. Correlation matrix of all included variables (Study 2)
2 1.Willingness to work 2.Job satisfaction 3.Affective OC 4.Normative OC 5.Continuance OC 6.Social relations 7.Environment 8.Interest (Business) 9.Interets (Clients) 10.Interest (Administrative) 11.General interest 12.Ledership/management 13.Goal setting 14.Spontaneous creativity 15.Creativity inhibition 16.Accepted risks 17.Risk burden 18.Job pressure 19.Lack of support 20.Intrinsic motivation 21.Extrinsic motivation 22.Pay satisfaction 23.Beliefs about one 's job 24.Percieved competence 25.Perceived autonomy 26.Perecived relatedness 27.Percieved control 28.Age 29.Education 30.Tenure .64 .50 .21 -.12 .41 043 .40 .26 048 .54 .52 .35 .34 -.16 .51 -.15 .27 .33 .14 .20 .28 .65 .52 .46 .23
Al

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20 21

22

23

24

25

26

27 28

29

30

I-.l

0 -....l

.09 -.11 .11

043 .25 -.11 .51 AD .25 .11 .36 .53 .55 .41 .13 -.08 .54 -.08 .23 .45 .09 .13 .29 .58 .48 .35 .26 .39 .10 -.11 .02

.65 .29 .35 .31 .09 .13 045 .36 .38 .40 .08 -.03 .27 -.09 .12 .20 .01 .16 .38 .50 .27 .18 .24 .25 .18 .10 .34

.58 .16 .19 .12 .07 .23 .24 .18 .35 -.02 -.06 .07 -.05 .09 .10 .06 .22 .15 .13 .06 .13 .07 .05 -.16 .05 -.27

-.10 -.06 -.21 .13 -.05 -.06 -.09 .12 -.14 .19 -.33 -.17 -.13 -.15 .15 .12 -.01 -.16 -.34 -.30 -.13 -.15 -.09 -.09 -.17

.34 .10 .01 .20 .25 .51 .38 .02 -.12 047 -.20 .21 047 .11 .06 .23 .35 042 .23 .57 .31 .07 -.16 -.04

.25 .10 .26 .26 .32 .15 .04 -.01 Al -.09 .23 .32 .10 .22 .26 .36 .17 .12 .20 .32 .07 .08 -.01

.64 .51 .23 .24 .13 042 -.30 .33 -.06 .16 .05 .49 .25 .11 .21 .32 .16 .05 .21 .02 -.01 .15

.60 .24 .11 .07 048 -.21 .18 -.02 .01 .13 .55 .28 .12 .16 .25 .14 .01 .15 -.02 -.07 -.01

.37 .34 .27 .25 -.15 .32 -.11 .18 .11 .29 .14 .30 .34 .32 .30 .23 .34 .06 -.11 .12

.37 .29 .20 -.02 .34 .17 -.22 .17 .16 .26 .17 .37 .32 .25 .10 .26 .11 -.10 -.03

.51 .11 -.13 043 .10 -.32 -.51 .06 -.04 .39 -.41 -046 -.41 -.22 044 .04 .01 -.04

.01 -.07 .19 -.11 .23 .37 .07 .13 .38 .23 .24 .17 .14 .26 -.02 -.05 -.05

-.15 .15 -.21 -.11 -.18 .52 .13 .03 .29 .23 .33 .05 -.22 .09 -.10 .06

-.17 -.05 -.06 -.06 -.16 -.03 -.05 -.12 -040 -.21 .09 -.16 .06 .10 .02

.35 .53 .53 .13 .02 .37 .56 .58 049 .37 .51 .13 -.16 .00

.39 .29 .04 .06 .01 .09 .10 .05 .13 .01 .12 .22 -.08

.65 .04 .02 .21 .13 .23 .30 .04 .33 .14 -.04 .06

.16 .22 042 .28 .35 .27 .33 040 .06 .02 .02

.33 .09 .22 .25 .21 .03 .14 .06 .06 .09

.12 .15 .06 .03 .10 .03 -.14 .00 .08

.24 .25 .29 .23 042 .03 .01 .00

.56 .58 .34 .54 -.07 .25 -.20

.61 043 049 -.02 .18 .04

.35 .64 .04 .11 -.08

.24 .03 .00 .09 -.18 .07 .10 .01 .21 .09

Note. OC

=

organizational commitment

Willingness to work was explained to 690/0 by a number of explanatory variables. As in Study 1, the explanatory variables were selected after performing a backward elimination procedure
7.

The factors tl1at accounted for

the highest explanatory values were, first of all, beliefs abollt one 's job and perceived competence, which were significant at the 1% level. Job satisfaction, risk burden, spontaneous creativity, work interest (business related work tasks) and affective organizational commitment were significant at the 5% level, whereas development of general work interest and perceived work environment were significant at the 10% level. The results of the analysis are presented in Table 18.

Table 18 Regression model obtained after performing a backward elimination procedure with willingness to work as the dependent variable, N== 160
Explanatory variables b
.30*** .20*** .16** -.16** .13** .13**
.12** .10* .10*

Opinion about one 's job Perceived competence Job satisfaction Risk burden Spontaneous creativity Interest in business related work tasks Affective organizational commitment General development of work interest Perceived work environment
2

.22

.16
.09
-.14

.18 .09 .08 .08
.10

R

adj

.69

*p ~ .10, ** p ~ .05, *** p ~ .01.

7

Other stepwise procedures were performed and similar results were obtained.

108

The results show that as in Study 1, attitudes about one 's job (opinion about one 's job), perceived competence, perceived risks (risk burden), work interest and affective organizational commitment contributed to explaining the variance in willingness to work. The variance in willingness to work was also in this study explained to a high extent.

4.2.1.2.1 Conclusion

The result showed that the expla11atory variables that accounted for the variance in willingness to work were the following: Opinion about one 's job, perceived competence, job satisfaction, risk burden (risks negatively related to work motivation), spontaneous creativity, work interest (in business related work tasks), affective organizational commitment, physical work environment, and increased general work interest over the years. The variance in willingness to work was explained to 69% by these explanatory factors. Thus, the level of explanation of the dependent variable, willingness to work, was high. There is only limited room for future improvement here, because the dependent variable is affected by reliability deficiencies. It seems like an individual 's beliefs or opinion about the job influence his or her willingness to work a great deal. In previous studies, it has been argued that beliefs directly affect intentions (Weber & Gillespie, 1998). Montgomery (1998) stated that if the strength of a behavioral intention is equivalent to volitional strength, then it follows from reasoned action theory (for a review, see Fishbein & Ajzen, 1975) that volitional strength is dependent on the individual 's beliefs. It should be noted that in a great number of studies, Ajzen and Fishbein have demonstrated a fairly close relationship between behavioral intentions and actual behavior (see e.g., Fishbein & Ajzen, 1975; Ajzen & Fishbein, 1980). Beliefs have been found to affect attitudes. Like work attitudes, beliefs are difficult to influence or change. One way might be to introduce new positive information about the job (Pinder, 1998).

109

As in Study 1, the results of this study showed that the more competence an employee perceives him- or herself to have at work or for a specific work task, the higher the motivational level is. It has been found that if the perceived competence is fulfilled in the workplace, this will result in increased job satisfaction (Ilardi, Leone, Kasser, & Ryan, 1993). Moreover, ifpeople experience unpleasant events (risk burden) and situations or if the workplace is perceived to be dangerous, this has shown to affect people negatively, not only in this study but in others as well (e.g., Roberts, 1993). If the work environment instead is perceived to be pleasant, this will affect motivation positively, as the result of the study showed. To let people come up with their own ideas and give them support in their creative processes was an important factor in explaining willingness to work. This supported the findings of Tollgerdt-Andersson and Sjoberg (1992).

In Study 1, it was found that work interest had a significant impact on work motivation. In this study, work interest, both general work interest as well as interest for specific work tasks (business), were important determinants of willingness to work. Sjoberg and Lind (1994) have reported the same and this indicates that work interest is an important variable to explain willingness to work Affective organizational commitment, another variable which was reported to have a significant impact on willingness to work in Study 1, was shown to be the dimension of the three organizational commitment dimensions suggested by Allen and Meyer (1990) that has the strongest relation with willingness to work also in this study. In previous studies, a quite strong association has been found between affective commitment and job performance (Meyer & Allen, 1997). Affective commitment has also been found to have an impact on turnover and absenteeism (Somers, 1995).

110

As the result showed, willingness to work was explained with a variety of explanatory factors. In the present study, variables of intrinsic values, such as perceived competence, spontaneous creativity, and work interest stand for a quite large amount of explanatory power together with job attitudes, like job satisfaction, affective commitment, and opinion about one 'sjob. There were two variables that represent the environment, and tl1ese were perceived physical work environment and risk burden. This reflects the complexity of work motivation and shows that it is important to integrate numerous variables in order to understand and predict work motivation.

4.2.2 A structural model of willingness to work

4.2.2.1 Introduction

As the results showed in the two previous studies, based on the data from preschool employees as well as employees in an insurance company, several of the variables, such as work interest, perceived risks, affective organizational conunitment, and perceived competence contributed to explaining the variance in willingness to work in both of the groups. All of the mentioned variables have been shown to have strong direct relationships to willingness to work, as shown in the two previous studies. Furthermore, in previous studies direct relationships between some of the variables included, such as work interest, has been shown to have a positive relationship with organizational comn1itment (Winer & Gati, 1986). A study of Cury, Biddle, Goudas, Sarrazin, and Durand (1996) found that perceived competence influenced students ' interests. It should be mentioned that the dimension of perceived risks included for further analysis was accepted risks, although risk burden was the dimension that contributed significantly to explaining willingness to work.

111

However, the decision was made to include accepted risks due to the results presented in the correlation matrices in Tables 15 and 17, because this factor was more strongly related to willingness to work than risk burden. The purpose of this study was to examine the indirect as well as direct relationships between willingness to work and the explanatory variables mentioned above. Participants are the same as in the two previous studies, preschool employees and employees of an insurance company. The pllrpose of including a second sample in this article was to examine the stability of the revised model for Study 1. In order to examine the direct as well as indirect relations in a model, structural equation model LISREL (Joreskog & Sorbom, 1996) was used. A Structural Equation Model (SEM) can be described as a multivariate technique combining aspects of multiple regression (examines dependence relationships) and factor analysis (representing unmeasured concepts factors with multiple variables) to estimate a series of interrelated dependence relationships simultaneously (Hair, Andersson, Tatham, & Black, 1998). Moreover, an additional characteristic of SEM techniques is its ability to represent unobserved concepts, latent variables, in these relationships and account for measurement errors in the estimation process (Hair et aI., 1998). A latent variable is a hypothesized and unobserved concept that can only be approximated by observable or measurable variables. The observed variables are known as manifest variables, which can be obtained through e.g., surveys, tests etc. The most obvious difference between SEM and other multivariate techniques is the use of separate relationships for each of a set of dependent variables. The characteristic of LISREL is that the entire LISREL model can be expressed in temlS of eight matrices, two defining the structural equations, two definil1g the correspondence of indicators and constructs, one for the correlation of exogenous ("independent") constructs, one for the correlation of endogenous ("dependent") constructs, and finally two detailing the correlated errors for the measurement of exogenous and endogenous variables (Hair et aI., 1998).

112

On the basis of the literature reviewed, the hypothesized model is presented in Figure 4.

+

+
Work interest

+

Organizational commitment (Affective)

.4~

+
Accepted risks

+

..+

W illingness to work

Perceived competence

Figure 4 . A hypothesized work motivation model

4.2.2.2 Study 1: Structural model of willingness to work
Table 19 reports intercorrelations for the indices. The underlying observed variables and their factor loadings on respective latent variables are presented for willingness to work and accepted risks in Appendix C1, affective organizational commitment in Table 1, work interest in Table 3, and perceived competence in Table 4.

113

Table 19 Intercorrelationsfor the complete set ofindices (Study 1)
1
1. Willingness to work 2. Affective organizational commitment 3. Perceived competence .59

2

3

4

.51

.36

4. Work interest (pedagogical tasks) 5. Accepted risks

.46

.20

.33

.42

.37

.43

.21

Note. All correlations are significant at p < .01 at least.

The correlations between willingness to work and the other variables are quite high, as expected. Many of the factors are rather highly correlated like accepted risks and perceived competence as well as accepted risks and affective commitment. Perceived competence is quite strongly correlated to both affective commitment and work interest.

4.2.2.2.1 Model estimation and modification

Estimation methods in structural equation modelling are developed under various assumptions. One is the assumptio11 that estimates and tests are based on large samples. The sample size (N = 179) in this study could be considered to be small. The adequacy of the test statistics is thus likely to be influenced by the sample size (Hu & Bentler, 1998). Therefore, the following fit indices were chosen because, according to Hu and Bentler (1998), they are less affected by sample size: Standardized Root Mean Square Residual (SRMR), Goodness of Fit Index (GFI), Comparative Fit Index (CFI), and Incremental Fit Index (IFI).

114

The model presented in Figure 4 was estimated using LISREL 8.30 (Joreskog & Sorbom, 1999) with the sample covariance matrix as input. The results of estimating the hypothesized model were, X2 (4) = 30.4, SRMR == .14, GFI = .94, CFI = .86 and IFI = .87. The result is shown in Table 20.

The

fit

indices

suggested potential

for

improvement through model

modification. LISREL modification indices suggested, firstly, to add a path linking accepted risks to affective organizational commitment, secondly to add a path between perceived competence and affective organizational commitment. The path between accepted risks and willingness to work was not significant and was deleted. This improved the model fit to X2 (4) = 4.55, p = .34, SRMR = .04, GFI = .99, CFI = 1.00, and IFI = 1.00. The revised model fits the data better than the initial model. The results of the revised model analyses are shown in Figure 5. As expected, work interest, intrinsic motivation, and affective organizational commitment had a significant impact on willingness to work, whereas accepted risks did not. As predicted, a positive and significant path was found between perceived competence and work interest as well as between work interest (pedagogical work tasks) and organizational commitment (affective). In the revised model the path between work interest and affective organizational commitment was not significant and was deleted. Approximately 52% of the variance in willingness to work, 20% of the variance in affective organizational commitn1ent and 11 % of the variance in work interest (in pedagogical tasks) was explained by the model.

115

Table 20 Hypothesized model (Study 1)
Structural paths
Unstandardized coefficient t-value a

Affective OC

~

WTW

.25 .22
WTW

7.93 5.18 3.85 1.75 2.75 4.60

Work interest ~ WTW Perceived competence Accepted risks
~ ~

.17 .13 .27 .32

WTW

Work interest ~ Affective OC Perceived competence
~

Work interest

a

t-values greater than 2.0 are significant. DC = Organizational commitment, WTW = Willingness to work, Work interest = Interest in pedagogical work tasks.

.20 (4.50)

Perceived competence

.32 (4.60)

Work interest (pedagogical tasks)

.22
.30 (3.15)

(5.2~ ..-----

-,

Willingness to work

...--.66 (3.88)
Accepted risks
~

.26_{---,7,:Y

Affective organizational commitment

Figure 5. Revised model of willingness to work for Study 1. Unstandardized coefficients and t-values within parantheses are presented.

For this sample the fit indices suggest a good fit of the data to the revised model.

116

4.2.2.3 Study 2: Structural model of willingness to work
In Table 21, intercorrelations of the idices included in the analysis are shown. Factor loadings on respective latent variables are presented for willingness to work and accepted risks in Appendix C2, for affective organizational commitment in Table 7, perceived competence in Table 9, and work interest in Table 8. Table 21 Intercorrelations ofthe indices included in the model (Study 2)
1
1. Willingness to work 2. Affective organizational commitment 3 Perceived competence 4. Work interest tasks) 5. Accepted risks (business .52

2

3

4

.58 .45

.26 .09 .32

.59

.29

.59

.34

Note. Correlations above r= .20 are significant on (p < .01).

All of the variables were highly correlated to willingness to work, which was expected. In addition, the relationship between perceived competence and accepted risks was very high. The rest of the relationships were quite high, around .30, except for the relationship between work interest and affective organizational commitment.

4.2.2.3.1 Model estimation

The revised model presented in Figure 6 was estimated using LISREL 8.30 with the sample covariance n1atrix as input. The results of estin1ating the revised

117

model are presented in Figure 6. Estimation of the revised model resulted in X2(df= 3) = 12.51, P = 0.01, SRMR = 0.05, GFI = 0.97, eFI = 0.95, and IFI = 0.95. Approximately 49% of the variance in willingness to work, 9% of the variance in affective organizational commitment, and 10% of the variance in work interest (in business related work tasks) was explained. Furthermore, the revised model fit the data quite well. Most of the paths were significant, except the paths between perceived competence and work interest and perceived competence and affective organizational commitment.

.31 (5.95)

Perceived competence

.41(4.28)

Work interest (business related tasks)

.18 (4.31) .23 (1.71)
Willingness to work

.23 (6.37) .41 (1.89)
Accepted risks

•

Affective organizational commitment

Figure 6. Revised model of willingness to work for Study 2. Unstandardized coefficients and t-values within parantheses are presented.

118

4.2.2.4 Conclusion
In Study 1, a positive relationship was observed between affective

organizational conunitment, work interest in pedagogical work tasks, accepted risks, perceived con1petence and willingness to work. The hypothesized relationship between perceived competence and work interest in pedagogical work tasks was significantly positive. The relationship betwee11 work interest and affective orga11izational commitment was significant in the hypothesized model but not in the revised model and was therefore deleted.

One relationship that was expected to be significant, the one between accepted risks and willingness to work, was not supported in this study and deleted. The link that was suggested to be added in the model was the one between accepted risks and affective organizational commitment. Another added link was the one between perceived competence and affective organizational commitment.

In study 2, which was used to confirm or examine the stability of the model, the fit indices showed a good fit of the data. In this second study, most of the paths were supported. There were some paths that were not supported and those were the relationships between work interest and affective organizational commitment and between perceived competence and affective organizational commitment.

As the results indicate, the level of willingness to work is increased when employees find their work task interesting. This supports previous findings by Sjoberg and Lind (1994) and Tollgerdt-Andersson and Sjoberg (1992). The more attachment to the organization an employee experiences, the more motivated he or she is. A person who perceives him/herself as competent at work will experience higher work motivation than a person who does not perceive him/herself as having competence. Deci and Ryan (1987) found that the higher the perceived competence, the greater the job satisfaction and job performance.

119

In the model, perceived competence was a factor that had impact not only on willingness to work but also on many of the other variables in the analyses. Consistent with the theoretical proposition of perceived competence, Cury, Biddle, Famose, Goudas, Sarrazin, and Durand (1996) fOllnd that it was an important factor for determining interest. However, as Sansone (1986) stated, competence perceptions are only important to interest if con1petence itself is valued in that context. In the revised model, it was shown that the more the risks are accepted by an employee the more affectively committed to the organization he or she is. The perception of one 's competence has a positive impact also on affective organizational commitment. To sum up, the results showed that the indices of fit suggest a good fit of the data to the revised model in both of the studies. This gives indications of a quite stable model. relationships. It also supports tentative conclusions regarding causal

4.3 Results of construct validity studies
4.3.1 Introduction
The purpose of this section was to examine the construct validity of willingness to work, which was investigated in three groups, pre-school employees (Study 1), employees in an insurance con1pany (Study 2), and salespeople in an insurance company (Study 3). Validity is the degree to which an instrllment actually measured what it was supposed to measure. It is always relevant to question the validity of an instrument. It is particularly relevant when new instruments are being developed or when previously validated instruments are translated and used in other countries (e.g., Lincoln & Kalleberg, 1990).

120

4.3.1.1 The relationship between willingness to work and workrelated behaviors
Work motivation has been defined as the process by which behavior is energized, directed, and sustained in organizational settings (Steers & Porter, 1991). Therefore, a goal of work motivation theories is to be able to explain the factors or processes that affect the direction, intensity and persistence of behavior in the workplace. Behaviors that have often been examined are withdrawal behaviors, such as absenteeism and turnover, because of the costs for organizations (Steers & Porter, 1983). In the present study, intention to quit is investigated. In prior studies, turnover intention has been found to be an immediate precursor to actual turnover (e.g., Hom, Caranikas-Walker, Prussia & Griffeth, 1992). A work-related behavior that reflects how hard an individual works is the number of hours (each day, week etc) he or she puts into work. The assumption here is that the more time an individual voluntarily puts into work, the higher the motivational level is. Wallace (1997) fOllnd a positive association between work motivation and number of hours worked. A positive relationship has been found between job performance and work motivation (for a review, see Pinder, 1984; Locke & Latham, 1990a; Locke & Henne, 1986; Maeher & Braskamp, 1986).

In Study 1, 179 pre-school employees participated. The work-related behaviors studied were absenteeism, number of hours worked, and intention to quit. The construct validity was studied by examining the relationship between willingness to work and the work-related behaviors.

121

Table 22 Intercorrelations among the variables. N= 179 (Study 1)
Variables
1

2

3

1. Willingness to work 2. Absenteeism 3. Number of hours worked 4. Intention to quit - .12
.31 * - .56**

- .14
.11 - .03

*p < .05 , **p

References: 2 Moreover, the factors that in the Sjoberg and Lind (1994) study were found to explain work motivation were primarily work interest, perceived risks, creativity, and organizational commitment 7 (Sjoberg & Lind, 1994) 2.3.2.2 The concept of will Volition or will is an old psychological construct with strong ties to modem philosophy and a range of colloquial meanings (Pervin, 1992) 12 Abstracts no longer carried a specific entry for 'will ' or 'volition ' (Nilsson, 1998) 13 behavioral control is said to come not from active awareness but from subtle cues in the environment and from thought processes and information not readily accessible to consciousness (Park, 1999)