Citation
Challenge and Hindrance Stressors in the Workplace: Tests of Linear, Curvilinear, and Moderated Relationships with Employee Strains, Satisfaction, and Performance

Material Information

Title:
Challenge and Hindrance Stressors in the Workplace: Tests of Linear, Curvilinear, and Moderated Relationships with Employee Strains, Satisfaction, and Performance
Creator:
PODSAKOFF, NATHAN P. ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Employees ( jstor )
Job performance ( jstor )
Job satisfaction ( jstor )
Labor ( jstor )
Locus of control ( jstor )
Modeling ( jstor )
Parametric models ( jstor )
Psychological stress ( jstor )
Stress strain relationships ( jstor )
Stress tests ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Nathan P. Podsakoff. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
5/31/2010
Resource Identifier:
659815082 ( OCLC )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

1 CHALLENGE AND HINDRANCE STRESSORS IN THE WORKPLACE: TESTS OF LINEAR, CURVILINEAR, AND MODERATED RELATIONSHIPS WITH EMPLOYEE STRAINS, SATISFACTION, AND PERFORMANCE By NATHAN P. PODSAKOFF A DISSERATION PRESENTED TO THE GRADUATE SCHO OL IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007

PAGE 2

2 © 2007 Nathan P. Podsakoff

PAGE 3

3 To the faculty, friends, and family who made this project possible through the ir mentoring, love , and ability to keep from being to o

PAGE 4

4 ACKNOWLEDGEMENTS I thank my committee for their guidance throughout the dissertation process . I thank the students and employees who participated in my research for their time and effort. I thank my parents for thei r limitless support. I thank the Heritage Club, Associate Dean Andy McCollough, Dr. Henry Tosi, and Dr. Jeffery LePine for providing financial support for this project.

PAGE 5

5 TABLE OF CONTENTS page LIST OF TABLES ................................ ................................ ................................ ........................... 8 LIST OF FIGURES ................................ ................................ ................................ ....................... 10 ABSTRACT ................................ ................................ ................................ ................................ ... 12 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .................. 14 2 LITERATURE REVIEW ................................ ................................ ................................ ....... 20 ................................ ............................. 21 Ch allenge (Positive) and Hindrance (Negative) Stressors in the Workplace ......................... 23 Challenge and Hindrance Stressors Measured as Positive and Negative Life Events .... 24 Challenge and Hindrance Stressors Measured in Work Settings ................................ .... 26 Challenge and Hindrance Stressors Operationalized in Meta Analytic Studies ............. 28 Limitations of Previous Research on Challenge and Hindrance Stressors ...................... 29 Construct contamination and deficiency ................................ ................................ .. 29 Nature of the relationships between stressors and their measures ........................... 32 Potential effects of method variance ................................ ................................ ........ 37 Potential Curvilinear Stressor Employee Criterion Relationships ................................ ......... 38 Research Examining Curvilinear Stressor Relationships ................................ ................ 40 Limitations of Previous Research Examining Curvilinear Stressor Relationships ......... 47 Hypotheses ................................ ................................ ................................ .............................. 51 of Stress and Coping and Curvilinear Stressor Criteria Relationships ................................ ................................ ................................ ... 51 Relationships among Challenge Stressors and Employee Criteria ................................ .. 53 Relationships among Hindrance Stressors and Employee Criteria ................................ . 58 Moderators of the Stressor Criteria Relationships ................................ .......................... 61 The moderating effects of negative affectivity on stressor criteria relationships ........... 61 The moderating effects of employee ability on stressor criteria relationships ........ 63 The moderating effects of employee locus of control on stressor criteria relationships ................................ ................................ ................................ .......... 65 Interactive challenge hindrance stressor effects ................................ ....................... 67 3 STUDY 1 ................................ ................................ ................................ ................................ 83 Step 1: Conceptual Definition of the Constructs ................................ ................................ .... 83 The Appraisa l of Challenging Work Stressors ................................ ................................ 88 The Appraisal of Hindering Work Stressors ................................ ................................ ... 93 Nature of the Relationships between Challenge an d Hindrance Stressors and Their Measures ................................ ................................ ................................ ...................... 98 Step 2: Generate Items to Measure the Constructs ................................ ............................... 100

PAGE 6

6 Step 3: Assess the Content Vali dity of the Measures ................................ ........................... 102 Step 4: Formally Specify the Measurement Model ................................ .............................. 107 Step 5: Collect Data and Refine the Measures ................................ ................................ ..... 108 Sample ................................ ................................ ................................ ........................... 109 Measures ................................ ................................ ................................ ........................ 109 Challenge and hindrance stressors ................................ ................................ ......... 109 Challenge and hindrance related work stress ................................ ....................... 110 Analytical Procedures ................................ ................................ ................................ .... 110 Results ................................ ................................ ................................ ........................... 111 Confirmatory factor analysis of the challenge and hindrance stressor dimensions ................................ ................................ ................................ .......... 111 Second order CFA of the ch allenge and hindrance stressor measures .................. 114 Reliability, convergent validity, and discriminant validity at the item level ......... 117 As sessment of convergent and discriminant validity at the dimension level ......... 119 Step 6: Specify the Nomological Network and Test the Hypothesized Relationships ......... 122 Sample ................................ ................................ ................................ ........................... 122 Measures ................................ ................................ ................................ ........................ 123 Challenge and hindrance stressors ................................ ................................ ......... 123 Negative affectivity ................................ ................................ ................................ 123 Job related strains ................................ ................................ ................................ ... 123 Stressor related emotions ................................ ................................ ....................... 124 Motivation ................................ ................................ ................................ .............. 124 Job satisfaction ................................ ................................ ................................ ....... 124 Task performance ................................ ................................ ................................ ... 125 Organizational citizenship behavior ................................ ................................ ....... 125 Counterproductive work behavior ................................ ................................ .......... 125 Analytical Pro cedure ................................ ................................ ................................ ..... 126 Results ................................ ................................ ................................ ........................... 126 Criterion related validity of the dimensions ................................ ........................... 126 Criterion related validity of the second order stressor constructs ......................... 129 Comparison of second order stressor constructs ................................ .................... 131 Criterion related validity controlling for negative affectivity ................................ 131 Supplemental analysis ................................ ................................ ............................ 133 4 DISCUSSION OF STUDY 1 ................................ ................................ ............................... 161 Assessment of Challenge and Hindrance Stressor Measure Construct Validity .................. 161 Challenge and Hindrance Stressors as Higher Order Fact ors ................................ .............. 164 The Nature of the Relationships between the Stressor Constructs and Their Dimensions .. 165 Limitations ................................ ................................ ................................ ............................ 168 Conclusions ................................ ................................ ................................ ........................... 170 5 STUDY 2 ................................ ................................ ................................ .............................. 171 Method ................................ ................................ ................................ ................................ .. 171 Participants ................................ ................................ ................................ .................... 171 Procedure ................................ ................................ ................................ ....................... 171

PAGE 7

7 Measures ................................ ................................ ................................ ........................ 172 Challenge and hindrance stressors ................................ ................................ ......... 173 Job related strains ................................ ................................ ................................ ... 173 Job satisfaction ................................ ................................ ................................ ....... 173 Task performance ................................ ................................ ................................ ... 174 Organizational citizenship behavior ................................ ................................ ....... 174 Counterproductive work behavior ................................ ................................ .......... 174 Negative affectivity ................................ ................................ ................................ 174 Employee ability ................................ ................................ ................................ .... 174 Locus of control ................................ ................................ ................................ ..... 175 Analysis ................................ ................................ ................................ ................................ 175 Results ................................ ................................ ................................ ................................ ... 176 Creating Scale Scores for the Challenge an d Hindrance Stressor Constructs ............... 176 Construct Validity ................................ ................................ ................................ ......... 178 Partitioning of Variance Components ................................ ................................ ........... 180 Tests of the Linear Effects Hypotheses ................................ ................................ ......... 180 Linear effects when challenge and hindrance stressors are modeled as having formative indicators ................................ ................................ ............................ 181 Linear effects when challenge and hindrance stressors are modeled as having reflective indicators ................................ ................................ ............................ 182 Tests of the Curvilinear Effects Hypoth eses ................................ ................................ . 183 Curvilinear effects when challenge and hindrance stressors are modeled as having formative indicators ................................ ................................ ................ 184 Curvilinear e ffects when challenge and hindrance stressors are modeled as having reflective indicators ................................ ................................ ................. 185 Tests of the Cross Level Moderating Effects Hypotheses ................................ ............ 186 Moderator effects when challenge and hindrance stressors are modeled as having formative indicators ................................ ................................ ................ 187 Moderator effects when challenge and hindrance stressors are mo deled as having reflective indicators ................................ ................................ ................. 191 Tests of the Interactive Stressor Effects Hypotheses ................................ .................... 194 6 DISCUSSION OF STUDY 2 ................................ ................................ ............................... 219 Assessment of Hypothesized Linear Effects ................................ ................................ ........ 220 Assessment of the Hypothesized Curvilinear Stressor Effects ................................ ............. 222 Assessment of the Hypothesized Moderator Effects ................................ ............................ 225 Assessment of the Results Obtained from Formative versus Reflective Indictor Modeling Techniqu es ................................ ................................ ................................ ........ 226 Limitations ................................ ................................ ................................ ............................ 227 Conclusions ................................ ................................ ................................ ........................... 232 7 IMPLICATIONS AND FUTURE RESEARCH ................................ ................................ .. 234 Theoretical Contributions ................................ ................................ ................................ ..... 235 Future Research ................................ ................................ ................................ .................... 238

PAGE 8

8 LIST OF TABLES Table page 2 1 Summary of selected studies examining positive versus negative forms of stress. ........... 69 2 2 Summary of studies testing curv ilinear stress criteria relationships . ................................ . 79 3 1 Summary of challenge hindrance stressor dimensions identified in previous research. . 138 3 2 Construct conceptual domains for challenge and hindrance stressors. ............................ 140 3 3 Demographic characteristics of the first undergraduate content validation sample. ....... 141 3 4 Means and standard deviations from the first content validation of the challenge stressor items. ................................ ................................ ................................ ................... 142 3 5 Means and standard deviations from the first content validation of the hindrance stressor items. ................................ ................................ ................................ ................... 143 3 6 Demographic characteristics of the second undergraduate content validation sample. ... 145 3 7 Means and standard deviations from the second content validation of the challenge stressor items. ................................ ................................ ................................ ................... 146 3 8 Summary of means and standard deviations from the second content validation of the hindrance stressor items. ................................ ................................ ................................ .. 148 3 9 Demographic characteristics of working MBA sample (Study 1) ................................ ... 150 3 10 Summary of model fit for the challenge and hindrance stressor construct CFAs. .......... 151 3 11 Results of first order CFA of the challenge stressor construct. ................................ ....... 152 3 12 Results of first order CFA of the hindrance stressor construct. ................................ ....... 153 3 13 Results of CFAs modeling challenge and hindrance stressors as second order constructs. ................................ ................................ ................................ ........................ 155 3 14 Means, standard deviations, reliabilities, and intercorrelations of the dimensions of the challenge and hindrance stressor constructs. ................................ ............................. 156 3 15 Correlations between the dimensions of the challenge and hindrance stressor stress measures. ................................ ................................ ................................ ................ 157 3 16 Correlations among the dimensions of the challenge and hindrance stressor constructs and employee strains, emotions, job satisfaction, and motivation. ................ 158

PAGE 9

9 3 17 Results of str uctural models with challenge and hindrance stressors modeled as second order constructs (controlling for negative affectivity). ................................ ........ 159 3 18 Results of structural models examining relationships b etween the dimensions of the challenge and hindrance stressor constructs and employee strains, emotions, job satisfaction, motivation, and performance criteria. ................................ .......................... 160 5 1 Demographic characteris tics of insurance employee sample (Study 2) .......................... 212 5 2 Variable means, standard deviations, reliabilities and intercorrelations .......................... 213 5 3 Parameter estimates and variance components of null models for strains, job satisfaction, and performance criteria ................................ ................................ .............. 214 5 4 Results of HLM analyses testing linear stressor relationships ................................ ......... 214 5 5 Results of HLM analyses testing interindividual and cross level interactive relationships with emotional exhaustion. ................................ ................................ ......... 215 5 6 Results of HLM analyses testing interindividual and cross level interactive relationships with frustration. ................................ ................................ .......................... 215 5 7 Results of HLM analyses testing interindividual and cross level interacti ve relationships with job satisfaction. ................................ ................................ ................... 216 5 8 Results of HLM analyses testing interindividual and cross level interactive relationships with task performance. ................................ ................................ ............... 216 5 9 Results of HLM analyses testing interindividual and cross level interactive relationships with organizational citizenship behavior. ................................ ................... 217 5 10 Results of HLM analyses testing interindividual and cross level interactive relationships with counterproductive work behavior. ................................ ...................... 217 5 11 Results of HLM analyses testing interactive challenge stressor hi ndrance stressor relationships. ................................ ................................ ................................ .................... 218

PAGE 10

10 LIST OF FIGURES Figure page 3 1 Overview of the scale development and validation process. ................................ ........... 135 3 2 CFA of second order challenge stressor construct having formative indicators represented by its dimensions. ................................ ................................ ......................... 136 3 3 CFA of second order hindran ce stressor construct having formative indicators represented by its dimensions. ................................ ................................ ......................... 137 5 1 Plot of the curvilinear relationship between challenge stressors (modeled as having formative indicator s) and organizational citizenship behaviors. ................................ ..... 196 5 2 Plot of the curvilinear relationship between challenge stressors (modeled as having reflective indicators) and emotional exhaustion. ................................ ............................. 197 5 3 Plot of the curvilinear relationship between challenge stressors (modeled as having reflective indicators) and job satisfaction. ................................ ................................ ....... 198 5 4 Plot of the curvilinear relationship between hindrance stressors (modeled as having reflective indicators) and task performance. ................................ ................................ .... 199 5 5 Plot of the moderating effect of locus of co ntrol on the linear relationship between hindrance stressors (modeled as having formative indicators) and frustration. ............... 200 5 6 Plot of the moderating effect of ability on the curvilinear relat ionship between challenge stressors (modeled as having formative indicators) and job satisfaction. ....... 201 5 7 Plot of the moderating effect of ability on the linear relationship between hindrance stressors (modeled as having formative indicators) and organizational citizenship behaviors. ................................ ................................ ................................ ......................... 202 5 8 Plot of the moderating effect of negative affectivity on the linear relationship between c hallenge stressors (modeled as having reflective indicators) and emotional exhaustion. ................................ ................................ ................................ ....................... 203 5 9 Plot of the moderating effect of locus of control on the linear relationship between challenge st ressors (modeled as having reflective indicators) and job satisfaction. ........ 204 5 10 Plot of the moderating effect of locus of control on the linear relationship between challenge stressors (model ed as having reflective indicators) and organizational citizenship behavior. ................................ ................................ ................................ ........ 205 5 11 Plot of the moderating effect of negative affectivity on the linear relationship between challenge stresso rs (modeled as having reflective indicators) and counterproductive work behavior. ................................ ................................ ................... 206

PAGE 11

11 5 12 Plot of the moderating effect of ability on the curvilinear relationship between challenge stressors (modeled as having reflective indicators) and task performance. .... 207 5 13 Plot of the moderating effect of ability on the curvilinear relationship between hindrance stressors (modeled as having reflective indicators) and task performance. .... 208 5 14 Plot of the moderating effect of locus of control on the curvilinear relationship between hindrance stressors (modeled as having reflecti ve indicators) and task performance. ................................ ................................ ................................ .................... 209 5 15 Plot of the moderating effect of hindrance stressors on the linear relationship between challenge stressors (modeled as having reflective indicato rs) and emotional exhaustion. ................................ ................................ ................................ ....................... 210 5 16 Plot of the moderating effect of hindrance stressors on the linear relationship between challenge stressors (modeled as having reflective indicators) and frustration. . 211

PAGE 12

12 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy CHALLENGE AND HINDRANC E STRESSORS IN THE WORKPLACE: TESTS OF LINEAR, CURVILINEAR, AND MODERATED RELATIONSHIPS WITH EMPLOYEE STRAINS, SATISFACTION, AND PERFORMANCE By Nathan P. Podsakoff May 2007 Chair: Jeffery LePine Major: Business Administration My research involved two studies. Study 1 was designed to develop and validate multi dimensional scales of challenge and hindrance stressors having formative indi cators. To this end, I used a six step process and three different samples to examine the construct validity of newly developed challenge and hindrance stressor measures. Generally speaking, the results of Study 1 provided support for the construct validity of the stressor measures. More specifically, the measures developed to assess challenge and hindrance stressors were found to be reliable, distinct from one another, and to exhibit similar relationships with employee strains, but opposite relationships with employee emotions, job satisfaction, motivation, and to a lesser extent performance. In addition, the results from Study 1 provided support for the modeling of the challenge and hindrance stressors as second order constructs, rather than as distinct dimensions. Study 2 was designed to assess the potential linear and curvilinear relationships between challenge and hin drance stressors and employee strains, satisfaction, and performance. Study 2 u s ed the newly developed work stressor scales , a sample of 164 employees of an insurance branch , and multi level modeling to test these hypotheses. Generally, the results of this study provided mixed support for hypotheses regarding the linear relationships between challenge and

PAGE 13

13 hindrance stressors and employee strains and job satisfaction , and limited support for h ypotheses regarding the curvilinear stressor criterion relationshi ps. In addition, the study also p rovided limited support for hypotheses regardin g the moderating effects of employee ability and locus of control on these linear and curvilinear relationships. In conclusion, I discuss the theoretical and practical implicat ions of these studies and recommendations for future research.

PAGE 14

14 CHAPTER 1 INTRODUCTION If the number of articles written in the popular press and in academic journals is a ny indication, job related stress is an important topic for both practicing manage rs and academics. For example, a Google search using the key word s 37. 7 million websites , and a search of the Web of Knowledge database using these same keywords produced a listing of 4,693 academic articles. Comparing the se findings w ith searches conducted using keywords related to job satisfaction (Google, 15.8 mil lion hits; Web of Knowledge, 8, 062 articles), the most studied variable in the field of management (Judge, Thoresen, Bono, & Patton, 2001), indicates that inter est in work stress is substantial. Although reasons for the interest in the topic of work stress vary, it is obvious that one important reason is that employees in the workplace are experiencing in creas ed amounts of stress. For example, surveys conducted over the last four decades indicate that half to three 1996; N ational Institute of Occupational Safety and Health, 1999) and that employees generally perceive their stress levels to be much greater than what they experienced 5 to 10 years ago (The Dream, 1994; The Mitchum Report, 1990). Indeed, constantly increasing levels of work related stress have made this generation of employees more stressed than their predeces sors, to the extent that some have labeled this trend an epidemic (Jones & Bartlett, 1995; Marino, 1997). Unfortunately, in a work environment characterized by increasing workloads, broadened job scopes, limited economic resources, and corporate downsizing there is an expectation that this trend will continue in the future. Given these increasing levels of stress, it is probably not surprising that a primary concern for managers and researchers has been to identify the effects of this trend on employees.

PAGE 15

15 G enerally speaking, there are two primary perspectives explaining stressor effects on employee outcomes such as performance. The more popular of the two perspectives is that work stressors have unequivocal detrimental effects on individuals and organization s (Jamal, 1984, 1985; Westman & Eden, 1996). Proponents of this perspective note that employees experiencing high levels of stress cost their employers 46% more in health care expenses than do employees experiencing relatively low levels of stress (Goetzel , Anderson, Whitmer, Ozminkowski, Dunn, & Wasserman, 1998), and that work related stress costs businesses worldwide an estimated $150 billion annually (Donatell e & Hawkins, 1989). Empirical research in the stress domain generally supports this perspective in that work related stressors have been shown to be related to employee burnout, physical illnesses, dissatisfaction, poor performance, and turnover (Jackson & Schuler, 1985; Kahn & Byosiere, 1992; Sonnentag & Frese, 2003). In contrast, advocates of the second perspecti ve (C hampoux, 1980, 1981; Muse et al., 2003) believe that the relationships between work stressors and employee outcomes are more complex. Generally speaking, these individuals argue that work stressors can have either positive or negative effects on employees, depending on the level at which stress is experienced by the individual. Based on research conducted by Yerkes and Dodson (1908), the general proposition of this approach states that work stress is functional for individuals up to an optimal level because it increases arousal and engagement in work. However, after stressors exceed this optimal level, the individual becomes overloaded, anxious, and experiences dysfunctional effects, such as decreased performance. This perspective has be en interpreted as suggesting that the relationship between stress and performance (and other employee outcomes such as job satisfaction) resembles an inverted U (Yerkes & Dodson, 1908).

PAGE 16

16 Despite the intuitive appeal of this second perspective, empirical tes ts of the underlying hypotheses have not proved to be very supportive. Indeed, a review of the empirical studies conducted to test this theory by Muse et al. (2003) indicated that only two studies provided support for the inverted U shaped relationship bet ween stress and performance. This has led some scholars to criticize the hypothesis ( Westman & Eden, 1996) or to recommend that it be ab andoned altogether (Neiss, 1988 ). However, other researchers (Muse et al., 2003) are somewhat more optimistic, arguing t hat the lack of support for the inverted U theory may be a More recently, an alternative perspective describing the relationships between work stressors and e mployee outcomes has emerged. Based on a theory developed by Lazarus and Folkman (1984), this perspective argues that employees experience positive or negative forms of stress based on their appraisal of work demands as either challenges or hindrances, and that Boswell, Roehling, and Boudreau (2000) found that job stressors could be categorized into those that tend to be perceived as promoting personal growth and Olson Buchanan , & LePine, 2004; Ca vanaugh et al., 2000) and meta analytic reviews (LePine, Podsakoff , & LePine, 2005; Podsakoff, LePine , & LePine, 2007 ) have shown that this framework can be used to explain inconsistent relationships between stressors and employee attitudes, perceptions, a nd behaviors, such that challenge stressors generally have functional relationships with employee motivation, satisfaction, organizational commitment, task

PAGE 17

17 performance, and retention, whereas hindrance stressors tend to have dysfunctional rel ationships wit h these criteria. In addition, the challenge hindrance stressor framework may be used to explain why both of the perspectives on stressor effects may, in part, contain elements of truth. Put simply, challenge and hindrance stressors may differentially repr esent the two prevailing perspectives on stressor criteria relationships. Because hindrance stressors are comprised of work demands that are thought to be dysfunctional at virtually all levels, these stressors are expected to produce negative affective and behavioral reactions from employees that are commensurate with the perceived levels of stress that are experienced by them. In contrast, challenge stressors provide employees with the opportunity for growth and gain, and there is growing evidence suggesti ng that these stressors can have positive relationships with employee motivation, satisfaction, and performance (LePine et al., 2005; Podsakoff et al., 2007 ). However, despite their generally positive effects, it would be logical to expect that very high l evels of challenging job demands can also become overwhelming for employees. When employees perceive that they are overloaded, their deadlines are overly ambitious, their job scope is too broad, or they carry too much responsibility in the workplace, they are likely to experience more strain and exhibit decreased levels of satisfaction and performance. This suggests that challenge stressors may be functional for employees up to a certain level and then become dysfunctional after exceeding that level. In oth er words, these types of demands may be representative of the curvilinear inverted U shaped perspective on stressor effects. Despite the potential for the challenge and hindrance stressor constructs to explain the differential effects of stressful demands in the workplace, the current measurement of these constructs is limited is several ways. Specifically, measures of challenge and hindrance stressors

PAGE 18

18 suffer from deficiency, contamination, method effects, and measurement model misspecification. These limi tations suggest that the interpretation of the substantive relationships between these stressors and other constructs may be biased. To conduct proper tests of the linear and non linear relationships between challenge and hindrance stressors and important employee criteria, measures of these stressors that address these limitations need to be developed and validated. In addition, researchers have also criticized research designs traditionally used to examine curvilinear stressor criterion relationships. Th e work stress literature is generally characterized by cross sectional studies which provide limited s work related stressors at one point in time. Some researchers (Kahn & Byosiere, 1992; Muse, Harris, Field, 2003) have argued that these designs cannot adequately capture fluctuations in the stress levels experienced by individuals, and thus , are deficient means of t est ing curvilinear stressor effects. Consistent with the notion that stress should be assessed at multiple t imes and in different contexts, Lazarus and Folkman (1984) have called f or the use of repeated measure s designs to better understand how individual differences may influence react ions to stressors over time . My research incorporates th is recommendation by using experience sampling m ethod (ESM) techniques , which involve taking repeated subsequent attitudinal and behavioral responses (Csikszentmihalyi & Larson, 1987) . T he ESM design facilitates the analysis of within pe rson variation in perceived stressors, and the examination of how different individual characteristics influence reactions to these stressors. Therefore, the purpose of this dissertation is to: (a) develop and validate new scales designed to measure chall en ge and hindrance stressors, (b) test differential hypotheses regarding the potential linear and curvilinear relationships between challenge and hindrance stressors and employee strains, job satisfaction, and performance and (c) test the potential moderat ing effects

PAGE 19

19 of three individual differences (negative affectivity, ability, and locus of control) on the hypothesized linear and curvilinear relationships. T o meet these objectives, I proceed ed in the following manner. First, I provide d a brief overview of general stress theory, with special emphasis on the cognitive appraisal model developed by Lazarus a nd Folkman (1984). Then, I reviewed the emerging literature on challenge related (positive) and hindrance related (negative) forms of stress in the work en vironment and some of the limitation s of this research. Next, I review ed studies that were specifically designed to examine the curvilinear relationships between work stressors and employee outcomes. I have argued that t he challenge hindrance stressor fram ework may be used to explain i nconsistencies that have been reported in previous research examining these relationships. Following this, I drew on theory developed by Lazarus and Folkman (1984) to develop specific hypotheses regarding the linear and curvi linear relationships between challenge and hindrance stressors and employee outcomes, and the potent ial moderating effects of employee negative affect, ability, and locus of control on these relationships. I conducted two studies designed to test these hyp otheses. In Study 1, I develop ed and validate d multidimensional measures of challenge and hindrance stressors using a s ix step process and three samples. In Study 2, I a sample of 164 insurance office employees, an ESM design, and m ulti level modeling tech niques to test the linear, curvilinear, and moderated relationship hypotheses. Finally, I discuss the implications of the findings of my studies and make recommendations for future research.

PAGE 20

20 CHAPTER 2 LITERATURE REVIEW Much of the early research in the work stress domain was driven by a stimulus response model of the stress process. Generally speaking, this model argues that stressors in the work environment elicit a variety of physical, psychological, and behavioral strains on the part of employees (Je x, 1998). Therefore, strains such as anxiety, exhaustion, frustration, and somatic symptoms are hypothesized to be the proximal outcomes of work related stressors such as role conflict, role ambiguity, and workload. Consistent with approach, research has s upported behaviors (Kahn, Wolfe, Quinn, Snoek, & Rosenthal, 1964). For example, meta analytic reviews of the role stressor literature (cf. Abramis, 1994; Fishe r & Gittleson, 1983; Jackson & Schuler, 1985) have reported that role ambiguity and role conflict are generally negatively related to employee satisfaction, organizational commitment, job involvement, and, to a somewhat lesser extent, performance. H owever, the stimulus response approach to work stressors has some limitations. First, this approach assumes that all stressors have basically the same effects on all people. According to this approach, all stressors are viewed as stimuli that have direct effects response which are not mediated in any way by cognitive processes occurring inside the individual. Thus, this approach assumes that all people react to stressors in basically the same manner. Related to this, the stimulus response view o f stressors does not acknowledge the potential effect of individual differences in the stress response, even though there is fairly compelling evidence that stressors do not affect all people in the same way (Lazarus & Folkman, 1984; Brief, Burke, George, Robinson, & Webster, 1988).

PAGE 21

21 In response to the limitations of the stimulus response model, Lazarus and Folkman (1984) in appraisal of the perceived demands and the subsequent coping proc esses exhibited by the individual in response to these appraisals serve as mediators of stressor strain relationships. According to this model, environmental stressors acquire meaning and significance through a cognitive appraisal or evaluative process. Th most readily understood as the process of categorizing an encounter, and its various facets, with respect to its significance for well process, ind ividuals assess whether a stressor represents a potential challenge, harm, or threat . Appraisal of a stressor as a challenge are typically appraised as harmful if psychological or physical damage has already been sustained by the individual or another person who is loved or valued. Finally, a threat appraisal is assigned to conditions or events in which harms or losses are antic ipated, but have not yet taken place. Lazarus and Folkman (1984) indicate that harm/loss and threat appraisals may be conjoined, because every harm appraisal includes the potential to elicit threat and because people not only assess current loss, but also their ability to function in the future. They also note that these appraisals are important to the stress process because they lead to very different affective reactions from those produced by challenge related stressors. For example, demands that are app raised as challenging are characterized by feelings of eagerness, excitement, and joy, whereas demands appraised as harmful or threatening are associated with fear and anger.

PAGE 22

22 The cognitive appraisal process is also important because it leads individuals to engage in efforts to manage or cope with the demands they experience. Specifically, Lazarus and Folkman specific external and/or internal demands that are appraised a s taxing or exceeding the resources receive challenge or threat appraisals elicit their own distinct patterns of coping. Generally speaking, challenge apprai sals are associated with vigilant or active forms of coping, because they encourage approach, openness, and problem focused coping (Lazarus & Folkman, 1984). Problem focused forms of coping include strategies similar to those used during problem solving su ategory of coping, which includes behaviors such as making a plan of actio n or exhibiting increased effort . Therefore, demands appraised as providing the potential for growth or gain tend to encourage positive coping behaviors such as preparation, allocati ng time for projects or tasks, and increased effort. In contrast, the appraisal of demands as threats typically elicits more complex forms of the person inward 191). Lazarus and Folkman (1984) refer to this form of coping as emotion focused, indicating udes generally consistent with escape or avoidance strategies of coping identified by Latack (1986), which include staying clear of the situation, trying n ot to get involved , and retaliation. Not

PAGE 23

23 surprisingly, these types of responses are not expected to promote effective social functioning and typically result in negative consequences (Koeske, Kirk, & Koeske, 1993). Therefore, demands appraised as threateni ng employee growth or gain are likely to result in negative coping behaviors such as avoidance of these demands, telling oneself that the demand is unimportant, negative forms of appraisal and coping processes. Challenge (Positive) and Hindrance (Negative) Stressors in the Workplace At about the same time that Lazarus and Folkman (1984) were developing their cognitive evaluation model of stress, researchers in the organizational behavior domain (e.g., Bhagat, McQuaid, Lindholm, & Segovis, 1985; Sarason & Johnson, 1979; Scheck, Kinicki , & Davy, 1995, 1997) began to recognize the possibility that not all stressors in the workplace have dysfunctional effects on employee attitudes, perceptions, and behaviors, and that some stressors may even have functional effects on these employee outcome variables. Although much of the early Boswell et al., 2004; Cavanaugh et al., 2000) has specifically utilized the c hallenge/hindrance stressor framework to examine the effects of stressors in the workplace. A summary of the research on challenge related (positive) and hindrance related (negative) stressors is provided in Table 2 1. Generally speaking, this table report s the authors, the sample from which data was obtained, the operationalizations of the stressors, the reported findings, and potential limitations of each study. An examination of this table reveals that there have been three general ways of operationalizi ng challenge and hindrance stressors in the previous research. The first of these approaches measures the positive and negative evaluation of various

PAGE 24

24 life events. Although this approach did not technically label stressors as challenges or hindrances, it is included here because it is a dire ct predecessor to the challenge /hindrance stressor framework. The second approach is specifically designed to measure challenge and hindrance stressors in workplace settings and their relationships with employee criteria. Finally, the third approach uses meta analytic data to determine whether challenge and hindrance stressors have differential relationships with employee criteria. I will discuss research using each of these approaches in the sections that follow. Challenge and Hindrance Stressors Measured as Positive and Negative Life Events et al., 1985; Kanner, Coyne, Schaefer, & Lazarus, 1981; Sarason & Johnson, 1979; Sch eck et al., 1995, 1997) employed individual evaluations of various life events to obtain measures of positive and negative life stress. For example, Sarason and Johnson (1979) had 44 naval personnel indicate the positive or negative impact that 57 life eve nts (marriage, personal illness, pregnancy, etc.), and 30 work related events (promotion, change in work schedule, conflict with coworkers, etc.) had on their life during the previous year. Respondents rated each event on a a very great impact) and then scale scores were created from these assessments. These authors reported that negative organizational stressors were negatively related to employee satisfactio n with work and coworkers, whereas positive organizational stressors were positively related to both forms of employee satisfaction. Three studies (Bhagat et al., 1985; Scheck et al., 1995, 1997) used the Psychiatric Epidemiology Research Interview (PERI) Life Events Scale (Dohrenwend, Ashkenasy, Krasnoff, & Dohrenwend, 1978) in combination with the response format used by Sarason and Johnson (1979) to assess positive and negative stress experienced by employees. Generally speaking, the

PAGE 25

25 PERI scale is a mul ti family, social activities, finances, relationships, compensation, and job. In all three of these studies, individuals rated the negative and positive impact these events had on their lives over a specific period of time. For example, in their study of 282 administrative, health care, and clerical personnel, Bhagat et al. (1985) reported that positively evaluated job related events that occurred during the previous three years we re positively related to employee work satisfaction and organizational commitment, and negatively related to job strains and job alienation; whereas negatively evaluated job related events that occurred during the same time period had opposite, dysfunction al relationships with these criteria. Scheck et al. (1995, 1997) used similar methods to examine negative and positive life events in two different studies. In their longitudinal study of 135 employees working in a technology firm, Scheck et al. (1995) fo und that positively evaluated life stressors (which included job related events) that occurr ed during the previous year had positive relationships with employee job satisfaction and organizational commitment, whereas negatively evaluated life stressors (wh ich also included job related events) that occurr ed during that same period had negative relationships with these employee job attitudes. A follow up cross sectional study conducted by Scheck et al. (1997) with 218 employees of another technology based org anization found that positively work relationships, and compensation) that happened during the previous year were positively related to problem focused forms of coping and subjecti ve well being; whereas negatively evaluated occupational stressors were positively related to emotion focused forms of coping and strain, and negatively related to employee s well being.

PAGE 26

26 Finally, taking a slightly different tack, Kanner et al. (1981) exam ined the effects of minor daily events, described as either hassles or uplifts, on affective feelings. Kanner et al. argued that ., 1981, p. 3), and include losing things, getting into arguments, concerns about money, and being from manifestations of love, relief at hearing good news, th work, and relaxing. In their study of 100 participants, Kanner et al. obtained four separate ratings of the frequency and i ntensity of 117 daily hassles and 135 daily uplifts over the course of a 10 month period. Generally speaking, these authors reported that the intensity and frequency of daily uplifts were positively related to the positive affect of the respondents, and th at the Challenge and Hindrance Stressors Measured in Work Settings At the beginning of the 21 st century, a new paradigm for assessing positive and negative work stressors em erged. This approach is characterized by the identification of demands that tend to be appraised by employees as either challenges or hindrances in the workplace. Based on the employee development literature, Cavanaugh et al. (2000) developed a two dimensi onal framework that classified specific work demands as challenge related stressors (i.e., demands that generally promote feelings of fulfillment and challenge in the workplace) or hindrance related stressors (i.e., demands that interfere with an individua goals related to personal development). Using a sample of over 1800 upper level managers,

PAGE 27

27 Cavanaugh and her colleagues reported that stress from challenge related job demands (e.g., workload, time pressure, and responsibili ty) was positively related to job satisfaction and negatively related to job search, and that stress from hindrance related job demands (e.g., role ambiguity, organizational politics, red tape, and job insecurity) was negatively related to job satisfaction , and positively related to job search and voluntary turnover. Building on the work of Cavanaugh et al. (2000), Boswell and her colleagues have conducted two additional studies (Boswell et al., 2004; Bingham et al., 2005) using this framework. In the fir st study, Boswell et al. attempted to expand the criterion related validity of variables in a sample of 481 university employees. These authors reported that: (a) challenge and hindrance stressors were factor analytically distinct from each other, (b) challenge stressors had a positive relationship with employee loyalty and negative relationships with job withdrawal, job search, and intentions to quit, and (c) hindr ance stressors had a negative relationship with employee loyalty, and positive relationships with job search and intentions to quit. In a second Developmental Challenge Pr ofile to examine differences in challenge and hindrance stressor effects across different cultures. Data obtained from 1,377 U.S. managers and 1,871 European managers indicated that although the effects of the hindrance related stressors were fairly simila r across these samples, the effects of challenge related stressors were somewhat different. More specifically, they found that hindrance related job demands were positively related to job search and negatively related to job satisfaction in both the U.S. a nd European manager samples, and challenge related job stressors were positively related to job satisfaction in the European sample.

PAGE 28

28 Challenge and Hindrance Stressors Operationalized in Meta Analytic Studies The third paradigm that has been used to examin e the effects of challenge and hindrance stressors is based on the ex amination of meta analytic data. Because the research using scales specifically designed to assess challenge and hindrances stressors is relatively new , t his approach focuses on coding pr imary studies that have tested the relationships between dimensions of these work stressors and other criterion variables . Although not necessarily a research paradigm, this meta analytic approach a llow s researchers to determine whether work stressor emplo yee criterion relationships differ as a function of the challenging versus hindering nature of the stressors. For example, LePine et al. (2005) used the challenge and hindrance stressor framework to explain inconsistencies in previous research examining wo rk stressor employee performance stressor constructs as a guideline, LePine et al. identified and categorized measures of job and role demands, pressure, time u rgency, and workload as challenge stressors, and measures of constraints, hassles, resource inadequacies, interpersonal conflict, supervisor related stress, organizational politics, and a variety of role str essors as hindrance stressors. Using data taken f rom 101 independent samples, these authors reported that although challenge and hindrance stressors were positively related with employee perceptions of strains, challenge stressors had positive relationships with employee motivation and performance , where as hindrance stressors had negative rel ationships with these criteria. Podsakoff et al. ( 2007 ) also used meta analytic techniques in a second study examining relationships between work stressors and employee attitudes, perceptions, and withdrawal behaviors . Based on theory linking role stressors with employee turnover (Schaubroeck, Cotton, & Jennings, 1989), these authors tested a path model hypothesizing differential challenge and hindrance stressors effects on turnover and withdrawal behaviors through emp loyee job

PAGE 29

29 satisfaction, organizational commitment, and turnover intentions. Using a coding scheme similar to that developed by LePine et al. (2005), Podsakoff et al. coded and analyzed stressor criterion relationships in 185 independent studies. These auth ors reported that challenge stressors had positive relationships with job satisfaction and organizational commitment and negative relationships with employee intentions to quit and actual turnover; whereas hindrance stressors had negative relationships wit h employee job satisfaction and organizational commitment, and positive relationships with employee intentions to quit, turnover, and withdrawal behaviors such as absenteeism and tardiness. Limitations of Previous Research on Challenge and Hindrance Stres sors Taken together, the findings of the studies summarized in Table 2 1 are encouraging in that they suggest that individuals differentiate between challenging (positive) and hindering (negative) forms of stress in the workplace, and that these two forms of stress have markedly different effects on the way people feel and behave in this context. However, despite these encouraging results, there are several important limitations to the studies reported in this table. Generally speaking, these limitations fa ll into three categories: (1) construct deficiency and contamination, (2) the assumed nature of the relationships between the stressor constructs and their measures, and (3) the potential effects of method variance on the relationships between the stressor s and other constructs. In the following sections, I will explain each of these limitations and their implications for research in this domain. Construct contamination and deficiency The first potential limitation in many of the studies reported in Table 2 1 relates to the fact that some of the measures of challenging (positive) and hindering (negative) work stressors are contaminated by other constructs. For example, the PERI scale (Dohrenwend et al., 1978) was not only designed to measure major life even ts at work, but also those that relate to school, love

PAGE 30

30 and/or marriage, children or families, home life, crime and/or legal matters, finances, social activities, and health. Indeed, items measuring work related events make up only one fifth (21 of 102 item s) of the original PERI scale. Although alternative measures of this scale reported by Scheck et al . (1995, 1997) put greater emphasis on work related events, these scales still contain measures of positive or negative stress related to love and marriage, family, social activities, finances, relationships, and/or compensation. As a result, all of these measures are contaminated with stressors that occur outside of the workplace. Similar problems exist for the daily life events measure used by Kanner et al . (1981). Generally speaking, the hassles and uplifts scales include daily events related to family, finances, health, job, personal characteristics, social activities, travel, weather, and so on; and job related al., 1981). Of course, given that these scales were designed to identify daily events in all aspects of life, it should not be too surprising that they represent very contaminated measures of challenge o r hindrance related work stressors. Nevertheless, the fact that these scales contain items that go well beyond the work environment makes their use as measures of challenge and hindrance stressors in the workplace highly problematic. Another limitation of many studies identified in Table 2 1 is that their measures of challenge and hindrance stressors are deficient in that they do not include all the important dimensions that are subsumed under the conceptual domain of these stressor constructs. For example , the PERI scale, which was used in three of the studies identified in this table (Bhagat et al., 1985; Scheck et al., 1995, 1997), does not include several relevant dimensions of hindrance related stressors, such as role ambiguity, role conflict, interper sonal conflict with coworkers,

PAGE 31

31 organizational politics, and resource inadequacies; or dimensions of challenge related stressors such as time pressure, job complexity, or job responsibility. It is also possible that the scales developed by Cavanaugh et al. (2000) to measure challenge and hindrance related job demands are somewhat deficient due to the methods used to develop these measures. Although these authors used items taken from popular work stressor scales (e.g., the Job Demands and Worker Health sca le developed by Caplan, Cobb, French, Harrison, & Pinneau, 1975; the Stress Diagnostic scale developed by Ivancevich & Matteson, 1983; and the Job Stress Index developed by Sandman, 1992) to develop their measures, these scales may not include other work s tressors that are considered to be challenging or hindering to dimensions (or concepts within dimensions) of self reported work stress. . . [and] the dimensionality o f the self reported work stress construct is in need of further theorizing and that items reflecting resource inadequacies, interpersonal conflicts, and job co mplexity are not represented. Finally, the challenge and hindrance stressor measures used in the meta analytic studies conducted by LePine et al. ( 2005) and Podsakoff et al. (2007 ) may be deficient because these studies depended on the operationalizations provided in the primary studies that they summarized. Due to constraints of the primary samples included in these meta analyses, neither of these studies included relationships between measures of job complexity, interpersonal conflict, or job insecurity and the focal criteria. Therefore, it is likely that the measures of the challenge and hindrance stressor constructs used in these studies are also deficient in these respects. In summary, previous measures of work related challenge and hindrance stressors may

PAGE 32

32 suffer from contamination and deficiency which may affect the interpretation of the reported relationships these constructs have with other criteria. Nature of the relationships between stressors and their measures The second potential limitation in studies reported in Table 2 1 is the ambiguous or incorrect specification of the nature of the relationship between the challenge and hindrance stressor constructs and their measures. Virtually all of the studies reported in Table 2 1 assumed that the mea sures used to assess challenge and hindrance stressors are reflections or manifestations of the underlying constructs being measured. This is consistent with classical test and therefore should be highly correlated, interchangeable, and have similar relationships with antecedents and consequences (Bollen & Lennox, 1991; Crocker & Algina, 1986; MacKenzie, Podsakoff, & Jarvis, 2005). Although this is the dominant approach to co nstruct measurement in the field of organizational behavior, an alternative perspective on the relationships between constructs and their measures has recently gained some prominence in the literature (Bollen, 1989; Bollen & Lennox, 1991; Edw ards & Bagozzi , 2000 ; Podsakoff, MacKenzie, Podsakoff, & Lee, 2003; Podsakoff, Shen, & Podsakoff, 2006). The advocates of this perspective note that, in some cases, it is more appropriate to view measures of a construct as defining characteristics that combine to form t he construct, rather than as reflections of it. In this case, the measures of the construct are formative indicators are not necessarily: (a) highly correlat ed with each other, (b) interchangeable, and/or (c) expected to have similar relationships with other constructs (Bollen & Lennox, 1991; MacCallum & Browne, 1993; MacKenzie et al., 2005). To help scholars determine whether a construct should be conceptua lized as having reflective or formative measures, researchers (Bollen & Lennox, 1991; Podsakoff et al., 2003 b )

PAGE 33

33 have developed three questions to ask when considering the nature of the relationships between a focal construct and its measures: (1) are the me asures viewed as defining characteristics of the construct or manifestations of it?; (2) are changes in the measures expected to cause changes in the construct or is the construct expected to cause changes in the measures?; (3) would a change in the value of one of the measures be expected to be associated with a change in the value of all of the other measures? Generally speaking, if the measures of a construct are viewed as manifestations of the focal construct, changes in the construct are expected to ca use changes in the measures, and a change in one measure is expected to be related to a change in all of the other measures, then the measures should be conceptualized as reflective indicators of the focal construct. However, if the measures are viewed as defining characteristics of the construct, changes in the measures are expected to cause changes in the construct, and a change in one measure is not necessarily expected to be associated with a change in all the other measures, then the measures are more appropriately specified as formative indicators of the focal construct. Unfortunately, most of the early studies identified in Table 2 1 do not explicitly discuss the nature of the relationships between the positive and negative stress constructs and thei r measures. For example, Sarason and Johnson (1979), Kanner et al. (1981), and Bhagat et al. (1985) do not discuss how the positive or negative evaluations of specific events are related to the constructs of positive and negative life or job stress. Instea d, these authors indicate that scale scores were created by summing the negative and positive evaluations for the events, respectively. However, it would appear from the language used by these authors that they intended these scales to serve as reflections of the underlying stressor constructs. For example, Sarason and Johnson (1979) indicated that , obtain measures reflecting the extent of desirable and undesirable change wit h the working

PAGE 34

34 nature of the relationships between positive and negative stressors and the measures of these constructs in their study, it would appear that the y intended these constructs to be treated as having reflective indicators. More explicit assumptions about the reflective nature of the relationships between the stressor constructs and their measures were made by Scheck et al. (1995, 1997), Cavanaugh et al., (2000), Boswell et al., (2004), and Bingham et al. (2005). Consistent with the assumptions of classical test theory, all of these authors used confirmatory factor analytic techniques to model their measures as reflections of the underlying challenge (positive) or hindrance (negative) stressor constructs. Unfortunately, there are good reasons to believe that the specification of these measures as reflective indicators of the stressor constructs may be problematic. Indeed, several researchers (Cohen, Co hen, Teresi, Marchi, & Velez, 1990; Edwards & Bag oz z i, 2000; Law , Wong, & Mobley, 1998 ; Turner & Wheaton, 1995) have argued that stressors are more appropriately viewed as determinants of an overall stress construct, rather than as reflections of it. For e xample, in their review of latent variable measurement and modeling, Cohen et al. (1990) correlated (p. 186). These the development of an events checklist is conceptually distinct from test construction because items are specifically not alternative estimates of a single under Indeed, other than socio economic status (SES), stress may be the most common example of a construct that should be measured as having formative indicators (Edwards & Baggozi, 2000).

PAGE 35

35 The empirical ev idence seems to support the notion that challenge (positive) and hindrance (negative) stressors should be treated as having formative indicators, in three different ways. First, the results of confirmatory factor analysis on the PERI scale reported by Sche ck et al. (1995) clearly indicated that evaluations of positive and negative stress events occurring in different aspects of life (i.e., love and marriage, family, finance, relationships, social activities, job, and compensation) were not highly correlated and did not have strong loadings on the latent positive or negative stressor constructs (average standardized loadings equal to .34 for positive stressors, and equal to .56 for negative stressors). Second, reliability coefficients reported by Bhagat et al dimensions of the PERI scale may be multi dimensional in nature or more appropriately modeled as having formative indicators. Although low reliabilities do not necessar ily indicate that a construct should be modeled as having formative indicators, they do suggest that the measures may contain other systematic components, and may be multidimensional. Finally, a review of the literature indicates that different forms of c hallenge and hindrance related work stressors are not necessarily highly correlated with each other. For example, meta analytic reviews of the role stressor literature indicate that employee perceptions of role ambiguity and role conflict share , on averag e, less then 20% of their variance with each other ( meta analytic r = .37, Fisher & Gitelson, 1983; meta analytic r = .42, Jackson & Schuler, 1985). In addition, analysis of studies using the Organizational Constraints Scale a nd the Interpersonal Conflict as Work Scale (both of which would be considered forms of hindrance stressors) rep orted that these variables also, on average, shared about 20% of the variance (meta analytic r = .44) . Primary studies examining other work stre ssors have reported that measures of role ambiguity and interpersonal conflict (Spector,

PAGE 36

36 Dwyer, & Jex, 1988), role ambiguity and situational constraints (Jex & Gudanowski, 1992; Spector et al., 1988), and workload and time urgency (Friend, 1982) also share 20% or less of their variance with each other. Therefore, empirical evidence supports the theoretical argument that dimensions of the challenge and hindrance stressor scales need not be highly correlated, are not necessarily interchangeable, and are more appropriately modeled as formative indicators of the stressor constructs. Of course, recognizing the distinction between reflective and formative indicators not only has implications for the manner in which the relationship between a construct and its mea sures should be modeled. It also has implications for the structural relationships that are observed between constructs. For example, a recent Monte Carlo simulation reported by MacKenzie et al. (2005) indicated that the effects of measurement model misspe cification on parameter estimates may be quite severe. More specifically, by comparing the parameter estimates from models with constructs having properly specified (formative) measurement models with parameter estimates from models having constructs with misspecified (reflective) measurement models, these authors reported that unstandardized estimates were biased up to an average of 490%. Although the magnitude and sign of the biases did vary, MacKenzie et al. reported that the parameter estimates in their misspecified model were severally biased regardless of whether the misspecified construct was an exogenous or endogenous variable in the model , or whether the measures were weakly or strongly correlated. Therefore, the findings of this study suggest that constructs that are misspecified as having reflective indicators when they should be modeled as having formative indicators could severely bias the estimation of substantive relationships with other constructs.

PAGE 37

37 In summary, the above discussion sug gests two things. First, the relationships between challenge and hindrance stressors and their measures have been inappropriately specified in past research such that the stressor constructs are modeled as having reflective indicators, when they should have been mo deled as having formative indicators. Second, the continued misspecification of the stressor constructs as having reflective indicators will bias the empirical results of research examining these constructs and lead to inaccurate inferences regarding their relationships with other constructs. Potential effects of method variance The third potential limitation of the studies listed in Table 2 1 is that several of them may have been affected by method variance. According to Bagozzi and Yi (1991), method varia nce: construct of interest. The term method refers to the form of measurement at different levels of abstraction such as content of the specific items, scale type, re sponse format, and general context (Fiske, 1982, pp. 81 84) . . . method effects might also be interpreted in terms of response biases such as halo effects, social desirability, acquiescence, leniency effects, or yea or nay Generally sp eaking, Podsakoff, MacKenzie, Lee, and Podsakoff (2003) indicate that method variance may result from four different sources: rater characteristics, item characteristics, item context, and measurement context. In the context of the studies reviewed above, there are two sources of method variance that appear particularly likely to limit the ability of researchers to draw accurate inferences. The first source is based on characteristics of the items that have been used to measure work stressors in some of the prior research (e.g., Bhagat et al., 1985; Sarason & Johnson, 1979; Scheck et al., 1995). For example, Cavanaugh et al. (2000) have noted that many of the early operationalizations of positive and negative stressors may have biased the observed relationsh appraising a particular demand or circumstance (e.g., promotion) as positive and then evaluating

PAGE 38

38 a positive outcome (e.g., job satisfaction), a finding of an association may be inflated as a result constructs with items that are negatively or positively worded or that have response formats that require participants to evaluate items as negat ive or positive, and constructs that represent positive or negative outcomes may be influenced by this type of method bias. The second possible source of method variance in the studies reported in Table 2 1 results from characteristics of the individuals who are assessing work stressors and strains (e.g., physical or somatic illness, job dissatisfaction, absences, etc.) in these studies. Several researchers (Brief et al., 1988; Burke, Brief, & George, 1993; Payne, 1988; Schaubroeck, Ganster, & Fox, 1992) h perceives stressful events, as well as the way he or she responds to these events, and therefore should be controlled in studies examining these relationships. Con sistent with this position, Brief et al. (1988) found that controlling for negative affectivity significantly decreased the overall satisfaction, somatic complai nts at work, life satisfaction, and symptoms of depression. has been viewed by some as a surrogate of negative affectivity), the general lack of control for negativ e affectivity in the other studies reported in Table 2 1 may bias the interpretation of substantive relationships with challenge and hindrance stressor constructs. Potential Curvilinear Stressor Employee Criterion Relationships As noted previously, there a re two general perspectives on the relationships between work stressors and employee performance. The more popular of the two perspectives argues that stress is unequivocally negative for employees in the workplace (Jamal, 1984, 1985; Westman & Eden, 1985) . Generally speaking, this perspective lin ks increased levels of stress to employee

PAGE 39

39 dissatisfaction, poor performance, absenteeism, and turnover, suggesting that stress has detrimental effects on individual employees and the organization. In contrast, the second perspective hypothesizes that the relationship between stressors and performance is more complex. Supporters of this approach (Champoux, 1980, 1981; Janssen, 2001) argue that stress ors ha ve a curvilinear, inverted U shaped relationship with employee performance such that stressors have functional effects up to some optimal level, but after exceeding that level, become theory of arousal and performance as th e foundation for tests of this perspectiv e (cf. Sullivan & Bhagat, 1992), but have more recently used activation theory (Gardner, 1986; Gardner & Cummings, 1988; Scott, 1966) to argue for the curvilinear relationships between job related demands and employ ee psychological, affective, and behavioral responses. Unfortunately, studies examining curvilinear relationships between job demands and employee criteria have produced somewhat inconsistent results. For example, studies conducted by Champoux (1980, 1981 ), Edwards and Harrison (1993), Xie and Johns (1995), and Janssen (2001) have provided empirical support for curvilinear relationships between job demands and employee exhaustion, job satisfaction, internal work motivation, and performance criteria. In con trast, studies conducted by Jamal (1984, 1985), Friend (1982), and Westman and Eden (1996) have not found support for the curvilinear relationships between job stressors and employee performance criteria. The discrepancy in the findings makes it difficult to draw concrete conclusions concerning the linear and curvilinear relationships between job demands and employee outcomes. However, the recent research conducted on challenge and hindrance stressors provides a framework that may explain these inconsisten cies. Because hindrance stressors are comprised of

PAGE 40

40 characteristics in the workplace that are thought to be dysfunctional at almost any level (Jackson & Schuler, 1 985; Jamal, 1984, 1985 ), one would expect employees to experience negative effects from these stressors, regardless of the level at which they occur. In contrast, challenge stressors are comprised of job demands that have been found to be functional at some level, with respect to employee motivation, satisfaction, and performance (LePine et al., 20 05; Podsakoff et al., 2007 ). However, challenging job demands such as workload, time pressure, job complexity, and responsibility can, at extreme levels, also have dysfunctional effects on employee criteria. When employees perceive that their workload is t oo high, their deadlines are too aggressive, their responsibilities are excessive , and/or their job is too complex, they are likely to experience more strain, express increased dissatisfaction with their job, and exhibit lower levels of performance. In oth er words, it is possible that job demands that correspond to hindrance stressors in the work environment represent the negative, linear approach to stressor effects; whereas job demands that correspond to challenge stressors in the work environment represe nt the curvilinear perspective to stressor effects. If this is the case, then recognizing differences in the nature of work stressors may explain the inconsistent findings of previous studies on job stressor employee outcome relationships, and provide mana gers with a better understanding of the positive and negative effects of stress in the workplace. Research Examining Curvilinear Stressor Relationships A summary of the research examining the potential curvilinear effects of job related stress(ors) on emp loyee criteria is provided in Table 2 2. Generally speaking, for each study, this table reports the authors, the sample from which the data was obtained, the operationalizations of the stressors and employee criteria, the findings, and potential limitation s. An examination of the studies presented in this table reveals that tests of the curvilinear relationships between job demands and employee responses have produced somewhat mixed results.

PAGE 41

41 On the positive side, several studies have found support for a cur vilinear relationship between job demands and employee attitudes and performance criteria. For example, Champoux (1980, 1981) conducted a series of studies that provided preliminary support for the curvilinear relationships between job scope and employee a ffective responses. Champoux (1980) argued that although the breadth of skills required to complete a job should initially stimulate activity and lead to increased satisfaction and motivation in employees, jobs with too broad a scope of required skills mig ht lead to excessive stimulation and decreased levels of satisfaction and motivation. In addition, Champoux (1980) argued that growth need strength (GNS) should moderate the curvilinear job scope n (p. 467). Using data gathered from three organizations (one R&D firm and two federal agencies; total n = 1425), Champoux found general support for the hypothesis that job scope had curvilinear relationships with: (a) general employee satisfaction and growth satisfaction in two of the three organizations, and (b) the internal work motivation of employees in one of the organizations examined. However, he found no su pport for the hypothesized moderating effects of GNS on the curvilinear relationships between job scope and the affective responses in any of the organizations he studied. Champoux (1981) conducted a follow up study reporting similar results gathered fro m four organizations. Unfortunately, the data reported for three of these organizations were the same as support for the curvilinear relationships between job scope and internal work motivation and growth satisfaction in data gathered from a sample of 1152 employees from an R&D organization. Although he did not test the potential moderating effect of GNS on these

PAGE 42

42 curvilinear relationships, he did report that superviso ry satisfaction moderated the curvilinear relationship between job scope and growth satisfaction, but did not moderate the curvilinear relationship between job scope and internal work motivation in this new sample. De Jonge and Schaufeli (1998) also found some support for the curvilinear hypothesis in their examination of the relationships between job demands on employee strains in a sample of 1400 Dutch health care workers. These authors examined the potential linear and non linear relationships between a composite measure of job demands that confounded aspects of challenge with job demands (working hard and strenuous work), and employee satisfaction, anxiety, and exhaustion. De Jonge and Schaufeli reported that although job demands had linear relationships with all three of the employee criteria, they only had a significant curvilinear relationship with employee anxiety. Thus, although the results of this stud y provide some support for the hypothesized curvilinear relationships between work stressors and employee outcomes, the evidence is mixed. More recently, Janssen (2001) conducted a study examining the moderating effects of employee justice perceptions on curvilinear job demand employee criteria relationships. In this study, he hypothesized that employee perceptions of fairness will interact with job demands to influence job satisfaction and job performance such that: (a) effort reward ratios that are perce (Janssen, 2001, p. 1041) and lower levels of job performance. For the purposes of his study, Janssen assessed job demands with an eight item scale that consisted primarily of dimensions of challenge stressors (e.g., pace

PAGE 43

43 of work, time pressure, workload), but that worded items in such a way that they could be interpreted as assessing from 134 Dutch managers, Janssen (2001) reported that job demands had curvilinear relationships with employee work and supervisory satisfaction and employee innovative performance (as rated by the manager). In addition, consistent with his hypotheses, he also found that employee perceptions of fairness moderated the curvilinear relationships between job demands and all of the employee performance and satisfaction criteria. Janssen concluded that job demands have a curvilinear relationship with employee responses and that these relationships may be contingent upon other factors in the work environment. Two studies (Edwards & Harrison, 1993; Xie & Johns, 1995) using a person environment (P E) fit approach have also provided general support for curvilinear stressor emp loyee outcome relationships. Generally speaking, the P E fit approach argues that a misfit between the person and the environment will lead to increased levels of physical, psychological, and behavioral onship between the demands experienced by an employee and his or her ability to cope with these demands (D A fit; Edwards & Cooper, 1990). Using this perspective, Edwards and Harrison (1993) reported a study examining several tests of potential curvilinear relationships between work stressors and employee responses. Specifically, these authors reanalyzed data collected by French , Caplan, and Harrison (1982) on 2,010 respondents in 23 occupations to examine the linear and curvilinear relationships between em ployee job demands (job complexity, role ambiguity, responsibility for persons, and quantitative work load), their perceived ability to cope with such demands, and affective and

PAGE 44

44 psychological strains (employee job dissatisfaction, work load dissatisfaction , boredom, depression, anxiety, irritation, and somatic complaints). Using regression techniques that modeled fit, deficient fit, excess fit, poor fit, quadratic fit, and interactive fit terms as predictors of employee strains, these authors found general support for the hypothesized curvilinear stressor strain relationships for the challenge related stressors (job complexity, responsibility and quantitative workload), but not for the one hindrance related stressor (role ambiguity) they included in their st udy. More specifically, Edwards and Harrison (1993) reported that: (a) job complexity had curvilinear relationships with job and work load dissatisfaction, boredom, and anxiety; (b) responsibility for persons had curvilinear relationships with employee job and work load dissatisfaction and boredom; and (c) quantitative workload had curvilinear relationships with job and work load dissatisfaction. However, role ambiguity had a curvilinear relationship with only one employee outcome (i.e., anxiety). Generally speaking, these findings suggest that even after controlling for P E fit and the main effects of an related job demands (i.e., job complexity, responsibility for others, and wor k load) had important curvilinear relationships on employee affective and psychological outcomes, whereas the one form of hindrance related job demands examined (i.e., role ambiguity) did not exhibit appreciable curvilinear effects. Xie and Johns (1995) also found some support for curvilinear stressor employee criteria effects using a P E fit approach. These authors examined the moderating effect of job demands job holder ability (D A) fit on potential curvilinear relationships between employee job scope and emotional exhaustion and anxiety. Using a sample of 418 full time employees holding 143 different types of jobs, Xie and Johns reported that all three of the measures of job scope they

PAGE 45

45 gathered (self reported; coded from the Dictionary of Occupational Titles; and coded from the Occupational Prestige Index) had curvilinear relationships with employee emotional exhaustion, but that none had a curvilinear relationship with anxiety. In addition, these authors also reported that D A fit significantly modera ted the curvilinear relationships between measures of job scope and exhaustion, but did not moderate curvilinear relationships between measures of job scope and employee anxiety. Consistent with the findings of Edwards and Harrison (1993), this study also suggests that challenge related job demands may have curvilinear relationships with employee strains, but that these relationships may be influenced by the fit of employee abilities to these demands. However, not all the studies designed to examine the cu rvilinear relationships between job stressors and employee criteria have been supportive of the hypotheses. For example, Jamal (1984) reported that he found little support for the inverted U hypothesis in a study of 440 Canadian nurses. Specifically, he ex ambiguity, role conflict, role overload, and resource inadequacy and their job effectiveness (job performance, job motivation, and patient care), absenteeism, tardiness, and anticipated turnover. Jama l reported support for only 4 of the 24 curvilinear relationships he examined. More specifically, he reported that: (a) role conflict and overload were not found to exhibit curvilinear relationships with any of the employee criteria; (b) role ambiguity was found to have curvilinear relationships with only two dimensions of employee effectiveness and employee absenteeism, and (c) resource inadequacy was only found to have a curvilinear relationship with absenteeism. However, Jamal did report that many of the job stressors had negative, linear relationships with of a negative linear relationship between stress and performance may be a surprise to many as it

PAGE 46

46 tends to be contrary to the Yerkes He explained these findings by arguing that role ambiguity, role conflict, role overload, and cies employees have for better performance, and concluded that a different set of stressors should be examined in the work environment to test for potential positive or curvilinear effects. Jamal (1985) reported similar findings in a sample of 227 middle managers and a sample of 283 blue collar workers from a large Canadian organization. Building on his earlier study, he examined the relationships between job demands (role ambiguity, role conflict, role overload, and resource inadequacies, and a composite index comprised of all four stressors) and job performance (operationalized as the quantity of performance, quality of performance, and effort exerted). Generally consistent with the findings of his earlier study, Jamal (1985) reported that even th ough all four stressors (and a composite stress measure) had negative, linear relationships with the three dimensions of performance, none of the 20 relationships between the stressors and the quantity and quality of performance across the two samples were curvili near, and that only role overload and resource inadequacy had a curvilinear relationship with effort expended. Role overload was found to have a curvilinear relationship with effort expended in both the managerial and blue collar sample, and resource inade quacies was found to have curvilinear relationship with this criterion in the blue collar sample. Therefore, when compared to the suggest that the nature of the rel ationships between work stressors and performance may depend on the nature of the stressor that is examined. Friend (1982) also found little support for the hypothesized curvilinear relationships between job demands and performance. He examined the relati onships between subjective

PAGE 47

47 workload and time urgency and the exam performance of 39 participants in a management training course. He reported that although both workloa d and time urgency had negative linear relationships with exam performance, neither of t hese demands had significant curvilinear relationships with performance. However, given the small size of the sample used in this study , it is likely that low power is somewhat responsible for the lack of support received for the curvilinear relationships between participant demands and performance. Finally, using a sample of 306 officer cadets in the Israeli Defense Forces to test propositions of the P E fit approach to stress, Westman and Eden (1996) reported that although a single item measure of work r elated demands had a linear relationship with a composite measure of cadet performance (including combat doctrine, navigation, leadership, etc.), there was no support for a curvilinear relationship between cadet stress and performance. Although these autho rs acknowledge potential limitations of their study, they argued that stress has a negative, linear effect on performance in work settings. After comparing their findings with those of previous re search, Westman and Eden concluded that the general lack of support for curvilinear stress effects in field research suggests that appropriate tests of this hypothesis should be restricted to the laboratory. Limitations of Previous Research Examining Curvilinear Stressor R elationships Despite the conclusions reach ed by Westman and Eden (1996), a careful examination of the findings of the studies reported in Table 2 2 suggests that these conclusions are premature for at least two major reasons. The first reason is that these studies provide general support for the n otion that some types of work demands have curvilinear relationships with employee psychological, affective, and behavioral responses, whereas other types of stressors have negative, linear relationships with these criteria. For example, it appears that in those studies that focused primarily on hindrance related stressors (Jamal, 1984, 1985), or that used a composite

PAGE 48

48 measure that confounded the c hallenge and hindrance stressors (Westman & Eden, 1996), linear stressor outcome variable relationships were mos t often observed. However, in those studies that focused either primarily on challenge stressors (Champoux, 1980, 1981; Janssen, 2001; Xie & Johns, 1995), or distinguished between challenge and hindrance stressors in their analyses (Edwards & Harrison, 199 3), curvilinear relationships were often observed. The one obvious exception to this is the study reported by Friend (1982), who found no evidence of curvilinear relationships, even though he examined two forms of stressors (workload and time urgency) gene rally conceptualized as dimensions of the challenge stressor domain. However, as noted earlier, the small sample size included in this study may have been responsible for the findings that Friend (1982) observed. Therefore, when taken together, the evidenc e suggests that even though the relationships between hindrance related stressors and employee outcome variables seem to be linear, the relationships between challenge related stressors and these criteria may be curvilinear in nature. The second major rea is that the studies reported in Table 2 2 have some important limitations which may have contributed to inconsistencies in the findings regarding the curvilinear effects. The first limitation r elates to the operationalizations of the job demands constructs examined in these studies. For example, Westman and Eden used a one item global measure of stress to assess the demands experienced by the officer cadets in their study. In addition to concern s about the reliability and validity of one item measures (Churchill, 1979; Schwab, 1980; Venkatraman & separate challenge related from hindrance related forms of st ressors. Indeed, given that this measure was designed to represent a global assessment of job demands, it is likely that

PAGE 49

49 participants considered both challenge and hindrance stressors when responding to this item, thereby producing a confounded measure of perceived stress. Related to this, other studies (Jamal, 1984, 1985) have focused predominately on hindrance stressors, thereby providing inadequate tests of the potential curvilinear stressor strain relationships. The second limitation of the studies rep orted in Table 2 2 is that they do not systematically examine the curvilinear relationships between various job demands and specific employee psychological, attitudinal, and behavioral responses. This results from two problems. The first is that tests of t he hypothesized curvilinear relationships have been restricted to a relatively small set of job demands and employee criteria. For example, although several studies (Champoux, 1980, 1981; De Jonge & Schaufeli, 1998; Edwards & Harrison, 1993) have examined the curvilinear relationships between job scope and facets of job satisfaction, relatively few studies have examined the curvilinear relationships between time urgency and employee performance (Friend, 1982) or between resource inadequacies and absenteeism or tardiness (Jamal, 1984). In addition, no study conducted to date has examined the potential curvilinear relationships between organizational politics, interpersonal conflict, or job responsibility, and any employee criteria. The second problem is that very few studies have replicated the specific job demand employee criteria relationships that have been studied. For example, the findings reported in performance of 39 management trainees have not been replicated with a larger sample. This presents a problem because the results of a single study, whether it supports or does not support the curvilinear hypothesis, may be sample specific. The final limitation of the st udies reported in Table 2 2 is that all of them are restricted in their use of between subjects designs. Although study designs examining between subjects

PAGE 50

50 relationships comprise the vast majority of research conducted in the stress literature, and the OB f ield as a whole, there are good reasons to believe that a study combining both within and between subject components would provide a more appropriate test of the curvilinear effects of stressors on employee outcomes. Lazarus and Folkman (1984) have argued that in order to gain a better understanding of the effe cts of stress on human behavior, It is necessary to observe the same person again and again . . . [which] requires comparing the person with himself or herself at different times and under different conditions. This intraindividual perspective contrasts with interindividual comparisons of that person with other persons under common conditions . . . An ideal alternative is to observe individuals repeatedly intraindividually and do interindividual compa risons (pp. 299 300). This suggests that a research design that permits the combination of between subject and within subject analyses should be superior to a design that uses only one or the other. Although within subject research designs are becoming mo re popular in the stress domain (Alliger & Williams, 1993; Fuller, Stanton, Fisher, Spitzmuller, Russell, & Smith, 2003; Sonnentag & Zijlstra, 2006), as well as the general OB literature (Ilies & Judge, 2002; Judge & Ilies, 2004; Judge, Scott, & Ilies, 200 6), no study to date has used this type of design to examine the potential curvilinear relationships between job demands and employee outcomes. In summary, there is evidence suggesting that many of the previous studies examining job demand employee criter ia relationships are potentially limited by inappropriate stressor operationalizations, a fairly restricted set of job demands employee criteria relationships, and the between subjects nature of their research designs. In addition, when the previous studie s are examined carefully, it appears that those studies that have focused on challenge related stressors generally have reported more consistent curvilinear relationships than those that have focused on hindrance related stressors. Therefore, a more approp riate test of the curvilinear relationships between job demands and employee criteria will require: (a) using measures that specifically assess challenge and hindrance stressors, (b) a comprehensive examination of the linear and

PAGE 51

51 curvilinear relationships b etween these stressors and employee psychological strains, job attitudes, and performance criteria, and (c) the analysis of both within and between subjects relationships between job stressors and employee outcomes. Hypotheses In the following section, I will describe how theory developed by Lazarus and Folkman (1984) can be used to explain the linear and curvilinear relationships between challenge and hindrance stressors and job strains, satisfaction, and performance. In addition, I will also provide the rationale for the potential moderating effects of three individual differences (negative affect, ability, and locus of control) on the stressor employee criterion relationships. Finally, I will develop hypotheses for the interactive effects of challenge a nd hindrance stressors on the employee criteria. Criteria Relationships According to activation theory (Gardner, 1986; Gardner & Cummings, 1988; Scott, 1966), reticular activation system (RAS), and is related to the total stimulation that an individual receives from internal (e.g., heart rate and other physiological responses), exter nal (environmental stimuli and reinforcing events) and cortical (e.g., cognitions) sources. Although often viewed as synonymous with each other, Scott (1966) has noted that activation can be distinguished from arousal in that activation is a function of th e central nervous system whereas arousal is primarily a function of the peripheral (autonomic) nervous system. This means that arousal can be both a determinant and manifestation of activation level (Gardner, 1986; Gardner & Cummings, 1988; Scott, 1966).

PAGE 52

52 A t least three propositions can be derived from activation theory that may help explain the differential relationships between challenge and hindrance stressors and employee perceptions of strains, job attitudes, and job performance. The first of these prop ositions is that each person has a characteristic level of activation (CLA) that they are motivated to maintain (Scott, 1966). An experienced during the course o f their life. This suggests that although the CLA of an individual 86). Scott (1966) has referred to these behaviors as impact modifying behaviors and noted that they may include those behaviors that increase stimulation (i.e., impact increasing behaviors) or those t hat decrease stimulation (i.e., impact decreasing behavior). The second proposition is that for any given stressor that a person encounters, there should be an inverted U shaped relationship between experienced activation level and the level of task per formance. This suggests that positive or negative deviations from the CLA will result in decrements in performance. This also suggests that when deviations from the CLA occur, high levels of employee performance can only be maintained if the impact modifyi ng behaviors do or exhibiting increased effort. In this case, the impact increasing behaviors resulting from a deviation from the CLA will probably result in continued levels of high performance. In contrast, avoiding the task com pletely, it is unlikely that high performance will be sustained. In others

PAGE 53

53 words, the impact decreasing behaviors in this case would be expected to result in lower levels of performance. Finally, the third proposition is that the total stimulation that on e experiences from a job demand is not only the result of the demand itself, but also of other events that have been associated with it. In other words, in addition to the direct influence that job demands such as work load, job complexity, role ambiguity, arousal, these job demands are also expected to influence arousal because of the reinforcing events with which they are associated (Gardner & Cummings, 1988). For example, jobs that are ambiguous not only complete his or her work, but also because these demands make it unclear as to what the employee should do to receive rewards for his or her performance, and therefore are negativ ely reinforcing. As another example, increased levels of job responsibility should not only increase arousal because they place increased demands on the individual, but also because this type of demand is typically associated with positive reinforcers such as promotion, increased compensation, and recognition. Therefore, the overall effect of any work stressor on an activation level as well as the relationship s between these job demands and the reinforcing events that are associated with them. Relationships among Challenge Stressors and Employee Criteria The theoretical framework provided above can be used to explain the differences in the relationships betwee n challenge and hindrance stressors and employee strain, job satisfaction, and performance. In the case of challenge related stressors, low levels of job demands are generally expected to cause increased levels of arousal, because of: (a) the direct effect s that these demands have on the individual, as well as (b) the positive reinforcing events that have

PAGE 54

54 previously been associated with these job demands. Assuming that an individual can sustain arousal at their CLA, he or she may be able to perform at a hig h level for an extended period of time. However, if the job demand continues to increase and as a result, is associated with other reinforcing events that cause arousal to increase beyond the CLA, one would expect that performance decrements will occur. Fo r example, demands related to job scope, or the variety of activities one is required to perform on a job, are hypothesized to have an initially motivating effect on employees (Hackman & Oldham, 1980). That is, employees who have jobs that require a broade r range of activities and are more complex in nature are initially expected to experience lower levels of strain, express higher levels of satisfaction with their job, and exhibit higher levels of performance than will employees in jobs that are less compl ex or are comprised of a monotonous set of tasks (Hackman & Oldham, 1980). According to my activation theory framework, these beneficial employee outcomes are a result of the direct stimulating effect that demands related to job scope have on employee arou sal and the positive reinforcement that is associated with job complexity (e.g., opportunities for promotion, recognition, feelings of accomplishment and pride). However, job complexity may also result in overstimulation if demands related to the variety of tasks one is required to perform become excessive. For example, an employee whose job requires too broad a variety of tasks or is too complex in nature is likely to experience increased levels of strain and decreased levels of job satisfaction and emplo yee performance. Consistent with the activation theory framework described above, employee strain and dissatisfaction may be a result of the overstimulation due to arousal levels that exceed the CLA. There are three possible reasons why arousal levels abov e the CLA may cause decrements in employee performance. The first is that high levels of arousal have been shown to cause

PAGE 55

55 impaired information processing capabilities (Easterbrook, 1959; Hockey, 1979; Humphries & Revelle, 1984). This suggests that employee s experiencing high levels of challenge stressors are less able to process the information needed to perform their job effectively. Second, when arousal levels exceed the CLA, employees may exhibit avoidance or withdrawal (impact decreasing) behaviors that are inconsistent with high levels of performance. Finally, challenge stressors that consistently produce high levels of stimulation should eventually acquire associations with negatively reinforcing events. For example, an employee whose job continues to increase in complexity, responsibility, and/or workload should eventually associate these job demands with physical and mental fatigue (e.g., increased susceptibility to illnesses), conflicts in maintaining work life balance (e.g., disagreements with spous e or family members spent at work or away from home), and other negative reinforcers. Taken together, all of these reasons provide potential explanations for how high levels of challenge stressors may lead to decreased levels of employee task performance. In addition, there are reasons to believe that there are curvilinear relationships between challenge stressors and voluntary behaviors related to employee performance in the workplace, such as OCB and CWB. Similar to their relationship with task performan ce, challenge stressors are expected to have an inverted U shaped relationship with OCB. To the extent that challenge develop themselves and (b) feelings of s atisfaction with the job, employees should feel an obligation to give back to the organization in the form of OCB (Gouldner, 1960). For example, organizations that challenge their employees with jobs that require a variety of skills and abilities and high levels of responsibility should also have employees who are willing to reciprocate the organization through the performance of OCB. However, if the challenging demands become so

PAGE 56

56 extreme that the employee does not believe he or she can meet them, behaviors that are not part of the formally specified job description may be the first to be withdrawn (Hui, Organ, & Crooker, 1994). When employees feel that they are overloaded with role related activities and have tight deadlines, it is expected that they that wi ll focus more attention and energy on the tasks related to their role in the organization and forgo extra role activities. In contrast, the curvilinear relationship between challenge stressors and CWB is expected to be U shaped such that extremely low an d high levels of challenging demands will result in retaliatory behaviors on the part of employees, but that an intermediate level of challenge stressors will result in lower forms of these behaviors. When employees are not challenged by the demands of the ir job (e.g., they have no work to occupy their time, no deadlines to meet, no job responsibilities, and/or only perform routine tasks), they may exhibit CWB to cope with the monotony of the work experience. For example, employees may deal with the unchall enging, routine aspects of a job through horseplay with coworkers and physical or mental withdrawal workloads are perceived to be too high, whose deadlines ar e too aggressive and who simply have too much responsibility and complexity in their jobs are also likely to cope with this situation through withdrawal and other CWBs. Although these extremes lie at different ends of the continuum, employee responses to v ery high and very low levels of challenge stressors may be a result of feelings that the organization does not appreciate them or that they are being unfairly treated. In the case of very low levels of challenge stressors, employees may be unsatisfied with their jobs because they perceive that: (a) they are not using the skills and abilities they have been trained to use, (b) they deserve a job that is challenging and provides them with the potential for growth and self development, and/or (c) there are oth er jobs or managers that would provide

PAGE 57

57 them with a more challenging work environment. Likewise, in the case of very high levels of challenge stressors, employee s may feel dissatisfied because: (a) they are overwhelmed by the sheer number of different abili ties and skills necessary to perform the job, (b) the organizational unfairly expects them to perform a job that exceeds their abilities, and/or (c) other organizations would assign them to a job with more reasonable work demands. All of these factors may contribute to employee perceptions of injustice, which have been found to be some of the strongest predictors of CWBs such as withdrawal and theft (Colquitt, Conlon, Wesson, Porter, & Ng, 2001; Greenberg, 1990). In contrast, employees who experience interm ediate levels of challenging stressors are expected to exhibit lower levels of CWB because these employees will perceive their work to be motivating, satisfying, and fair, and subsequently, should not be expected to exhibit retaliatory behaviors that would harm the organization or people in it. In summary, I hypothesize that: H ypothesis 1: Challenge stressors have a positive, linear relationship with employee strains. H ypothesis 2: Challenge stressors have a positive, linear relationship with employee job s atisfaction H ypothesis 3: Challenge stressors have a positive, linear relationship with employee task performance. H ypothesis 4: Challenge stressors have a positive, linear relationship with employee OCB. H ypothesis 5: Challenge stressors have a negative, linear relationship with employee CWB. H ypothesis 6: Challenge stressors have a curvilinear, U shaped relationship with employee strains. H ypothesis 7: Challenge stressors have a curvilinear, inverted U shaped relationship with employee job satisfaction H ypothesis 8: Challenge stressors have a curvilinear, inverted U shaped relationship with employee task performance.

PAGE 58

58 H ypothesis 9: Challenge stressors have a curvilinear, inverted U shaped relationship with employee OCB. H ypothesis 10: Challenge stressors have a curvilinear, U shaped relationship with employee CWB. Relationships among Hindrance Stressors and Employee Criteria In contrast, hindrance related stressors are generally expected to have negative, linear relationships with employee satisfaction and performance, but a positive, linear relationship with job strains. This is expected to be the case for two basic reasons. The first reason is that employee functioning appears to be highest when these demands are not present in the ment. This would suggest that th e presence of these job demands job dissatisfaction, and poor performance. In other words, the arousal associated with employe e perceptions of hindrance stressors is generally expected to exceed the CLA of individuals and therefore the relationships between these stressors and employee satisfaction and performance is negative. This expectation is generally consistent with researc h on the effects that negative events have on brain activation levels. For example, in a study designed to measure event related potentials during the evaluative categorization of . . . negative as compared with positive stimuli, also consistent with research reported by Baumeister, Bratslavsky, Finkenauer, and Vohs, (2001 ) and Rozin and Royzman (2001) indicating that negative events are qualitatively different than positive events in that negative events have stronger effects on the activation centers of the brain, as well as on the emotions, learning, and information proc essing of individuals than do positive events.

PAGE 59

59 The second reason that hindrance stressors are expected to have negative, linear relationships with employee satisfaction and performance (and positive relationships with job strains) is that these types of st ressors are generally associated with negatively reinforcing events. For example, decisions in the organization that are perceived by employees to be politically driven (e.g., promotions or raises are given to those who are favored and not those who deserv e it) will generally result in high performing employees receiving fewer rewards than they deserve, perceptions of unfair treatment, and subsequent dissatisfaction. Similarly, when employees are provided inadequate resources to complete their assigned task s, this is likely to be associated with a number of negative consequences including decreased expectancies that effort will lead to performance and valued outcomes, lower levels of self esteem, and/or a decreased sense of accomplishment because of the empl consistent with research reported by Jamal (1984, 1985) who found that hindrance stressors such as role ambiguity, role conflict, and resource inadequacies were generally found to have significant negativ e, linear (but not significant curvilinear) relationships with employee performance, absenteeism, and anticipated turnover. It is also consistent with meta analytic evidence reported by LePine et al. (2005) indicating that hindrance related stressors were negatively related to employee task performance. Hindrance stressors are also expected to have negative, linear relationships with employee OCB. Because perceptions of hindrance stressors are negatively related to job satisfaction and positively related t o strains, it is expected that high levels of these demands will lead to lower levels of employee OCB. The dissatisfaction that accompanies hindering demands such as role ambiguity and conflict, inadequate resources, organizational politics, administrative hassles, interpersonal conflict, and job insecurity should make employees less willing to perform extra -

PAGE 60

60 role behaviors that are not part of their job description. This hypothesis has been supported by meta analytic research reported by Podsakoff, MacKenzi e, Paine, & Bachrach (2000), in that role ambiguity and role conflict were found to have negative relationships with dimensions of the OCB construct (e.g., altruism and courtesy). In addition, employees that experience hindrance stressors will also be less likely to exhibit OCB because of the debilitating effects of strains associated with these job demands. For example, Cropanzano, Rupp, and Byrne (2003) reported that emotional exhaustion was negatively related to OCB aimed at the organization and OCB aime d at the supervisor in a sample of 232 employees drawn from a variety of industries including human services, manufacturing, and fitness. In contrast, hindrance stressors are expected to have a positive, linear relationship with CWB through their effects on strains and job dissatisfaction. Research has consistently shown that employees who are dissatisfied with their jobs are much more likely to retaliate against their organizations or the people in them (Colquitt et al., 2001; Dalal, 2005). In addition, t he job related strains associated with hindrance stressors are also expected to result in higher levels of employee CWB. Indeed, Spector and his colleagues (Chen & Spector, 1992; Fox & Spector, 1999; Fox, Spector, & Miles, 2001; Storms & Spector, 1997) hav e reported findings from several empirical studies which indicate that job related strains such as anger, anxiety, frustration, and depression were associated with aggression, hostility, sabotage, withdrawal, and other forms of employee CWB. Therefore, it is expected that the strains and dissatisfaction associated with hindrance stressors will lead to behaviors that are counterproductive for the organization and those in it. This has been supported by met a analytic evidence reported by Podsakoff et al. (200 7) which indicates that hindrance related stressors were positively related to employee withdrawal behaviors. Therefore, I hypothesize that:

PAGE 61

61 H ypothesis 11: Hindrance stressors have a positive, linear relationship with strains. H ypothesis 12: Hindrance stre ssors have a negative, linear relationship with job satisfaction. H ypothesis 13: Hindrance stressors have a negative, linear relationship with task performance. H ypothesis 14: Hindrance stressors have a negative, linear relationship with OCB. H ypothesis 15 : Hindrance stressors have a positive, linear relationship with CWB. Moderators of the Stressor Criteria Relationships In addition to the main effects of challenge and hindrance stressors on employee criteria variables that I have hypothesized above, I als o expect three individual difference variables (e mployee affect, ability, and locus of control) to moderate these relationships. In the section that follows, I provide the rationale for these expectations. The moderating effects of negative affectivity on stressor criteria relationships Negative affect is the first hypothesized moderator of the relationships between challenge and hindrance stressors and employee outcomes. Although the role of employee affect in the stress process is not completely clear (cf . Barsky, Thoresen, Warren, & Kaplan, 2004; Spector, Zapf, Chen, & Frese, 2000), some researchers (Bolger & Zuckerman, 1995; Spector et al., 2000), environment and h ave more pronounced responses to stressors than will employees low in NA. In contrast to exposure based (Bolger & Zuckerman, 1995) or perception based (Spector et al., 2000) models of NA stressor odels (Barsky et al., 2004) hypothesize that high NA employees react more strongly to stressors in the workplace, which will result in higher levels of exhaustion, somatic illnesses, and frustration. rgument that affective disposition is the tendency to experience positive or negative emotions in response to the environment. General

PAGE 62

62 support for the moderating effects of NA on stressor strain relationships have been found in empirical studies conducted by Parkes (1990) and Moyle (1995). Therefore, consistent with the exacerbation model, I expect that employees with higher levels of NA will be more vulnerable to the negative effects of hindrance stressors such that these stressors will exhibit stronger po sitive relationships with strains and CWB , and stronger negative relationships with satisfaction, in role behavior, and OCB than will employees with lower levels of NA. Although much less research has been directed at examining the relationships between ch allenge stressors and job satisfaction and performance, it is expected that NA will also influence the strength of these relationships. Because high NA individuals tend to perceive the world is a more negative light, it is expected that demands related to challenge stressors will lead to more detrimental levels of strain for these individuals than individuals who have lower levels of NA. Using the same rationale, one would also expect that high NA employees experiencing challenge stressors will also express lower levels of satisfaction with their jobs, exhibit lower levels of task performance and OCBs , and exhibit higher levels of CWB than will employees with lower levels of NA. Therefore, I hypothesize: Hypothesis 16 : Employee negative affectivity moderates the relationships between challenge stressors and (a) strains, (b) satisfaction, (c) task performance , (d) OCB, and (e) CWB. More specifically, the positive relationship between challenge stressors and strains and the negative relationship between challen ge stressors and CWB will be stronger; and the positive relationships between challenge stressors an d satisfaction, task performance , and OCB will become weaker for individuals with high NA than for individuals with low NA. H ypothesis 17 : Employee negati ve affectivity moderates the relationships between hindrance stressors and (a) strains, (b) satisfaction, (c) task performance , (d) OCB, and (e) CWB. More specifically, the positive relationships between hindrance stressors and job strains and CWB; and the negative relationships between hindrance stressors and satisfaction, task performance , and OCB will be stronger for individuals with high NA than for individuals with low NA.

PAGE 63

63 The moderating effects of employee ability on stressor criteria relationships Consistent with the D A fit approach to stress (Edwards & Cooper, 1990; Edwards & Harrison, 1993), I also expect that employee ability will moderate the relationships between work stressors and employee outcomes. French et al. (1982) have argued that when there is a misfit between environmental demands and the abilities of employees, it should result in higher levels of strain, less satisfaction, and lower performance. This suggests that employees who have higher levels of ability will experience less strai n, more satisfaction, and outperform employees who have lower levels of ability but experience a similar level of these demands. In other words, employee ability is expected to buffer the effects of challenging work demands on strains such that the positiv e relationship between challenge stressors and strains will be weaker for higher ability employees than for employees with lower levels of ability. In addition, since employees with higher abilities should be able to cope with work demands more effectively that employees with lower abilities, I expect that there will be a stronger positive relationship between challenge stressors and employee satisfaction for employees who possess higher abilities than for those that possess lower abilities. For similar rea sons, employee ability is also expected to buffer the negative effects that hindrance stressors have on employee perceptions of strains and job satisfaction. Because employees with higher ability are better able to adapt to the demands of their jobs, the p ositive relationship between hindrance stress and strains and the negative relationship between hindrance stressors and job satisfaction should be weaker for high ability employees than for lower ability employees. Finally, employee ability is also expec ted to moderate the relationships between challenge and hindrance stressors and task performance, OCB , and CWB. Because employee ability is a critical determinant of task performance in virtually all jobs (Hunter & Hunter, 1984), employees with higher leve ls of ability are expected to be better equipped to cope with the challenging

PAGE 64

64 demands they experience, and perform at higher levels than employees with lower levels of ability. Similarly, employee ability is also expected to decrease the negative effects o f hindrance stressors on employee performance, such that employees with high ability will suffer less of a performance decrement from these types of demands than will employees with lower levels of ability. Because employees with higher ability are able to cope more effectively with task related behaviors, they should also have more time and resources to devote to functional extra role behaviors . This would suggest that employees that have high ability and who experience challenge stressors or hindrance str essors will be more likely to exhibit OCB than will employees with lower levels of ability who find themselves subjected to these same work demands. Finally, because employees who have high ability are better able to cope with work demands than people with lower ability, they should also be less likely to perceive that demands being placed on them are unfair. Subsequently, I expect that the relationship between challenge stressors and CWB will be weaker for high ability people than for low ability people; w hereas the opposite should be true in the case of the relationship between hindrance stressors and CWB. That is, high ability employees exposed to hindrance stressors will be less likely to exhibit CWB than will employees of low ability who are subjected t o the same level of hindering job demands. Thus, I hypothesize that: H ypothesis 18 : Employee ability moderates the relationships between challenge stressors and (a) strains, (b) satisfaction, (c) task performance, (d) OCB, and (e) CWB. More specifically, t he positive relationships between challenge stressors and satisfaction, task performance , and OCB will be stronger for individuals with high ability than those with low ability; whereas the negative relationship between challenge stressors and CWB and the positive relationship between challenge stressors and strains will be weaker for individuals with high ability than for those individuals with low ability. H19 : Employee ability moderates the relationships between hindrance stressors and (a) strains, (b) satisfaction, (c) task performance, (d) OCB, and (e) CWB. More specifically, the negative relationships between hindrance stressors and satisfaction, task performance , and OCB and the positive relationships between hindrance

PAGE 65

65 stressors and strains and CWB w ill be weaker for individuals with high ability than for individuals with low ability. The moderating effects of employee locus of control on stressor criteria relationships The third individual difference that I expect to moderate the relationships betw een challenge and hindrance stressors and employee outcomes is locus of control. According to Rotter (1966), an individual with an internal locus of control generally believes that he or she is in control of events and outcomes of importance, such that the y are contingent upon his/her own behavior; whereas an individual with an external locus of control generally believes that he or she is not in control of events or outcomes in their life such that they are not contingent upon his/her actions, but upon cha nce, luck, fate, or the actions of others. Because Lazarus and Folkman (1984) have argued that the extent to which people feel they have control or mastery (or a lack thereof) over a demanding situation will influence the way stress affects them, it is exp More specifically, I expect that employees with an internal locus of control will experience less negative effects from the work demands they experience, because they are more likely to exert direct action against the stressor than are employees with an external locus of control (Kahn & Byosiere, 1992; Sonnentag & Frese, 2003). In the case of challenge stressors, employee with an internal locus of control are expected to experience less strain, exhibit lower levels of CWB, express higher levels of satisfaction with their jobs, and exhibit higher levels of in role and citizenships be haviors than employees who are exposed to challenge stressors but have an external locus of control. Consistent with these expectations, Brissett and Nowicki (1973) found that individuals with an internal locus of control reported lower levels of frustrati on than did individuals with an external locus of control.

PAGE 66

66 Although less research has been conducted on the moderating effects of employee locus of control on stressor employee behavior relationships, there are good reasons to believe that these effects e xist (Jex, 1998). For example, Storms and Spector (1987) reported data taken from 160 mental health care employees indicating that employee locus of control moderated the relationship between self reported frustration at work and employee sabotage such tha t employees with an internal locus of control had weaker relationships between these variables than did employees with an external locus of control. Consistent with propositions put forth by Allen and Greenberger (1980), Storms and Spector argued that indi viduals with lower levels of perceived control may attempt to modify their environment through acts of destruction or deviance. These types of behavior may offer employees an increased sense of control over their environment while also reducing employee te nsion and frustration. In addition, Storms and Spector also argued that employees with an internal locus of concerns; Jex, 1998) than employees with an external l ocus of control. Butterfield (1964) reported that frustrated subjects in an experiment who had an internal locus of control provided more constructive responses that did subjects having an external locus of control who were in the same situation. Finally, the study reported by Brissett and Nowicki (1973) found that participants having an internal locus of control were more likely to put forth increased effort when frustrated than were those having an external locus of control. For similar reasons to those s tated above, I also expect that employees with an internal locus of control who are exposed to hindrance stressors will also experience lower levels of strain, exhibit lower levels of CWB, be less dissatisfied with their jobs, and suffer lower decrements i n task performance and OCB than will

PAGE 67

67 employees exposed to similar levels of these stressors who have an external locus of control. In summary, I hypothesize that: H ypothesis 20 : Employee locus of control moderates the relationships between challenge stres sors and (a) strains, (b) satisfaction, (c) task performance, (d) OCB, and (e) CWB. More specifically, the positive relationships between challenge stressors an d satisfaction, task performance , and OCB, and the negative relationship between challenge stres sors and CWB will be stronger for individuals with high internal locus of control as opposed to those with high external locus of control; whereas the positive relationship between challenge stressors and strains will be weaker for individuals with high in ternal locus of control as opposed to those with high external locus of control. H ypothesis 21 : Employee locus of control moderates the relationships between hindrance stressors and (a) strains, (b) satisfaction, (c) task performance, (d) OCB, and (e) CW B. More specifically, the negative relationships between hindrance stressors and satisfaction, in role behavior, and OCB, and the positive relationships between hindrance stressors and strains and CWB will be weaker for individuals with high internal locus of control than for individuals with high external locus of control. Interactive challenge hindrance stressor effects Finally, the interaction between work demands related to challenge stressors and work demands related to hindrance stressors is also ex pected to have an influence on employee outcomes. Consistent with resource based (Hobfoll, 1989), P E fit (Edwards & Cooper, 1990), and arousal based (Yerkes & Dodson, 1908) approaches to stress, I expect that employees will experience higher levels of str ain and lower levels of satisfaction and performance when their work related demands are too high and exceed their capacity to cope with these demands. The combination of both of these types of demands is likely to result in employees feeling that they are overwhelmed, which may lead to anxiety, exhaustion, and frustration over the inability to meet demands. It may also be the case that employees experiencing high levels of challenging and hindering stressors will experience strong deleterious effects beca use they perceive this situation to be unfair. It seems logical to assume that employees experiencing extremely challenging

PAGE 68

68 workloads and deadlines in the complex jobs they have will feel more strain, less satisfaction, and lower levels of performance when their ability to meet these demands is hindered by inadequate resources, unclear or conflicting information regarding their role in the organization, and/or feelings that their job may be insecure. This is generally consistent with research in the organiz ation justice literature (Colq uitt et al., 2001), which has found the employee perceptions of unfair treatment are related to strains, dissatisfa ction, task performance, and OCB. In addition, and also consistent with this literature, I expect that heighten ed levels of employee frustration and/or perceptions of unfair treatment associated with simultaneously high levels of challenge and hindrance stressors will lead employees to exhibit higher levels of CWB than those experiencing lower levels of hindrance s tressors. In other words, employees experiencing a combination of high levels of hindrance related demands and challenge related demands in the workplace should exhibit increased levels of employee strain and CWB, and l ower levels of satisfaction, task per formance, and citizenship behaviors, than will employees experiencing high levels of challenge stressors that are unobstructed by hindrance stressors. Therefore, I hypothesize that: H ypothesis 22 : Hindrance stressors will moderate the relationship betwee n challenge stressors and (a) strains, (b) satisfaction, (c) task performance, (d) OCB, and (e) CWB. More specifically, the positive relationships between challenge stressors and satisfaction, task performance and OCBs will be weaker for employees exposed to higher levels of hindrance stressors than for those exposed to lower levels of hindrance stressors; the negative relationship between challenge stressors and CWB will be stronger for employees exposed to higher levels of hindrance stressors than for tho se exposed to lower levels of hindrance stressors; and the positive relationships between challenge stressors and strains will be stronger for employees exposed to higher levels of hindrance stressors than for those exposed to lower levels of hindrance str essors.

PAGE 69

69 Table 2 1. Summary of selected studies examining positive versus negative forms of stress. Researcher(s) Sample Conceptual definition of stressor (s) Operationalization Findings Limitations Sarason & Johnson (1979) 44 male naval personnel No conceptual definition was provided for positive stressors. The Life Experience Survey (LES) is a 57 item index that allows participants to indicate the positive or negative impact that various events had on their life. Examples include: Marriage/divorce Personal illness Death of a family member Pregnancy 30 item organizational stress inventory including: Change in work schedule Transfer Conflict with worker New commanding officer Positive life stress was positively related to promotion satisfaction an d negative life stress was negatively related to work, supervisor, and pay satisfaction. Positive organizational stress was positively related to work and people satisfaction, and negative organizational stress was negatively related to work and people sat isfaction. No theoretical framework for which stressors should be positive or negative. Positive and negative stress scales are confounded with stressors occurring outside the work environment. Did not control for the potentially biasing effects of resp affect. No conceptual definition was provided for negative stressors. Kanner, Coyne, Schaefer, & Lazarus (1981) 100 respondents ranging in age from 45 to 64 years of age positive experiences such as joy derive d from manifestations of love, relief at hearing good news, the pleasure 135 item uplifts index developed by the authors. Examples include: Relaxing Spending time with family Using skills well at work Praying Th e frequency and intensity of daily uplifts was positive related to positive affect and the frequency of daily hassles was positively related to negative affect. Hassles and uplifts scales are confounded with stressors occurring outside the work environment . Hassle scale is confounded by items that have negative wording. Uplifts scale is confounded by items that have positive wording. irritating, frustrating, distressing demands that to some degree characterize everyday transactions with the 117 item Hassles index developed by the authors. Examples include: Losing things Traffic jams Inclement weather Arguments

PAGE 70

70 Table 2 1. Continued Researcher(s) Sample Conceptual definition of stressor (s) Operationalizati on Findings Limitations Bhagat, McQuaid, Lindholm & Segovis (1985) 282 full time, white collar administrati ve, health care and clerical personnel stress as resulting from those life event changes that produce a state of challenge c oupled with disruptive pleasure for suggest that positive stress is a dynamic condition in which an individual is confronted with an opportunity for being (or having or doing) what he or she 83 item version of Psychiat ric Epidemiology Research Interview (PERI) Life Events Scale using (1979) response format. Negative job stress Negative personal life stress Total negative life stress Positive job stress was negatively correlated to job strains and job alienation and positively related to satisfaction with work and commitment. Negative job stress was positively correlated with job strains, feelings of alienation, and turnover intentions, and was negatively related to job satisfaction and commitment. Positive and negative stressor scales are confounded with stressors occurring outside the work environment. Low internal consistency reliabilities of the stressor scales may be attenuating empirical relationships with criteria. The use of reflective indi cator modeling may lead to measurement model misspecification and inaccurate parameter estimates. Did not control for the potentially biasing negative affect. other hand, is defined as resulting from life event changes that produce excessive and undesirable constraints and/or demands on the 83 item version of Psychiatric Epidemiology Research Interview (PERI) Life Events Scale using (1979) response format. Negati ve job stress Negative personal life stress Total negative life stress

PAGE 71

71 Table 2 1. Continued Researcher(s) Sample Conceptual definition of stressor (s) Operationalization Findings Limitations Scheck, Kinicki & Davy (1995) 135 employees of a high tech firm involved in R&D and manufacturi ng (1985) definition of positive stressors. 27 item version of PERI Life Events Scale using Sarason and (1979) response format. Love and marriage (5 items) Family (3 items) Social act ivities (4 items) Finances (3 items) Relationships (3 items) Compensation (2 items) Job (7 items) Positive stress events were positively related to job and life satisfaction and commitment. Negative stress events were negatively related to job and lif e satisfaction and commitment. Positive and negative stressor scales are confounded with stressors occurring outside the work environment. Seven job events that were described may not capture complete domain of positive and negative work stressors. The u se of reflective indicator modeling may lead to measurement model misspecification and inaccurate parameter estimates. Did not control for the potentially biasing negative affect. (1985) definition of negativ e stressors. 27 item version of PERI Life Events Scale using Sarason and (1979) response format. Love and marriage (5 items) Family (3 items) Social activities (4 items) Finances (3 items) Relationships (3 items) Compensation (2 items) Job (7 items)

PAGE 72

72 Table 2 1. Continued Researcher(s) Sample Conceptual definition of stressor (s) Operationalization Findings Limitations Scheck, Kinicki & Davy (1997) 218 employees of a high tech firm (1985) definition of positive st ressors. 13 item version of PERI Life Events Scale using Sarason and (1979) response format. Job (7 items) Work relationships (3 items) Compensation (2 items) Positive stress events were positively related to problem focused forms of coping and subjective well being; whereas negative stress events were positively related to emotion focused forms of coping and employee strain, and negatively related to subjective well being. Ratings of the positive or negative nature of various stressors may be i nfluenced by personality, affect, memory, etc. Seven job events that were described may not capture complete domain of positive and negative work stressors. The use of reflective indicator modeling may lead to measurement model misspecification and inaccu rate parameter estimates. Did not control for the potentially biasing negative affect. (1985) definition of negative stressors. 13 item version of PERI Life Events Scale using Sarason and (1979) res ponse format. Job (7 items) Work relationships (3 items) Compensation (2 items)

PAGE 73

73 Table 2 1. Continued Researcher(s) Sample Conceptual definition of stressor (s) Operationalization Findings Limitations Cavanaugh, Boswell, Roehling & Boudreau (200 0) 1886 high level managers related self reported stress as self reported work stress associated with challenging job 6 item challenge stressors scale developed by the authors. The number of projects or assignment s I have. The amount of time I spend at work. The volume of work that must be accomplished in the allotted time. Time pressure I experience. The amount of responsibility I have. The scope of responsibility my position entails. Challenge related stressors w ere positively related to job satisfaction, negatively related to job search, and unrelated to voluntary turnover. Hindrance related stressors were negatively related to job satisfaction, positively related to job search and voluntary turnover. Use of expl oratory factor analysis and reflective indicator modeling may have led to deficiency in the challenge and hindrance stressor measures. The use of reflective indicator modeling may lead to measurement model misspecification and inaccurate parameter estimat es. challenge related self reported stress, stress associated with job demands or work circumstances that involve excessive or undesirable constraints that interfere with or ability to achieve desired goals (deman ds that produce distress) is called hindranc e related self (p. 67). 5 item hindrance stressor scale developed by the authors. The degree to which politics rather than performance affects organizational decisions. The inability to clearl y understand what is expected of me on the job. The amount of red tape that I need to go through to get my job done. The lack of job security I have. The degree to which my career

PAGE 74

74 Table 2 1. Continued. Researcher(s) Sample Concept ual definition of stressor (s) Operationalization Findings Limitations Boswell, Olson Buchanan, & LePine (2004) 461 university employees holding clerical, computer support, maintenanc e, and administrati ve positions findings is consisten t with research suggesting that stress may be distinguished as to whether it is appraised as hindering or promoting mastery, personal growth, or future gains ( Folkman & Lazarus, 1985 ; Lazarus & Folkman, 1984 ). Scholars have the latter type of stress (henceforth, challenge stress ) with respect to important attitudes and b Used the six item challenge stressor scale developed by Cavanaugh et al. (2000). Challenge related job demands were negatively related to job search in both samples and positively related to job satisfaction in the European sample. Hindr ance stressors were positively related to job search and negatively related to job satisfaction in both samples. Use of confirmatory factor analysis and reflective indicator modeling may have led to deficiency in the challenge and hindrance stressor mea sures. The use of reflective indicator modeling may lead to measurement model misspecification and inaccurate parameter estimates. Did not control for the potentially biasing effects of negative affect. findings is consist ent with research suggesting that stress may be distinguished as to whether it is appraised as hindering or promoting mastery, personal growth, or future gains ( Folkman & Lazarus, 1985 ; Lazarus & Folkman, 1984 ). Scholars have differentiated the former type of stress (henceforth, hindrance stress Used the five item hindrance stres sor scale developed by Cavanaugh et al. (2000).

PAGE 75

75 Table 2 1. Continued Researcher(s) Sample Conceptual definition of stressor (s) Operationalization Findings Limitations LePine, Podsakoff & LePine (2005) 82 articles reporting primary data from 101 independent samples included stressful demands viewed by managers as obstacles to be overcome in order to learn and achieve. (p. 764 766). Stressors categorized as challenges: Job/role demands Pressure Time urgency Workload Measures of challenge stressors were positively related to employee strains, motivation, and performance. Measures of hindrance stressors were positively related to employee strains, but negatively related to motivation and performance. Use of primary studies may lead to deficient measures of challenge and hindrance stressors because previously these studies may not capture all dimensions of the focal variables. Challenge and hindrance stressors measures used in primary stu dies may have questionable construct validity. Did not control for the potentially biasing negative affect. included stressful demands viewed by managers as unnecessarily thwarting personal growth and goal attainment. 765 766). Stressors categorized as hindrances: Constraints Hassles Resource inadequacies Role ambiguity Role/interpersonal conflict Role dissensus Role interference Role clarity (reverse coded) Role overl oad Supervisor related stress Organizational politics

PAGE 76

76 Table 2 1. Continued Researcher(s) Sample Conceptual definition of stressor (s) Operationalization Findings Limitations Bingham, Boswell & Bourdreau (2005) 1377 high level managers in the Uni ted States and 1877 managers in Europe related job demands generally denote the aspects or features of a job that simulate development and include such requirements as working fast, having much work to do, and managing a heavy workload. Such job demands generally associate with positive 654). You were responsible for numerous products, technologies or services. You were responsible for multiple functions or groups. To succeed in the job, you had to dismantle the strategy your p redecessor had established. You were trying something the organization had never tried before . You had to deal with diverse clients, customers, or markets. The job included launching new organizational ventures (e.g., new product lines, acquisitions). Some of your key direct reports lacked the experience to do their job without your close supervision. Challenge related job demands were related to job search in both samples and positively related to job satisfaction in the European sample. Hindrance stressor s were positively related to job search and negatively related to job satisfaction in both samples. Use of confirmatory factor analysis and reflective indicator modeling may have led to deficiency in the challenge and hindrance stressor measures. The use of reflective indicator modeling may lead to measurement model misspecification and inaccurate parameter estimates Did not control for the potentially biasing effects of affect.

PAGE 77

77 Table 2 1. Continued. Researcher(s) Sample Concep tual definition of stressor (s) Operationalization Findings Limit ations related job demands are elements of the job that may be seen as overly taxing and exceeding the possible resources a manager may possess. Obstacles include difficult soc ial relationships (i.e., unsupportive bosses) and adverse business conditions. The obstacles often inhibit managerial development and are associated with negative psychological Your view of what the strategy is for your business differs significantly from top Your business or unit faces poor economic conditions. You boss is unaware of his or her shortcomings. There was no well though out direction or plan relating your part of the business to the overall strategy. Your boss is not highly regarded in the company. Your boss gives you useful advice and support ( reverse coding ).

PAGE 78

78 Table 2 1. Continued Researcher(s) Sample Conceptual definition of stressor (s) Operationalization Findings Limitations Podsakoff, LePin e , & LePine ( 2007 ) 157 articles reporting primary data from 183 independent samples factor was comprised of items reflecting high levels of workload, time pressure, job scope, and responsibility and was related stress because employees tended to view these demands as creating challenge and/or the opportunity for personnel development 4). Challenge stressors: Job or role demands Pressure to complete tasks Time urgency Quantitative and subjectiv e workload Challenge stressors were positively related to strains, job satisfaction, and organizational commitment, and negatively related to intentions to leave and turnover. Hindrance stressors were positively related to strains, intentions to leave tur nover, and withdrawal behavior, but were negatively related to job satisfaction and organizational commitment. Use of primary studies may lead to deficient measures of challenge and hindrance stressors because previously these studies may not capture all dimensions of the focal variables. Challenge and hindrance stressors measures used in primary studies may have questionable construct validity. Did not control for the potentially biasing negative affect. fa ctor was comprised of items measuring role ambiguity, role conflict, organizational politics, and concerns about job security, and was because employees tended to view these job demands as obstacles to personal growth and (p. 4). Hindrance stressors: Situational constraints Hassles Organizational politics Resource inadequacies Role ambiguity Role conflict Role overload

PAGE 79

79 Table 2 2. Summary of studies testing curvilinear stress criteria relations hips . Author(s) Sample Job demand(s) Dependent and moderator variables Findings Limitations Champoux (1980) Three samples: 1. 1198 members of an R&D organization 2. 66 members of a federal agency 3. 161 members of a federal agency Job scope Growth need s trength Internal work motivation Growth satisfaction Supervisory satisfaction Gen erally speaking, this study found that job scope ha d curvilinear effects on general satisfaction and growth satisfaction in two of the three organizations and curvilinear e ffects on internal work motivation in only . suggest that the curvilinear relationship between job scope and affective response has important implications for job redesign Examined a limited set of job demands. Examined a limited set of criterion variables. Between subjects design. Champoux (1981) Four samples: 1. 1152 members of an R&D organization 2. 1193 members of an R&D organization 3. 66 members of a federal agency 4. 160 members of a federal agency Job scope Growth need strength General satisfaction Internal work motivation Growth satisfaction Gen erally speaking, this study found that: (a) job scope had curvilinear relationships with growth satisfaction in three of the four organizat ions examined, (b) job scope had curvilinear relationships with internal work motivation in half of the samples examined, and (c) supervisory satisfaction moderated the curvilinear relationship between job scope and growth satisfaction in one of the organizations that was examined. This study uses data reported in Champoux (1981). Examined a limited set of job demands. Examined a limited set of criterion variables. Between subjects design. Friend (1982) 39 management personnel in a t wo week management training course in economics Subjective workload Time urgency Performance (Exam scores) This study found that subjective workload and time urgency had negative, linear effects on exam performance, but did not have the hypothesized cu rvilinear effects on the relations provides no evidence for the inverted U hypothesis relating stress, activation, and/or motivation to Examined a limited set of job deman ds. Examined a limited set of criterion variables. Low sample size may contribute to weak power to detect curvilinear relationships. Between subjects design.

PAGE 80

80 Table 2 2. Continued Author(s) Sample Job demand(s) Dependent and moderator variables Fi ndings Limitations Edwards & Harrison (1993) Data taken from French et al. (1982): 2010 employees from 23 occupations Job complexity Ro le ambiguity Responsibility for persons Quantitative work load Job dissatisfaction Work load dissatisfaction Bor edom Depression Anxiety Irritation Somatic complaints Generally speaking, job complexity, responsibility for persons, and quantitative work load were found to exhibit curvilinear relationships with job dissatisfaction, work load dissatisfaction, boredo m, and other strains. Role ambiguity was not found to exhibit any significant curvilinear relationships with the measures of strains. Between subjects design. All criteria are self reported and relationships may be influenced by method bias. Jamal (198 4) 440 nurses in two hospitals in Canada Role ambiguity Role conflict Role overload Resource i nadequacy Overall performance Job performance Job motivation Patient care Absenteeism Tardiness Anticipated t urnover Only 2 of the 12 test s gave support to non findings of a negative linear relationship between stress and performance may be a surprise to many as it tends to be contrary to the Yerkes Dotson Law (Yerkes & Dotson, 1908) and activation the ory (Scott, Role overload, role conflict, and resource inadequacy had predominately negative, linear relationships with employee effectiveness, absenteeism, and anticipated turnover. Examined a limited set of (hindrance) stre ssors. Between subjects design. Limited range of occupations. Jamal (1985) 227 middle managers and 283 blue collar workers in a Canadian organization Role ambiguity Role conflict Role overload Resource i nadequacy Composite measure of stressors Per formance Quality Quantity Effort exerted Although all four stressors had negative, linear relationships with the three dimensions of performance, none of the 20 relationships between the five stressors and the quantity and quality of performance across the two samples were curvilinear, and that only role overload and resource inadequacy had a curvilinear relationship with effort expended. Role overload had a curvilinear relationship with effort expended in both the managerial and blue collar samp le, and resource inadequacies had a curvilinear relationship with this criterion only in the blue collar sample. Examined a limited set of (hindrance) stressors. Between subjects design. Limited range of occupations.

PAGE 81

81 Table 2 2. Continued Author(s) Samp le Job demand(s) Dependent and moderator variables Findings Limitations Xie & Johns (1995) 418 Full time employees in 143 different jobs Job scope Anxiety Emotional e xhaustion Perceived D A f it Job scope had a curvilinear relationship with employe e exhaustion, but a negative, linear relationship with employee anxiety. D A fit moderated the relationship between job scope and employee anxiety. Between subjects design. All criteria are self reported and relationships may be influenced by method bia s. Westman & Eden (1996) 306 Israeli Office Cadets Work Demands Ability Performance Navigation Leadership Combat d octrine Tactics Discipline Found suppor t for the negative linear relationship between stress and performanc e, but no support for the inverted U hypothesis for this relationship . Between subjects design. One item measure of stress may confound challenge and hindrance stressors. Limited range of occupations; the demands of cadets may not be representative of g eneral employees. De Jonge & Schaufeli (1998) 1437 Dutch health care workers (nurses and nurses aides) Job demands Job satisfaction Job related anxiety Emotional e xhaustion Job demands were found to have a curvilinear U shaped relationship with anxiet y. However, job demands were also found to have a negative, linear relationship with satisfaction, and a positive, linear relationship with exhaustion. Between subjects design. A limited range of occupations were studied. Measure of job demands may confou nd challenge stressors with measures of other constructs.

PAGE 82

82 Table 2 2. Continued Author(s) Sample Job demand(s) Dependent and moderator variables Findings Limitations Janssen (2001) 134 Dutch employees and managers in the food sector Job demands F airness p erceptions Standard job performance Leader Rated Innovative job performance Self rated innovative job performance Work satisfaction Supervisory satisfaction Job demands had a curvilinear relationship with leader rated innovative job performan ce, work satisfaction, and supervisory satisfaction. In addition, a significant interaction between the squared job demands term and fairness perceptions was reported for all forms of job performance, work satisfaction, and supervisor satisfaction. Betwee n subjects design. Measure of job demands may confound challenge stressors with measures of other constructs.

PAGE 83

83 CHAPTER 3 STUDY 1 The purpose of Study 1 was to develop and validate measures of challenge and hindrance stressors that addressed the limitat ions identified in previous studies. Specifically, I created multi dimensional scales of challenge and hindrance stressors and modeled these measures as second order constructs having formative indicators. I used a six step construct development and valida tion process and reported analyses from three samples designed to examine the content, convergent, discriminant, and nomological validity of the challenge and hindrance stressor measures. The construct development and validation procedures used in this stu dy were developed from both classical and contemporary methods discussed by Campbell and Fiske (1959), Cronbach and Meehl (1955), Churchill (1979), Schwab (1980), Fornell and Larcker (1981), Bollen (1989), Bollen and Lennox (1991), Diamatopolous and Winklh ofer (2000), MacKenzie et al. (2005), and others. Generally speaking, the six step process I followed is outlined in Figure 3 1. In the following sections, I explain each step in detail and the actions that I completed in each of these steps. In the intere st of parsimony, I will present the steps for both constructs simultaneously. Step 1: Conceptual Definition of the Constructs In the first step of any construct validation process, it is necessary to provide a clear and concise definition of the conceptual domain of the focal construct. Because the conceptual definition of a construct is the foundation for all later steps of the process, it should be considered the most important stage in any construct validation. Although several researchers describing sca le development and validation procedures (Churchill, 1979; DeVellis, 1991; Schwab, 1980; Spector, 1992) emphasize the importance of this step, they typically do not

PAGE 84

84 eptual domain. However, virtually all researchers agree that the definition of a construct should be consistent with the previous literature. I took three steps to ensure that the challenge and hindrance stressor measures met this requirement. First, I co nducted an extensive review of the work stress literature. Although the size of this literature (in excess of 4,6 00 articles) precluded me from conducting an exhaustive search, I did his literature. Specifically, I identified 11 dimensions of work demands that have been consistently referred to as important in the literature. These include the following: (1) role ambiguity (Abramis, 1994; Fisher & Gitelson, 1983; Jackson & Schuler, 198 5; Kahn et al. , 1964; Rizzo, House, & Lirtzman, 1970), (2) role conflict (Abramis, 1994; Fisher & Gitelson, 1983; Jackson & Schuler, 1985; Kahn et al. , 1964; Rizzo et al., 1970), (3) workload, or role demands (Edwards & Harrison, 1 993; French et al., 1982; Jex, 1998; Spector & Jex, 1998), (4) resource inadequacies time pressure (Andrew & Ferris, 1972; Jex, 1998; Spector & Jex, 1998), (6) organizational politics (F erris & Kacmar, 1992; Harris & Kacmar, 2005; Kacmar & Ferris, 1991), (7) administrative hassles or red tape (Brough, 2005; Hart, 1999; Hart, Wearing, & Headey, 1994; Zohar, 1999), (8) interpersonal conflict (Jehn, 1994, 1995, 1997; Jex, 1998; Spector & Je x, 1998), (9) job complexity (Jex, 1998; Hackman & Oldham, 1980; Xie & Johns, 1995 ) , (10) job responsibility (Edwards & Harrison, 1993; French et al., 1982; Hackman & Oldham, 1980), and (11) job insecurity (Greenhalgh & Rosenblatt, 1984; Probst, 2005; Sver ke, Hellgren, & Naswall, 2002).

PAGE 85

85 Second, I reviewed all the studies that have specifically distinguished between challenge and hindrance stressors to identi fy which work stressors were grouped into each of these two categories . It is important to note that these previous studies have used a variety of different techniques to identify the stressors that should be identified as challenges versus those that should be identified as hindrances. For example, Cavanaugh et al. (2000) had two groups of subject matte r experts categorize job demands into those that represented challenge stressors from those that represented hindrance stressors. In order to develop measur es of challenge and hindrance stressors, LePine, LePine, and Jackson (2004) gathered critical incide nts of stressful events that were perceived by participants in their study to either promote or hinder growth and accomplishment. Items were created from these incidents and then sorted by the authors into categories representing challenge and hindrance r elated stressors. A second set of participants then categorized the items into challenges and hindrances, which were reported to be very similar to the categorizations conducted by the authors. In addition, these authors conducted a confirmatory factor ana lysis (CFA) and reported that the measures of the challenge and hindrance stressor constructs were distinct from each another and valid. In another study, LePine et al. (2005) had 43 part time MBA students with work experience rate various work stressors on the extent to which they were challenging and/or hindering to personal goal accomplishment, growth, and learning. These authors reported that the job demands that were identified as hindrance stressors were rated as significantly more hindering than the job demands identified as challenge stressors; and that the job demands that were identified as challenge stressors were rated as significantly more challenging than were the job demands identified as hindrance stressors. Finally, preliminary support for the nomological validity of these stressor measures was provided in two meta analyses (LePine et al., 2005;

PAGE 86

86 Podsakoff et al. , 2007 ), in that challenge stressors were generally found to have functional relationships with employee motivation, job satisfactio n, organizational commitment, intentions to leave, job performance, turnover, and withdrawal behaviors; whereas hindrance stressors generally had dysfunctional relationships with these criteria. As a final step in my attempt to ensure that conceptualizati ons of the challenge and hindrance stressor constructs were consistent with the previous literature, I categorized each of erature using the challenge/ hindrance stressor framework. T he results of this categorization process are reported in Table 3 1. In this table, I have indicated the dimensions of the challenge and hindrance stressor constructs that have been used in each of the previous st udies testing the challenge/ hindrance stres sor framework. As indicated in T able 3 1 , there is a great deal of consensus with respect to the categorization of some of the dimensions of the work stressors, but not for others. For example, workload and work pace have been included as dimensions of ch allenge stressors in virtually all the previous studies in this domain. Similarly, with only a few exceptions, role ambiguity, organizational politics, and administrative hassles have been included in all of the previous studies as dimensions of hindrance stressors. In contrast, job complexity has been included as a dimension of the challenge stressor construct only in recent studies by Boswell et al. (2004) , Zapata, Hurst, and Colquitt (u nder review), and Podsakoff, Rich, Saul, and LePine ( u nder review); a nd only one study (Zapata et al.) identified interpersonal conflict as a dimension of the hindrance stressor construct. However, in order to make certain that the domain of the challenge and hindrance stressors were adequately represented and that the meas ures of these constructs were comprehensive, I choose to include all four dimensions of the challenge stressor construct

PAGE 87

87 (i.e., workload , work pace, job complexity, and job responsibility) and all seven dimensions of the hindrance stressor construct (i.e., role ambiguity, role conflict, resource inadequacies, organizational politics, administrative hassles, interperson al conflict, and job insecurity ) identified in Table 3 1 . Although it is obvious that some of these dimensions have been acknowledged as chal lenge and hindrance stressors more o ften than others, the inclusion of the dimensions in this table is driven by theoretical, as well as empirical, considerations. In addition to being consistent with the previous literature, there are three other charact eristics of the conceptual domain that should be considered when developing a good construct definition. First, Churchill (1979) notes that, to adequately specify the conceptual at is included in specify the dimensionality of the construct (Spector, 1992). Identifying the important facets or aspects of a construct is central to understandin g its nature and its relationships with other constructs. Finally, researchers should specify the level of the construct that is being , group , organizational , or cultural level of analysis. For example, job satisfaction is considered an individual level construct, whereas team adaptability is a group level construct. Using these guidelines, I developed conceptualizations for challenge and hindrance stressors an d each of their respective dimensions. A summary of these conceptualizations is presented in Table 3 2. As indicated in Table 3 2, the challenge stressor construct is multi dimensional, comprised of four unidimensional facets (i.e., workload, work pace, j ob complexity, and job responsibilit y), and is defined as demands in the workplace that tend to be appraised as promoting the accomplishment of job tasks and the personal development of the individual. For the purposes of

PAGE 88

88 tly refer to characteristics of or events occurring in the specifically refer to individual level perceptions of characteristics or events in t he work environment. Th is definition does not include promotions or raises, because these are not work Rather, these events describe the acquisition of title or status in the organizatio n and additional compensation that one receives for performing his or her job. In contrast, the hindrance stressor construct is a multi dimensional construct comprised of seven unidimensional facets (i.e., role ambiguity, role conflict, organizational pol itics, resource inadequacies, administrative hassles, interpersonal conflict, and job insecurity ) and is defined as demands in the workplace that tend to be appraised as barriers or obstacles to the accomplishment of job tasks and the personal development of the indivi dual. Given this definition, hindrance stressors do not include layoffs or actual job loss, because these events signal a removal of the employee from the work environment, and thus cannot influence subsequent job tasks. In the following secti conceptual framework to elaborate on the distinctions between challenge and hindrance stressors. The A ppraisal of Challenging Work S tressors Although challenge and hindrance appraisals do not lie on opposite end s of a continuum, they do provide for very different perspectives on stressors in the workplace (Lazarus & Folkman, 1984). Thus, even though appraisals of challenges and hindrances may be correlated, there are specific work related demands that tend to be appraised as challenges by employees and other work related demands that tend to be appraised as hindrances by employees. Lazaurus and hindrance stressors and subsequ ent reactions to these demands.

PAGE 89

89 Lazarus and Folkman (1984) indicate that commitments, or expressions of what is important or meaningful to individuals, are particularly importan t to the appraisal process. These authors lves a strongly held commitment will be evaluated as meaningful to the extent that the outcome harms or threatens the commitment or facilitates its ls, motives, and values (Edwards & Cooper, 1990) and are based on his or her 299) . In essence, position in the organization (Biddle, 1979). Although the process of role development may be complex, once employees have taken on a specific r ole in the organization they generally recognize that they will be rewarded for fulfilling commitments related to this role, and that not meeting these commitments makes them subject to punishment, the withholding of rewards, or removal from the organizati on (Graen & Scandura, 1987). Therefore, employees who experience job demands that facilitate their ability to perform effectively within their role and develop additional competencies would be expected to receive more extrinsic rewards from the organizatio n than those employees who do not experience these job demands. In contrast, employees who encounter job demands that act as obstacles or barriers to the accomplishment of their role related responsibilities or the enhancement of skills necessary to perfor m their role would be expected to receive fewer extrinsic rewards from the organization than employees who do not experience these job demands.

PAGE 90

90 In addition to the extrinsic rewards associated with the accomplishment of role responsibilities, employees also receive intrinsic rewards from the roles they perform. Klapp (1969) has noted that the role an individual performs at work is associated with his or her social self worth, and self esteem; (b) structure and purpose to daily li fe; (c) opportunities for skill development and creative expression; and (d) a sense of security and a measure of our identity and the meaningfulness of his or her life. Thus, events that either facilitate or hinder the identity. More specifically, employees who encounter obstacles or barriers to fulfilling their role r esponsibilities should also feel that their work is less meaningful, perceive fewer opportunities for development and creativity, and diminished levels of self esteem, self worth, and security. In contrast, it is expected that employees who encounter job d emands that serve to promote the fulfillment of their role responsibilities should experience greater meaningfulness in their work, have more opportunities for personal development and creativity, and stronger perceptions of self esteem, self worth, and se curity. Under these circumstances, employees should feel more committed to meeting their role responsibilities because they are more consistent with their own goals, motivations, and values. In other words, job stressors will tend to be appraised as chal lenging if they provide an employee with the opportunity to obtain: (a) extrinsic rewards related to the accomplishment of role concept and

PAGE 91

91 development opportunities. Consistent with th is description, I have identified four demands in the work environment that will tend to be appraised as challenging to employees (s ee Table 3 2). workload , or role demands. Building on Spector (1998) notion of quantitative workload, this dimension refer s to the amount or quantity of work challenging because they e ngage employee i n the performance of duties necessary to thei r role and help clarify their identity within the organization. The completion of prescribed tasks also provides employees with the opportunity for intrinsic and extrinsic gain in the workplace. Accomplishing job tasks is a basic requirement for the evalua tion, reward, and promotion of employees, and employees who adequately perform tasks related to their roles will be more likely to receive greater extrinsic rewards than those that do not meet role expectations related to their tasks. In addition, meeting should lead to intrinsically rewarding feelings of accomplishment and pride (Lazarus & Folkman, 1984). The second job demand that should tend to be appraised as challenging by employees is work pace , which is defined as the relative rate or speed with which work must be completed to ployees as challenging because it motivates them to direct effort toward the complet ion of t heir role responsibilities within a specific period of time. Locke and Latham (1990) have noted that setting specific , challenging goals motivates employees to perf orm at higher levels than do goals in which employees are told to do their best. In addition, these authors also note that challenging goals motivate people to persist in the face of obstacles to complete their role related tasks. One might

PAGE 92

92 also expect tha for completing their role requirements. Assuming that these motivational processes also apply to time specific goals, employees will tend to view work pace as a demand that pr omotes the accomplishment of role related tasks and offers the opportunity for growth. Job complexity is the third demand that should tend to be appraised as providing challenge in the workplace. Consistent with job characteristics theory (Hackman & Oldha m, 1980), job complexity refers to the breadth or variety of job related activities performed by an organizational member, such that more c omplex jobs require a broader variety of tasks than do less complex jobs. For example, a retail salesperson who is re quired to meet and serve customers, stock inventory, and keep track of customer purchase patterns would be perceived as having a more complex job than one who is only req uired to perform only o ne of these tasks. Research on job characteristics generally in dicates that employees in complex, challenging jobs perform at a higher level (Fried & Ferris, 1987; Hackman & Oldham, 1980), learn more about their jobs and the results of their work (Fried & Ferris, 1987; Hackman & Oldham, 1980), and develop more innovat ive ideas (Amabile, Conti, Coon, Lazenby, & Herron, 1996; Oldham & Cummings, 1996) than those employees who work in less complex or less challenging jobs. In addition, jobs having complex components are generally accompanied by rewarding feelings of satisf action (Fried & Ferris, 1987). Finally, jobs that require a variety of skills tend to be compensated at higher levels than jobs that require a less complex set of skills (Brannick & Levine, 2002; McCormick, 1979). Therefore, employees should perceive the c omplexity of their job to be a challenging demand in the workplace. The fourth type of demand that should be perceived as challenging in the workplace is job responsibility , which is defined as the perceived accountability that an individual has for his/ her

PAGE 93

93 own work and the work of others. This demand is expected to be associated with challenge appraisals because it offers intrinsic and extrinsic rewards related to growth and gain in the organization. For example, employees are likely to perceive high res ponsibility jobs as providing greater opportunities for ext rinsic gain in the organization because these jobs are compensated at a higher level then are jobs with lower levels of accountability (McCormick, 1979). In addition, employees who perceive higher levels of accountability should be more committed to meeting their role related responsibilities. Finally, because high levels of responsibility signal the importance of the individual to other organizational members, this demand should also have a powerfu l, clarifying effect on the identity of the individual. Therefore, perceptions of job responsibility should be appraised as challenging demands by employees. The Appraisal of Hindering Work S tressors Con process, I argue that work stressors will tend to be appraised as hindering (threatening) if they prevent or deter an employee from: (a) accomplishing job related tasks, (b) gaining intrinsic and/or extrinsic rewards, and/or (c) developing him or herself. Following this general description, I have identified seven work demands that will tend to be appraised as hindering to employees. The first two demands that should be perceived as hindering are role ambiguity and role conflict . Consistent with Kahn et al. appropriate to take to fulfill role requirements ; whereas role conflict is defined as incompatibility between the expectations of parties or aspects of a set of role related tasks. Because it is very difficult for employees to complete role related tasks when information regarding these duties is not provided or is unclear, they are likely to perceive that role ambiguity threatens their role in the organization, t he completion of their tasks, and their potential for growth. Similarly, employees who receive conflicting requests or assignments from management tend to view these

PAGE 94

94 demands as a barrier to the achievement of personal and organizational goals. In addition, these take, threaten the identity of the individual within the organization, and result in negative affective responses (Jackson & Schuler, 1985). Therefore, organization unclear or create conflict in this role will generally be appraised as hindrances because these demands are obstacles to the performance of role related duties and the achievement of personal growth. Percept ions of organizational politics are the third type of job demand that should be appraised by employees as hindering to personal gain and growth. Consistent with the work of Ferris and his colleagues (Ferris & Judge, 1991 ; Ferris & Kacmar, 1992; Kacmar & Fe rris, 1991), organizational politics are defined as unsanctioned influence attempts designed to promote self interest at the expense of organizational goals. Work environments with high levels of organizational politics are characterized by perceptions tha manipulative and self serving (Kacmar & Ferris, 1991). Employees who perceive that indi viduals in the organization build themselves up by tearing other people down, are willing to agree with superiors just to get ahead, and are given raises and promotions based on favoritism rather than performance or policy (Kacmar & Ferris, 1991), are likely to appraise this source of stress as threatening to their individual role, their potential for growth, and the goals of the organization. Moreo ver, employees who attempt to respond to these perceptions have fewer resources available to complete their assigned tasks and assist coworkers. Indeed, perceptions of organizational politics have been shown to be negatively related to task performance (Wi tt, 1998) and organizational citizenship behaviors (Randall, Cropanzano, Bormann, & Birjulin, 1999). Finally, because demands related to organizational politics often make the behaviors

PAGE 95

95 related to reward allocation decisions unclear, the appropriate respon se to these demands is ambiguous and will likely result in employee withdrawal (Ferris & Kacmar, 1992 ). Therefore, perceptions of organizational politics will tend to be appraised by employees as hindering to their task accomplishment and individual develo pment. The fourth type of job demand that should be appraised by employees as hindering to personal gain and growth is administrative hassles . Admin istrative hassles are defined as the perception that the organization has excessive or unne cessary requireme nts , regulations, or rules that employees must deal with during the completion of their work. Examples of these types of overly restrictive rules and regulat ions. These stressors are viewed as hindering, because they ions of strains, and reduce their satisfacti on, motivation and performance o n the job. For example, employees who expend resources by constantly completing excessive or unnecessary paperwork are likely to feel hampered from performing their job and would b e expected to express greater levels of frustration and dissatisfaction. Similarly, employees who feel that their job performance is constrained by what they perceive to be unnecessary rules and regulations are likely to become upset with these constraints . This is generally consistent with the research reported by Brough (2005), who found that organizational hassles (which included administrative hassles) were negatively correlated with intrinsic job satisfaction, extrinsic job satisfaction, and feelings of well being in three samples comprised of a total of 723 respondents from the New Zealand police, fire, and ambulance services. It is also consistent with the results of a three wave longitudinal study conducted by

PAGE 96

96 Hart (1999), who reported that that a m easure of work hassles that included administrative hassles was negatively related to job satisfaction and life satisfaction across all three time periods in a sample comprised of 479 police officers. Finally, Zohar (1999) has reported that work hassles (i ncluding administrative hassles) had a positive relationship with the negative mood and fatigue of a sample of military parachute jump trainers, even after controlling for sleep loss and task complexity. Taken together, these findings suggest that employee s view administrative hassles as barriers or obstacles to task accomplishment and personal growth in organizational settings. The fifth job demand that should generally be appraised as hindering in the work context is resource inadequacies . Consistent with conceptualizations provided by Kahn et al. (1964) and availability of tools, equipment, materials, and/or supplies is insufficient to adequately complete role related tasks. Although resources could refer to any number of objects and/or processes (Wernerfelt, 1984), for the purposes of this conceptualization , resource inadequacies specifically refer to the insufficient availability of physical resources in the work env ironment n ecessary to complete job tasks. authors also reported that em ployees describing critical incidents identified resource constraints these findings suggest that situational constraints act as an obstacle to employee perceptions that increased effort on their part will lead to increased levels of performance and valued rewards. In

PAGE 97

97 addition, because economic resources and rewards have bee n linked to employee creativity (Amabile et al., 1996; Shalley, Zhou, & Oldham, 2004) it is likely that employees who have inadequate resources or expectations that constraints inhibit the relationships between effort and performance will not be as likely to develop novel, potentially useful ideas for the organization. Therefore, job demands reflecting resource inadequacies should tend to be viewed as obstacles to employee task performance and personal growth in the organization. The sixth demand that shou ld be appraised as hindering in the workplace is interpersonal conflict . Following the conceptualization provided by De Dreu and Van Vianen (2001), interpersonal conflict is defined as disagreements resulting from personal differences in style, preferences , attitudes, and personality. Jehn (1994, 1995, 1997) has hypothesized that although conflict related to procedures and policies , the distribution of resources, and the judgment and interpretation of tasks (i.e., task conflict) may result in better decisio n making and higher group performance, interpersonal (relationship) conflict has unequivocal deleterious effects on group functioning and performance (De Dreu & Weingart, 2003). For example, arguing with coworkers about differences in personal opinions, po litical preferences, values, and personalities is likely to occupy employee cognitions and effort that could be used to complete role requirements and develop personal skills. Consistent with the appraisal of demands as threats (Lazarus & Folkman, 1984), i nterpersonal conflict often leads to negative emotions, job dissatisfaction, loss of self worth and identity, and decreased task performance. Therefore, perceptions of interpersonal conflict should tend to be viewed by employees as obstacles to achievement and growth in the work environment. Finally, job demands related to perceptions of job insecurity are the seventh stressor that should be hindering to employee growth and gain in the workplace. Consistent with Greenhalgh

PAGE 98

98 lization, job insecurity is defined as the perceived likelihood of involuntary job loss. According to a recent study (Society of Human Resource Management, 2001), 43% of organizations in the U.S. reported layoffs in 2000 and 2001, averaging between 10 to 1 likelihood of involuntary job loss and employees perceive that they have little control over this source of stress, it is not surprising that recent estimates suggest tha t over one third of the American workforce is concerned about job insecurity (Belton, 1999). It is likely that this sense of powerlessness is perceived as a hindrance to growth and gain in the organization and leads to withdrawal or avoidance forms of copi ng. For example, employees who hear rumors about layoffs and perceive their jobs to be vulnerable are likely to feel strain, shirk job duties, and expend time and effort searching for another job. A meta analysis of 86 independent studies conducted by Sver ke et al. (2002) supports this expectation in that job insecurity generally had negative effects on employee attitudes, mental and physical health, and job performance. Finally, by making membership in the organization less stable, job insecurity directly threatens the roles and identities of employees in the organization. Therefore, work demands related to employee perceptions of job insecurity should be perceived as hindrances to the accomplishment of role duties and the personal growth of employees. Nat ure of the Relationships between Challenge and Hindrance Stressors and Their M easures The right hand column of Table 3 2 provides a summary of the relationships between the first and second order stressor constructs and their measures. As indicated in thi s table, challenge and hindrance stressor s are conceptualized as second order constructs with formative indicators represented by their r espective first order dimensions. A ll of the first order dimensions

PAGE 99

99 of the challenge and hindrance stressor constructs are conceptualized as having reflective indicators. The second order challenge stressor construct was conceptualized as having first order formative indicators for several reasons. The first reason is that the four dimensions that make up the challenge st ressor construct (role demands, work pace, job complexity and job responsibility) are expected to cause changes in the challenge stressor construct and not vice versa. For lter his or her perception of challenge stressors, regardless of whether his or her perceptions of work pace, job complexity, or responsibility remain the same. The second reason is that these four dimensions are conceptually distinguishable from each othe r, are not necessarily interchangeable, and are not necessarily highly correlated with each other. For example, some jobs (e.g., assembly line work) may have a high workload and/or work pace, but are not overly complex in nature. Likewise, an employee may be assigned tighter deadlines for tasks (i.e., increased work pace) or tasks that require a broader set of skills (i.e., job complexity), even though his or her job responsibility may not change at all. Finally, it is worth noting that some of the factors expected to influence work pace (the speed of machine processe s or the technology being used) that some of the antecedents (and consequences) of these forms of challenge stressors are likely to be different from each other. Thus, I believe that it is more appropriate to model the challenge stressor construct as having formative, rather than reflective, indicators. As indicated in Table 3 2, the dimensions of the hindrance stressor construct were also specified as formative indicators of a higher order construct, because they were viewed as defining characteristics of the construct rather than as manifestations of it. In other words, role

PAGE 100

100 ambiguity, role conflict, job insecu rity, organizational politics, administrative hassles, resource inadequacies, and interpersonal conflict form they experience on the job, and are not reflections of an underlying latent variable. Indeed, c hanges in the hindrance stressor dimensions are expected to cause changes in the higher order that resources are inadequate to complete his or her tasks will le perception of the level of hindrance stressors he or she experiences on the job, regardless of whether his or her perceptions of role ambiguity, role conflict, organizational politics, administrative hassles, interpersonal conflict, and job insecurity stay the same. Finally, a change in the perceived level of one of the dimensions is not necessarily expected to be associated with a commensurate change in the experienced level of another dimension of the hindrance stressor co nstruct. For example, an employee who perceives that his or her role in the organization is ambiguous is not necessarily expected to also perceive high levels of interpersonal conflict with c oworkers or that the managers engage in political decision making . This is consistent with previous empirical research which has found that dimensions of hindrance related job demands do not typically covary at high levels (Cavanaugh et al., 2000; LePine et al., 2005; Podsakoff et al., 2007 ). In summary, the dimensions of the hindrance stressor construct are viewed as defining characteristics of the construct, are expected to cause changes in the construct, and are not necessarily expected to covary at a high level, and therefore, are more appropriately conceptualized as formative indicators of a higher order construct. Step 2: Generate Items to Measure the Constructs In the second step of the construct development and validation process, I developed indicators of the challenge and hindrance stressor constructs. It is im portant to note that when developing items to measure a construct, researchers must take into consideration the purpose of

PAGE 101

101 these items in assessing the focal variable. Reflective indicators of a construct, developed under the prescriptions of classical tes t theory, are expected to represent a sample of the total population of possible items that could be used to measure the construct (Bollen & Lennox, 1991), and are expected to be highly correlated and interchangeable with each another (Bollen & Lennox, 199 1). Because the dimensions of the challenge and hindrance stressor constructs fall under this conceptualization, each will be measured with reflective indicators. In contrast, formative indicators are not based on classical test theory, are not necessar ily expected to be interchangeable, and may exhibit positive, negative, or non significant correlations wit h each other (Bollen, 1984 ). Therefore, valid measures of a construct having formative indicators must represent a census of the measures tapping int o the content domain of the dimensions of the focal construct. If formative measures do not represent a census of items, then the focal construct will be deficient in its content. Because the challenge and hindrance stressor constructs fall under this conc eptualization, special care was taken during the conceptualization stage to identify a census of dimensions that tapped the complete conceptual domains of each construct and ensure they were not deficient. I used multiple methods to assist in the develop ment of the measures of the challenge and hindrance stressor dimensions. First, I conducted a comprehensive review of the theoretical and empirical literature on stressors. Following this, I conducted a review of stressor scales that have been used in orga nizational behavior research. A search of this literature identified over 50 scales comprised of over 500 items measuring employee work stressors. Finally, I employed the recommendations of subject matter experts (SMEs) in the field of work stress to devel op measures of the dimensions of the challenge and hindrance stressors constructs.

PAGE 102

102 Generally speaking, when generating items measuring the challenge and hindrance stressor constructs, I followed recommendations (Peterson, 2000; Spector, 1992; Torangeau, Ri ps , & Rasinski, 2000) that items be worded as simply and precisely as possible, not be double barreled (i.e., capture two distinct concepts), and not have content eliciting socially desirable responses . In addition, in response to concerns regarding the wo rding of items used in previous work stressor measures (cf. Cavanaugh et al., 2000), I took two additional precautions. First, I attempted to frame items in a way that would allow raters to respond as objectively as possible. As a result, I avoided wording items in a way that suggested a demand was either negative or positive in nature. For example, I was careful to exclude item content that required employees to whether they could cope with the demand. Second, I selected a response format that did not require employees to evaluate whether the demand had a positive or negative impact on them. Instead, respondents were asked to assess each item using a 5 point Likert type scale that ranged Step 3: Assess the Content Validity of the Measures In the third step of the validation process, I assessed the content validity of the challenge and hindrance stressor measures. Acc concerns the degree to which an instrument assesses all relevant aspects of the conceptual or behavioral domain that the instrument is intended to measure, or how thoroughly it samples the relevant target domain (p. 104). For the purposes of this study, I used a modified version of Hinkin and the challenge and hindrance stressor items were representative of the distinct dimensions of the content domai n of the respective constructs.

PAGE 103

103 To this end, I administered a content validity assessment questionnaire to two samples of undergraduate students. The content validity questionnaire was comprised of a matrix of the challenge and hindrance stressor items (l ocated in the rows) and the definitions of the distinct dimensions of the challenge and hindrance stressors constructs (located in the columns). Participants were instructed to follow a series of directions. First, they were instructed to read the definiti ons provided in the columns. The challenge stressor content validity survey contained four definition columns and the hindrance stressor content validity questionnaire contained seven definition columns. Second, the participants were instructed to read the first item carefully. Third, each participant was asked to rate the extent to which this item was captured by each definition listed in the columns. The rating scale provided was a five point scale with anchors not at all captured by the completely captured by definition (column). In addition, participants were provided one additional blank column at the end of each row lab captured in a given item, but were not captured by any of the definitions. Finally, participants were asked to continue the rating process with each subsequent item until they had c ompleted the entire questionnaire. The first challenge stressor and hindrance stressor content validity questionnaires were comprised of 20 items and 35 items, respectively. Items were randomly distributed within each of these questionnaires to reduce the potential effects of priming. Participants in the first stage of the content validity procedure were 44 undergraduate students from a large public university. Students were enrolled in a management course and participated in the project for extra credit , and the opportunity to win one of three $25 cash prizes.

PAGE 104

104 Table 3 3 provides a summary of the demographic characteristics of this sample. As indicated in this table, the majority of participants in the first stage of the content validity assessment were fem ale (61%), spoke English as their first language (87%) , and on average, were approximately 20.5 years (SD = 0.76) old. In addition, participants in this study had an average of over three years of previous work experience (SD = 2.16). Tables 3 4 and 3 5 report the means and standard deviations of the content validity ratings provided by this sample. Table 3 4 reports the challenge stressor items in 20 rows and the construct dimension labels in four columns. Table 3 5 reports the hindrance stressor items i n 35 rows and the construct dimension labels in seven columns. In order to examine whether each item assessed the dimension that is was intended to capture, I conducted t tests of the ratings for each item on its intended factor versus ratings on all other factors. The results of these analyses are reported in Tables 3 4 and 3 5. The results reported in these tables indicate that the average ratings for the items on their intended factors were high (challenge stressor item M = 4.38; hindrance stressor item M = 4.09). In addition, the t tests reported in these tables indicated that the ratings on the intended factor versus the ratings on other non intended factors were significant ly different (p < .01) in virtually all cases. The single exception to this was hindrance average, participants rated this item almost as high on the role ambiguity dimension (M = 3.10), as on its intended role conflict dimension (M = 3.62), a nd indeed, these ratings were not found to be significantly different (p > .10) from each other. However, generally, this initial evidence suggested that the challenge and hindrance stressors items did tap their intended factors, and were not contaminated by other related stressor factors.

PAGE 105

105 Despite this generally supportive evidence, a closer examination suggested that the wording of some items could be refined, while others items might be removed. Specifically, I asked SMEs to examine the items in the cont ext of the initial content validity ratings and make suggestions for item purification. This resulted in minor wording changes to three challenge and four hindrance stressors items, and the removal of four challenge and six hindrance stressors items. Afte r making these changes, I administered the revised content validity questionnaire s to a second sample of undergraduates. The procedures for this second content validation were exactly the same as those for the previous sample except that the challenge and hindrance stressor questionnaires were broken into three separate surveys. The reason for separating the items into three smaller surveys was to encourage participants to take their time in responding to the questionnaire and to help reduce participant fa tigue during the rating process. Three separate participant sub samples (A, B, and C) were administered one challenge and one hindrance stressor content validity questionnaire. The instructions were exactly the same as those provided in the previous conten t validity questionnaire. Participants were 256 students enrolled in a management course at a large public university who participated in the study for extra credit and the opportunity to win one of thirteen $25 cash prizes. Table 3 6 reports the demograp hic characteristics for each participant sub sample in this study. As indicated in this table, sub sample A, B, and C were comprised of: approximately the same number of participants (86, 84, and 86, respectively), slightly more males (57%, 55%, and 54%, r espectively) than females, and individuals who spoke English as their first language (85%, 90%, and 86%, respectively). In addition, participants in all three sub samples averaged about

PAGE 106

106 21 years of age (21.18, 21.13, and 21.16, respectively) and had betwe en three and two thirds to four years of previous work experience (4.01, 3.44, and 3.68 years, respectively). Tables 3 7 and 3 8 report the means and standard deviations of the content validity ratings provided by these sub samples. Table 3 7 repor ts the challenge stressor items , broken into three separate surveys (six items per survey), and the construct dimensions provided in the four columns. Table 3 8 reports the hindrance stressor items, broken into three separate surveys (10 11 items per survey) , and the construct dimensions provided in seven columns. Consistent with the previous content validity analysis, I conducted t tests of the item ratings to determine whether each item captured its intended construct. As indicated in Tables 3 7 and 3 8, the ave rage ratings for the items on their intended factors were high (challenge stressor item M = 4.54; hindrance stressor item M = 4.23). In addition, the t tests indicated that the ratings were significantly different (p < .01) in all but two of the cases. The se cases both involved ratings of items that were designed to capture either the role ambiguity or role conflict dimensions. In the does not provide a clear idea of what I should be doing at w dimension (M = 4.12), but also the role conflict dimension ( M = 3.73). These ratings were not found to be significantly differ ent from each other (p > .05) . In the second case, hindrance st by participants as equally capturing its intended role conflict dimension, as well as the role ambiguity dimension (M = 3.48 for both, p > .05). Therefore, with a few exceptions, the data from the second content validation procedure provided further evidence that the challenge and hindrance stressors items did tap their intended factors, and were not contaminated by other related factors.

PAGE 107

107 To assess whether each sub sample performing the item ratings exhibited consistent response patterns, I included one challenge and one hindrance stressor item on each of the three surveys that was exactly the same. The t tests comparing the ratings on these items across the sub sam ples indicated that they were not significantly (p < .01) different from each other. samples did no t differ, on average, in their ratings of this item on its intended factor (M = 4.59, M = 4.49, M = 4.54; all t values non significant at p > .10); or in their ratings of this item on any of the other three factors (all t tests were non significant at p > .10). Similarly, participant sub samples did not differ in their ratings 7, M = 4.43; all t tests were non significant at p > .10); or to any of the other six factors (all t tests were non significant at p > .10). Therefore, given that the three sub samples shared very similar demographic characteristics and response patterns w ith respect to the shared items, I would expect the results for any one of the sub samples to be representative of the others. Step 4: Formally Specify the Measurement Model The fourth step of the construct validation process requires the formal specifi cation of the follow directly from the previous steps, this does not always seem to be the case. For example, n the contemporary strategy literature led these researchers to identify a number of instances in which constructs that were conceptually defined as having formative indicators were nevertheless modeled as having reflective indicators. Given the potentiall y biasing effect of measurement model misspecification (MacKenzie et al.,

PAGE 108

108 2005), it is important that the specification of the measurement model for a newly developed set of measures be explicit and consistent with the underlying theory. Because the challe nge and hindrance stressor constructs are comprised of multiple, distinct dimensions that independently come together to form the higher order constructs, both are modeled as second order constructs having formative first order indicators represented as di mensions; and each of these dimensions were modeled as having multiple reflective indicators. The scale of measurement for the first order dimensions was set latent construct to one of its indicators to 1.0. As noted by Ma cCallum and Browne (1993), the emits paths to: (a) at least two theoretically appropriate reflective indicators, (b) at least two independent latent constructs with reflective indicators, or (c) one reflective indicator and one latent construct with reflective indicators. To resolve this identification problem, I followed order stressor construct as h aving two reflective indicators. The scale of measurement for the second order composite latent respective g lobal reflective indicators to 1.0. The fully specified CFA mod els for the c hallenge and hindrance stressor constructs are depicted in Figure s 3 1 and 3 2, respectively . Step 5: Collect Data and Refine the Measures In the fifth stage of the development and validation of the challenge and hindrance stressor constructs, I collected data from a sample of 178 wo rking MBAs to test the reliability, dimensionality , and the convergent and discriminant validity of the measures. Participants were approached at the end of their classes and asked to complete a survey on the demand s they experience at work. The survey took approximately 20 minutes to complete and the MBAs

PAGE 109

109 received $15 for their participation. The survey included measures of the challenge and hindrance stressor dimensions, as well as the a mount of stress they experie nce at work. Sample Table 3 9 reports the demographic characteristics of the sample used in this study. As indicated in this table, the majority of the participants were Caucasian (78.1%), male (71.9%), married (57.9%), and were on average, approximately 29 years (SD = 4.46) of age. Participants held a range of occupations, but the majority could be classified as either managers/supervisors (23.0%), engineers/scientists (23.0%), or employees in the financial sector (16.3%). On average, the participants ha d approximately three and three quarters years of tenure (SD = 2.56) with their current organization, almost three years of tenure (SD = 2.60) in their current job, and approximately two years of tenure (SD = 1.75) with their current supervisor. As indicat ed in this table, the average GMAT score for this sample was reasonably high (M = 623.27, SD = 56.87). Measures Challenge and hindrance stressors For the purposes of this study, each dimension of the challenge stressor and hindrance stressor constructs wa s measured with four items. The selection of these items was based on the content validation s tudies reported in Step 2 . In addition to measures of the dimensions , participants also were asked to rate two global reflective indicators of the challenge stres sor As noted earlier, the reflective indicators were necessary to identify the higher order formative measurement model s (MacCallum & Browne, 1993; MacKenzie et al., 2003). All items measuring challenge and

PAGE 110

11 0 hindrance stressors were rated on a 7 point Likert sc C hallenge and h indrance related work stres s In order to help establish the convergent and discriminant validity of the newly developed stressor scales, challenge and hindrance related work stress were also measured using item and 5 item measures, respectively. Cavanaugh et al. reported these scales to have adequate reliability, to be factor analytically distinct from each other, and to have different ial relationships with employee job satisfaction and job search behavior. These measures ask participants to the rate the level of stress they experience from several challenging ral hindering assessing these constructs were rated on 5 point Likert like scales with anchors ranging from 1 Analy tical P rocedures The dimensionality and convergent and discriminant validity of the challenge and hindrance stressors scales were assessed in several stages. In the first stage, I conducted a confirmatory factor analysis (CFA) of the fir st order dimensions of the stressor constructs using AMOS 6.0. Following this, I examined the fit of the higher order challenge and hindrance stressor constructs. As I discussed in Step 1 , the higher order challenge and hindrance stressor constructs are mo re appropriately modeled as having formative indicators represented by their first order dimensions . B ecause this study is one of the first that is explicitly designed to report on the development of a higher order construct having formative indicators , I also examined the fit of higher order models treating the first order work stressor dimensions as reflective indicators of the challenge and hindrance stress constructs.

PAGE 111

111 In the next stage of my analysis, I assessed the convergent and discriminant validity of newly designed stressor measures. I assessed the convergent validity of the measures of the vc index developed by Fornell and Larcker (1981). This index represents the average variance accounted for in the items by their respective constructs (dimensions), and is computed by averaging the squared completely that vc should be higher than .50 for the indi cators of a construct to possess convergent validity; t his indicates that more variance in the indicators is accounted for by the underlying construct than by measurement error. These authors have also argued that one can assess the discriminant validity o f the measures of two constructs by comparing their respective vc values with the squared correlation between the constructs. S pecifically, if the vc values for each construct are higher than the squared correlation between the constructs, then constructs are discriminant form each other to the extent that the ir ind icat ors share more variance with each other than they do with indicators of another construct. Next, I assessed the convergent and discrimin a nt validity of the dimensions of the challenge and hindrance stressor constructs by first comparing the dimension level intercorrelations and estimating the correlations between these dimensions and the challenge and hindrance related stress measures developed by Cavanaugh et al. (2000). I supplemented this analysis by calculating means, standard deviations, and reli abilities using SPSS 12.0. Results Confirmatory factor analysis of the challenge and hindrance stressor dimensions Table 3 10 reports the results of the CFA models I included in my analyses. This table reports the chi squared values and associated degree fit index (CF

PAGE 112

112 root mean mean squared error of appr oximation The first model reported in Table 3 10 (Model A) included all the items assessing the dimensions of the challenge stressor construct. As indicated in this table, although the chi squ 2 (98, N = 178) = 182.31, p < .01], the majority of the fit indices indicated that the model adequately fit the data in that the GFI (.96), TLI (.96), and SRMR (.05) all met their recommended cutoff levels. T he only exception was the RMSEA (.07), which approached , but did not meet the value recommended by Hu and Bentler (1998). Although Model A provided evidence supporting the validity of the mea sures of the challenge stressor dimensions, I used SME suggestio ns, comparisons of the item correlations, and an examination of the magnitudes of the factor loadings to refine each dimension to include only three reflective items. I did this to ensure good model fit and to obtain parsimonious measures of the stressor d imensions. The results of the CFA estimating the revised measures of the challenge stressor construct (Model B) are provided in Table 3 10. As indicated in this table, the chi squared value for Model B was not significant [ 2 (48, N = 178) = 51.38, p > .01 ], and the GFI (1.00), TLI (1.00), SRMR (.04), and RMSEA (.02) all exceeded their recommended cutoff values. In addition, all of these values represented substantial improvements over those calculated for Model A. Therefore, this model provided support for the factor structure of the more parsimonious set of measures of the dimensions of the challenge stressor construct. Table 3 11 reports the means, standard deviations, unstandardized factor loadings, t values, and standardized factor loadings for the it ems included in Model B. As indicated in this table, all the challenge stressor items had significant unstandardized factor loadings (all p values < .01) on

PAGE 113

113 values ranged from .73 to .95, with an average of .87). Taken together with the fit indices reported in Table 3 10, these results suggest that Model B fit the data well. I conducted a second CFA analyzing the measures of the hindrance stressor construct. Table 3 10 reports the results of this CFA (Model C). As indicated in this table, the fit indices for Model C provide mixed support regarding its fit to the data. Specifically, the chi squared value [ 2 (303, N = 178) = 540.04, p < .01] for the model was significant, and the GFI (.94) and TLI (.93) and RMSEA (.07) values all approached, but did not meet, their recommended cutoff values. Indeed, only the value calculated for the SRMR (.06) indicated ade quate model fit. Given this evidence, and my desire to create parsimonious measures similar in length to those developed for the challen ge stressor constructs, I refined the measures of the hindrance stressor dimensions to each include three items. Similar to the process used with the challenge stressor construct, I used SME recommendations, item content, comparisons of item correlations, and factor loading magnitudes to refine each hindrance stressor dimension. Following this refinement process, I estimate d a CFA using the revised (three item) versions of the hindrance stressor dimensions. The fit indices for the revised measurement model (Model D) are reported in Table 3 10. As indicated in this table, Model D fit the data very well. Although the chi squa red value for this model was significant [ 2 (168, N = 178) = 233.04, p < .01], the GFI (.98), TLI (.97), SRMR (.04), and RMSEA (.05) values all exceeded their recommended cutoff levels and represented improvements over those reported for Model C. Table 3 12 reports the means, standard deviation s, unstandardized factor loadings, t values and standardized factor loadings for each of the items included in Model D. This table indicates

PAGE 114

114 that all the hindrance stressor items had significant unstandardized loadings (all p values < .01) on their respect ive dimensions, and that the standardized factor loadings were generally very high ( reported in Table 3 10, these results indicate that Model D fit the data well. After revising the dimensions of both the challenge and hindrance stressor constru cts and conducting separate CFAs, I estimated an overall CFA that included all of the challenge and hindrance stressor measures. Table 3 As indicated in this table, Model E fit the data well. The chi squared value for this model was significant [ 2 (440, N = 178) = 549.15, p < .01], and the GFI (.98), TLI (.97), SRMR (.04), and RMSEA (.04) all exceed their cutoff values. Therefore, these results provide additional support indicating that the new measures of the challenge and hindrance stressor dim ensions load on their appropriate constructs, but not on the other dimensions of the stressor constructs. Second order CFA of the challenge and hindrance stressor measures Having established support for the dimensionality for the new challenge and hindran ce stressor measures, I examined the fit of a second order measurement model in which the dimensions served as indicators of the higher order challenge and hindrance stressor constructs. I have argued that the higher order challenge and hindrance stressor constructs should be modeled as having formative first order dimensions. However, because I am not aware of any studies that have reported an analysis using this structure, and because most previous research using higher order measurement models has genera lly modeled these constructs as having reflective indicators, I estimated both models and compared the results. Consistent with my theoretical arguments, I first estimated a CFA with the challenge and hindrance stressors modeled as second order constructs having formative indicators represented by their respective dimensions. For the purpose of identifying this model, I added paths from the

PAGE 115

115 second order stressor constructs to the two global reflective indicators (MacCallum & Browne, 1993; MacKenzie et al., 2005 ). The fit indices for this CFA (Model F) are reported in Table 3 10. Generally, the evidence in this table suggests that Model F adequately fit the data. The chi squared value for this model was significant [ 2 (563, N = 178) = 744.09, p < .01], and the GFI (.96), TLI (.95) SRMR (.07), and RMSEA (.05) all met their recommended cutoff values. Table 3 13 reports the unstandardized parameter estimates, t values, and standardized parameter estimates for the challe nge stressor dimensions, the hindrance stressor dimensions, and the respective global reflective indicators estimated in Model F. As indicated in the left side of this table, the unstandardized parameter estimate for the second global reflective indicator of the challenge stressor construct was positive and significant (B = .98, p < .01), and the standardized parameter estimates for both the first and second global reflective indicators were large ( parameter estimates for the work pace and job complexity dimensions were significant and he workload and job responsibility dimensions did not have significant relationships wit h the high er order challenge provides support for the notion that the second order formative challenge stressor construct converged with reflective indicators designed to assess the overall construct, and that two of challenge stressors dimensions made significant empirical contributions to the second order construct the y form. Table 3 13 also provides evidence for the construct validity of the second order hindrance stressor construct. This table reports that the unstandardized parameter estimate for the second global reflective indicator of the hindrance stressor const ruct was positive and significant (B =

PAGE 116

116 .94, p < .01), and that both the first and second global reflective indicators had similarly large that the role confl ict, resource inadequacies, organizational politics, and job insecurity dimensions had significant, positive relationships with the second order hindrance stressor ctively, all p values < .05); whereas the role ambiguity, interpersonal conflict, and administrative hassles .05). Therefore, the evidence presented in this table suggests that the second order formative hindrance stressor construct converged with its global reflective indicators, and that a majority of the hindrance stressor dimensions made significant empirica l contributions to the second order construct they form. Following this analysis, I estimated a CFA that included both challenge and hindrance stressors as second order constructs having reflective indicators represented by their dimensions. The fit indice s for this model (Model G) are reported in Table 3 10. As indicated in this table, although the chi squared value for Model G was significant [ 2 (483, N = 178) = 619.28, p < .01], the GFI (.97), TLI (.97), SRMR (.06), and RMSEA (.04) all met their recomme nded cutoff values. Therefore, the second order reflective indicator model also seemed to fit the data well. The unstandardized parameter estimates, t values, and standardized parameter estimates for Model G are reported in the right side of Table 3 13. A s indicated in this table, all of the unstandardized estimates for the dimensions of the challenge and hindrance stressor constructs were significant and positive (all p values < .05). However, there was substantial variance in the standardized parameter e stimates for the dimensions of both the challenge stressor construct

PAGE 117

117 vc values fo r the higher order challenge ( vc = .39) and hindrance stressors ( vc = .35) constructs indicated that the dimensions do not meet the requirements for convergent validity, and moreover, that the first order dimensions do not fit presentations of their underlying second order stressor constructs. Based on the evidence provided above, it would appear that measurement models treating the challenge and hindrance stressor dimensions as second order reflective or formative indicators b oth fit the data fairly well in an absolute sense. Although the second order reflective model probably could be said to fit somewhat better in a relative sense, this model did not meet criteria necessary for establishing the convergent validity of the stre ssor constructs when modeled as having reflective indicators. In addition, it is important to remember that: (a) the determination of whether a construct should be modeled as having formative or reflective indicators is more of a conceptual matter than it is an empirical one (cf. Diamantopoulos & Winklhofer, 2001; MacKenzie et al., 2005); and (b) because the dimensions of both the challenge and hindrance stressor constructs are distinct from each other, are not necessarily interchangeable, and are likely to have different antecedents and consequences, they are more appropriately v iewed as formative indicators of their respective higher order constructs. Reliability, convergent validity, and discriminant validity at the item level In addition to establishing the underlying factor structure of the challenge and hindrance stressor measures, it is also important to demonstrate that these measures posses internal consistency reliability, as well as convergent and discriminant validity. Therefore, in the following sections, I turn my attention to these analyses. In the first section, I examine the reliability and convergent and discriminant validity of the challenge and hindrance stressor 1981) measure of

PAGE 118

118 vc . Following this, I examine the relationships between these measures at the dimension level. Finally, I examine the relationships between the dimensions of the challenge and hindrance stressors, the global reflective measures designed to tap these const ructs, and the challenge and hindrance related work stress measures designed by Cavanaugh et al. (2000). Table 3 14 reports the means, standard deviations, internal consistency reliabilities, average shared variance estimates ( vc ), and latent construct i ntercorrelations for both the challenge and hindrance stressor dimensions. The data reported in this table indicates that the internal consistency reliability estimates for all of the challenge stressor dimensions were above recommended level of .70; and the vc values exceeded Fornell provides evidence for the convergent validity of the measures of the challenge stressor dimensions. Evidence of the discriminant valid ity of these dimensions was also quite good, in that the variance shared among any pair of these dimensions was always less than average variance explained in the measures by their respective constructs (Fornell & Larcker, 1981). Thus, taken together, the reliability coefficients, vc values, and correlational analyses provide strong evidence of the convergent and discriminant validity of the dimensions of the challenge stressor construct. Table 3 14 also indicates that the internal consistency reliability estimates for the hindranc level of .70; all of the construct vc level of .50; and that the variance shared among any pair of these dimensions w as always less than average variance explained in the measures by their respective constructs (Fornell &

PAGE 119

119 Larcker, 1981) Thus, these findings also provide strong support for both the convergent and discriminant validity of the measures of the hindrance stre ssor dimensions. Assessment of convergent and discriminant validity at the dimension level In order to assess the dimension level convergent and discriminant validity of the challenge and hindrance stressors measures, I examined the relationships among th e challenge and hindrance stressors reported in Table 3 14. There are three patterns reported in this table that deserve attention. First, with the exception of the relationship between job complexity and work pace (r = .16, p > .05), all other relationshi ps between the four dimensions of the challenge stressor construct were positive and significant (all p values < .05). It is important to note, however, that the size of these correlations tended to be moderate in nature (average r = .34). The one exceptio n to this was the relationship between workload and work pace, which were correlated somewhat more strongly (r = .63). This correlation may explain why work pace, and not workload was found to load on the second order formative challenge stressor construct . Despite this finding, the data are generally supportive of the expectation that the challenge stressor dimensions are related, yet distinguishable from each other. Second, although the majority of the hindrance stressor dimensions have significant, posi tive correlations with each other (16 of 21 p values < .01), the size of their correlations were relatively moderate (average r = .30). The dimension that generally had the lowest correlations with the other hindrance stressor dimensions was job insecurity (average r = .11); whereas the dimension that generally had the highest correlations with the other hindrance stressor dimensions was role conflict (average r = .39). However, even the highest correlation reported in the table for any two dimensions of th is construct (r = .54 between role conflict and resource inadequacies) indicated that these measures shared at most, about 26% of their variance.

PAGE 120

120 Third, the dimensions of the challenge stressor construct tended not to be related with the dimensions of the hindrance stressor construct. Indeed, the average correlation between these constructs was essentially zero, (average r = .01) and only one relationship (work pace and role ambiguity r = .19) was found to be significant (p < .05). Thus, taken together, th ese findings indicate that the dimensions of the hindrance stressor construct generally share significant, but not overwhelming amounts of their variance with each other; and that the dimensions of these respective stressor constructs are empirically disti nct from each other. Following the examination of the correlations between the challenge and hindrance stressor dimensions, I calculated co rrelations between the stressor constructs and (a) the global reflective indicators of these constructs and (b) the challenge and hindrance related work stress scales developed by Cavanaugh et al. (2000). The results of these analyses are reported in Table 3 15. Generally speaking, the evidence presented in this table provides additional support for the convergent and discriminant validity of the dimensions of the challenge stressor construct. As indicated in the first two columns of this table, the dimensions of the challenge stressor construct generally had positive relationships with both of the global reflective ch allenge stressor indicators (average r = .26, all p values except one < .05), but had no relationships with both of the global reflective hindrance stressor indicators (average r = .01; all p values > .05). In addition, the workload and work pace dimensio ns had positive relationships with the challenge related work stress scale (r = .58 and r = .59, respectively, both p values < .01); whereas none of the dimensions of the challenge stressor construct exhibited a significant relationship with the hindrance related work stress scale (average r = .04; all p values > .05) . Table 3 15 also provides additional support for the convergent and discriminant validity of the dimensions of the hindrance stressor construct. The results reported in this table indicate th at

PAGE 121

121 the dimensions of this construct generally had strong, positive correlations with both of the global, reflective hindrance stressor indicators (average r = .38; all p values < .01) ; but that these same dimensions generally had negative correlations with both of the global, reflective challenge stressor indicators (average r = .19; 9 of 14 p values < .05) . This table also reports that all the dimensions of the hindrance stressor construct had strong, positive correlations with the hindrance related work stress scale (average r = .41; all p values < .01); but that these same dimensions generally had no relationships with the challenge related work stress scale (average r = .06; six of seven p values > .05) . Thus, Table 3 15 provides further evidence for th e convergent and discriminant validity of the challenge and hindrance stressor measures in that the dimensions of these constructs were found to have: (a) relationships with the global reflective indicators designed to tap their respective overall construc ts, (b) relationships with scales designed to assess the stress resulting from challenging and hindering work demands; but (b) either no relationships, or negative relationships with measures that were designed to assess the other stressor construct. In su mmary, the evidence presented above provides strong support for the item and dimension level convergent and discriminant validity of the measures of the challenge and hindrance stressor constructs. Specifically, the results reported from multiple CFAs, re liability analyses, and the comparison of correlations suggests that the stressors items do indeed load on their purported measures, that the three item measures of the stressor dimensions are distinct from one another, and that these dimensions represent the overall challenge and hindrance stressor constructs. Therefore, the next step in the validation procedure was to assess the nomological validity of these measures.

PAGE 122

122 Step 6: Specify the Nomological Network and Test the Hypothesized Relationships In thei r discussion of the importance of establishing the construct validity of the measures erlocking laws which speaking, the nomological network is represented by the relationships between a focal construct and other constructs that should be conceptually related to it. Therefore, in the sixth step of the construct development process I tested the nature of the relationships between the focal constructs and the other employee criteria. To examine the nomological validity of the challenge and hindrance st ressors measures, I gathered data to test the relationships between challenge and hindrance stressors and employee strains, emotions, job satisfaction, motivation, and employee performance. Consistent with the previous analysis, I contrasted differences ac ross construct mod eling techniques by estimating these structural relationships using one model in which the challenge and hindrance stressors constructs were modeled as having formative indicators and one model in which the stressor constructs were modele d as having reflective indicators. Sample The same sample of 178 working MBA students that was used in Step 5 was also used to analyze the nomological validity of the challenge and hindrance stressors measures. When completing the questionnaires containin g the measures of the challenge and h indrance stressors, the participants also completed self report m easures of job related strains, demand related emotions, job satisfaction, and motivation. After completing this survey, the participants were asked to pr ovide their immediate supervisor with a packet I gave to them. This sealed packet contained a brief explan ation of the study, a survey, a self addressed stamped envelope , and a $5

PAGE 123

123 bill that the supervisor was asked to keep after completing the survey. The survey asked each related behaviors, OCB, and CWB. After completing the questionnaire, the supervisor mailed me th e survey using the envelope included in the packet. I received usable data from 10 1 supervisors (56% of the sample). Measures Challenge and hindrance stressors The three item measures of the dimensions of the challenge and hindrance stressors used in Step 5 were employed in this part of the study. In addition, each overall stressor co nstruct was also assessed with the two global reflective indicators discussed in the previous step. Therefore, the challenge stressor construct was measured with a total of 14 items and the hindrance stressor construct was measured with a total of 23 items . Negative affectivity Employee trait based negative affectivity was measured using 10 items taken from the Positive and Negative Affect Schedule (PANAS) scale reported by Watson, Clark, and Tellegen (1988). This scale asks participants to indicate the e xtent to which they generally are this scale in the present study was .82. Job related strains Employee job related strains were measured with the 3 item fr ustration scale developed by Peters et al. (1980) and the 9 item emotional exhaustion scale developed by Maslach and , & Heinish, 1999; Jex & Gudanowski, 1992; Peters et al., 1980; Leiter & Jackson, 1988 ; Spector et al. , 1988 ; Wright & Bonett, 1997 )

PAGE 124

124 have found these scales to possess adequate internal consistency reliabilities . Both of the se Stressor related emotions related emotions were measured using Folkman and L 3 item threat, 3 item challenge, 5 item harm, and 4 item benefit emotion scales. These items ask participants to rate the extent to which the demands they experience at work generally cause them to feel a specific set of emotions. Example it research (Folkman & Lazarus, 1985; 1986; 1988) has shown these scales to possess acceptable reliabilities. In the present study, the alphas for these measures were .81 for the threat emotion scale, .76 for the challenge emotion scale, .77 for the harm emotion scale, and .85 for the benefit emotions scale. Motivation I assessed employee motivation using a four item scale dev eloped for this study. The items in this scale reflect dimensions of directed behavior and effort, and are consistent with other levels of effort to complete ta Job satisfaction I used the 3 item measure developed by Cammann, Fichman, Jenkins, and Klesh (1983) to assess general employee job satisfaction. This scale has been used in a number of previous st udies (McFarlin & Rice, 1992; McLain, 1995; Siegall & McDonald, 1995), and has been found esent study.

PAGE 125

125 Task performance Supervisors rated the task performance of their employees using a modified 4 item version of the scale developed by Williams and Anderson (1991). An example item from this scale is to be factor analytically distinct from OCB, and previous research (Funderberg & Levy, 1997; Morrison & Phelps, 1999; Thompson & Werner, 1997 ; Williams & Anderson, 1991 ) has reported it to possess good relia bility, with alphas ranging from .80 to .94. In the present study, the reliability coefficient of this scale was .89. Organizational citizenship behavior Supervisors rated the organizational citizenship behaviors of their employees using Lee (2002) 16 Willingly gives of his/her time to help coworkers who have work arch using this scale (Finkelstein & Penner, 2004; Jackson, Colquitt, Wesson, & Zapata Phelan, 2006; Piccolo , & Colquitt, 2006) has shown it to possess good reliabilities. In the present study, the coefficient alpha for this scale was .90. Counterproductiv e work behavior Supervisors rated the counterproductive work behaviors of their employees using 11 items taken from the workplace deviance scale developed by Bennett and Robinson (2000). Items from Works on a personal matter instead of work for the employer, Robinson reported this scale to be reliable. In the present study, coefficient alpha was .92. Items measuring the dimensions of the chal lenge and hindrance stressors constructs, employee strains, job satisfaction, motivation, and negative affectivity were rated on a 7 point

PAGE 126

126 supervisors rated items refle cting task performance, OCB, and CWB on 5 point Likert scales emotions constructs were rated on a 5 point Likert slight Analyt ical P rocedure I tested the criterion related validity of the challenge and hindrance stressor constructs using a three step process. First, I calculated the correlations between the dimensions of the challenge and hindrance stressor constructs and the employee criteria. Following this, I estimated the structural relationships between the second order stressor constructs modeled as having formative indicators and the other employee criteria. I compared these esti mates with results of a structural model that estimated the second order challenge and hindrance stressor constructs as having reflective indicators represented by their respective dimensions. Finally, I compared both of the models to a model in which the stressors dimensions were specified as individual predictors of each employee criterion. All models were estimated using AMOS 6.0. Results Criterion related validity of the dimensions Table 3 16 reports correlations of the challenge and hindrance stressor dimensions with employee strains, emotions, job satisfaction, and motivation. As expected, this table reports that dimensions of both stressor constructs generally had positive relationships with measures of employee strains. Specifically, the workload, jo b responsibility, and work pace dimensions of the challenge stressor construct were all found to have significant positive relationships with both employee frustration (r = .29, r = .18, and r = .29, respectively, all p values < .05) and emotional exhausti on (r = .47, r = .25, and r = .38, respectively, all p values < .01); and the role ambiguity,

PAGE 127

127 role conflict, resource inadequacies, interpersonal conflict, organizational politics, and administrative hassles dimensions of the hindrance stressor construct w ere all related to both employee frustration (r = .40, r = .48, r = .38, r = .32, r = .39, and r = .32, respectively, all p values < .01) and emotional exhaustion (r = .31, r = .37, r = .38, r = .31, r = .40, and r = .37, respectively, all p values < .01). Therefore, despite the fact that job complexity was not related to either emotional exhaustion or frustration (both r values = .01, both p values > .05) and job insecurity was positively related to emotional exhaustion (r = .18, p < .05), but not frustra tion (r = .09, p > .05), this table provides support for the notion that dimensions of both the challenge and hindrance stressor constructs were associated with employee s train. Table 3 16 also provides evidence supporting the criterion related validity of the dimensions of the challenge and hindrance stressor constructs with respect to the differential relationships these stressors were expected to have with employee outcomes. For example, the workload, job responsibility, and job complexity dimensions of the challenge stressor construct were found to have positive relationships with employee challenge related emotion (r =.21, r = .21, and r = .32, respectively, all p values < .05); whereas the role ambiguity, role conflict, resource inadequacies, organizat ional politics, and administrative hassles dimensions of the hindrance stressor construct had negative relationships with this type of employee emotion (r = .40, r = .30, r = .27, r = .38, and r = .20, respectively, all p values < .05). Although job c omplexity was the only challenge stressor dimension found t o be significantly correlated with benefit emotions (r = .26, p < .01), all of the dimensions of the hindrance stressor construct, except job insecurity (r = .10, p > .05), had negative correlatio ns with this criteria (average r = .31; six of seven p values < .01).

PAGE 128

128 I n addition, this T able 3 16 indicates that all the dimensions of the hindrance stressor construct, except for administrative hassles (r = .14, p > .05), had positive relationships wit h threat related emotions (average r = .24; six of seven p values < .05). In contrast, the workload and work pace dimensions of the challenge stressor construct were found to have significant positive relationships with this type of emotion (r = .36 and r = .29, respectively, both p values < .01); whereas job responsibility and job complexity had non significant relationships with this criteria (r = .12 and r = .03, respectively, both p values > .05). Finally, all of the dimensions of the hindrance stressor construct, except for job insecurity (r =.15, p > .05), were found to have positive relationships with harm related emotion (average r = .39; six of seven p values < .01); whereas none of the dimensions of the challenge stressor construct had significant relationships with th is criteria (average. r = .07; all p values > .05). Table 3 16 also indicates that the dimensions of the challenge and hindrance stressor constructs generally have differential relationships with employee job satisfaction and motivati on. Specifically, although all the dimensions of the hindrance stressor construct had negative relationships with employee job satisfaction (average r = .39; all p values < .01); the dimensions of the challenge stressor construct tended to have non signif icant relationships with this criterion (average r = .08; three of four p values > .05). The one exception to this was the job complexity dimension , which was found to have a significant positive relationship (r = .27, p < .01) with job satisfaction . In ad dition, this table reports that the workload, job complexity, and work pace dimensions of the challenge stressor construct had positive relationships with employee motivation (r = .25, r = .28, and r = .27, respectively, all p values < .01); whereas all of the dimensions of the hindrance stressor construct, with the exception of interpersonal conflict (r = .11, p > .05), had

PAGE 129

129 negative relationships with this criteria (average r = .23; six of seven p values < .01). In summary, the results reported in this t able generally support the criterion related validity of the stressor construc ts in that the dimensions of this construct tended to be positively relate d to strains, challenge related emotion , motivation , and to a lesser extent, job satisfaction; whereas t he dimensions of the hindrance stressor construct tended to be positively related to strains and threat and harm related emotions, and negatively related to challenge and benefit related emotions, job satisfaction, and employee motivation. Criterion rel ated validity of the second order stressor constructs To assess the criterion related validity of the second order stressor constructs , I estimated a structural model testing the paths between these constructs and employee strains, emotions, job satisfacti on, and motivation . Th e standardized parameter estimates calculated from this model are reported in Table 3 17. This table reports the results of the structural models in which the second order stressor constructs are modeled as having formative first orde r indicators presented by their respective dimensions . In general, the results reported in this t able provide support for the criterion related validity of these constructs. F or example, both challenge and hindrance stressors had positive relationships wit respectively, both p values < .0 1); frus tration. This generally indicates that both types of work demands were perceived as stressful and were correlated with some forms of adaptive employee responses. Table 3 17 also reports that challenge stressors had positive relationships with both challen ge and benefit .33, and .58, respectively, both p values < .01). Conversely, hindranc e stressors were positively

PAGE 130

130 related to threat and harm .01). Although it was somewhat surprising that challenge stressors were positively related to threat < .01), these demands did not have a significant relationship with harm Finally, the results reported in Table 3 17 provided support the expected relationships between the second order challenge and hindrance stressor constructs and employee job satisfaction and motivation. As indicated in this table, challenge stressors had positive whereas hindrance stressors were found .40, p < .01). Following this analysis, I estimated a second structural model e xamining the relationships between the second order formative challenge and hindra nce stressor constructs (modeled as having formative indicators) and the employee performa nce criteria. This was done separately because there were a substantial number of participants that did not receive performance evaluations from their supervisors and I wanted to maintain the highest level of power possible for each set of estimated relationships. The results of this analys i s are also reported in Table 3 17. As indicated in this table, all the parameter estimates for the structural paths between the ch allenge and hindrance stressor constructs and the performance criteria were in the expected directions, but the estimates for the relationships between the second order (formative) challenge stressor construct and task performance, OCB , and CWB were not fo .14, all p values > .05). Similarly, the relationships between the hindrance stressor construct (when modeled as having formative indicators) and task performance, OCB, and CWB were also found to be non sig

PAGE 131

131 .05). Although the results from this analysis did not provide much support for the criterion related validity of the stressor measures, this may be a result of the moderate s a mple size (N = 101) , or range restriction in these measures (see Limitations section in Chapter 4) Comparison of second order stressor constructs Although I have argued from a the oretical perspective that it is more appropriate to model the challenge and hindrance stressor const ructs as having formative indicators comprised of their dimensions, i t may also be informative to examine the relationships between these constructs and the criteria when the second order stressors were modeled as having reflective indicators. The standard ized parameter estimates fro m this analysis are reported in the bottom half of Table 3 17. Generally speaking, a comparison of the pattern of relationships reported in Table 3 17 indicates that the relationships between the challenge and hindrance stresso rs and the employee criteria are fairly similar when the second order constructs are modeled as having formative compared with reflective indicators. For example, all of the relationships between the stressor constructs maintained the same sign and relativ e magnitude across the two different modeling techniques. In deed, there were only three cases (i.e., the challenge stressor frustration, the hindrance stressor OCB, and the hindrance stressor CWB relationships) in which the significance of a relationship c hanged across the two analyses. Although this comparison suggests that empirical results are fairly similar, additional comparisons with other samples will be needed to be conducted before any definitive conclusions can be made . Criterion related validity controlling for negative affectivity As indicated in Chapter 2 , several researchers (cf. Brief et al., 1988; Burke et al., 1993; Schaubroeck et al., 1992) have argued that negative affectivity may confound the relationships between measures of work stress ors and employee strains, job attitudes, and behaviors.

PAGE 132

132 Therefore, to assess the potential confounding effect of this trait , I controlled for participant negative affectivity in the relationships between the second order stressor constructs and the employe e criteria examined in this study. The results of the structural relationships for the s econd order formative and reflective models , controlling for NA, are reported in Table 3 17. Generally speaking, there are two important findings in regards to this an alysis. First, controlling for NA did not have an effect on the significance o r sign of the vast majority of the relationships between the stressor constructs and the criteria. There were a few exceptions to this as, after controlling for NA: (a) the previ ously non significant relationships between the challenge stressor construct ( modeled as having formative indicators ) .20, respectiv ely, both p values < .05); (b) the previously positive relationship between the challenge stressor construct ( modeled as having formative indicators ) significant . 13, p > .05); and (c) the previo usly positive relationships between the hindrance stressor construct (when modeled as having formative indicators and when modeled as having reflective indicators ) and ively, both p values < .01) both became non Second, significant changes in the magnitude of the parameter estimates were more prevalent when the stressor constructs were modeled as having formative indicators than when they were modeled as having reflective indicators. More specifically, when the stressor constructs were modeled as having formative indicators, the magnitude of the relationships between the challenge stressor constructs and three of the employee criteria (i.e., fr ustration, emotional exhaustion, and harm emotions ) , and the magnitude of the relationships between the

PAGE 133

133 hindrance stressor construct and seven of the employee criteria (i.e., frustration, emotional exhaustion, challenge emotions, benefit emotions, threat e m otions, harm emotions, and CWB) changed by approximately 50% or more (either stronger or weaker) after controlling for NA. Conversely, when the stressor constructs were modeled as having reflective indicators, only the relationship between challenge stres sor s and harm emotions and the relationships between the hindrance stressor s and two employee criteria (i.e., harm emotions and task performance) were altered by at least 50% after controlling for employee NA. In summary, it seems that the relationships be tween challenge and hindrance stressors and employee strains, emotions, job satisfaction, motivation, and performance were fairly consistent, even after controlling f or NA; but that the effect s of NA may be more prevalent when the constructs are modeled as having formative indicators versus when there are modeled as having reflective indicators. Supplemental analysis To compare the relationships obtained from analyses using the second order stressor constructs with the relationships obtained from analysis using more tradit ional regression techniques, I c onducted a supplemental analysis. This analysis required that the employee criteria be regressed on all the first order dimensions of the challenge and hindrance stressor constructs. Therefore, this analysis assessed the unique effect of each stressor dimension, controlling for all other dimensions. Because the sample size for the performance data w as substantially different from the data gathered for all other criteria, two separate models were estimated. Th e first estimat ed the relationships between the stressor dimensions and all the employee criteria except for task performance, OCB, and CWB. The second estimat ed the relationships between the stressor dimensions and only the performance criteria. The resul ts of these analyses are reported in Table 3 18.

PAGE 134

134 Most importantly, Table 3 18 reports that the majority of the parameter estimates for the relationships between the first order dimensions and the employee criteria were non s ignificant. For example, none o f the hindrance stressor dimensions were found to have significant relationships with emotional exhaustion (all p values > .05), task p erformance (all p values > .05) and CWB (all p values > .05); and only one relationship between the stressor dimensions and frustration (six of seven p values > .05), challenge emotions (six of seven p values > .05), threat emotions (six of seven p values > .05), and motivati on (six of seven p values > .05), respectively, was found to be significant. Finally, a comparison o f t he findings reported in Table 3 18 with those reported in Table 3 17 indicates that the challenge and hindrance stressors (when modeled as either second order constructs) seemed to explain at least as much, and in several cases more, variance in most cr iteria than did the individual first order dimensions. For example, the c hallenge and hindrance stressors (when modeled as having reflective indicators) explained more variance in emotional exhaustion, challenge emotions, benefit emotions, threat emotions, job satisfaction, and motivation than did the combined effects of the first order dimensions. Although the perform ance data was an exception to this general finding, the limitations may make it difficult to generalize these findings .

PAGE 135

135 Figure 3 1. Overv iew of the scale development and validation process.

PAGE 136

136 Figure 3 2. CFA of second order challenge stressor construct having formative indicators represented by its dimensions.

PAGE 137

137 Figure 3 3. CFA of second order hindrance stressor construct having formati ve indicators represented by its dimensions.

PAGE 138

138 Table 3 1. Summary of challenge hindrance stressor dimensions identified in previous research. Studies Challenge stressors Hindrance stressors Evidence of construct validity Work load Work pace Job comple xity Job responsibility Role ambiguity Role conflict Resource inadequacies Organizational politics Administrative hassles Interpersonal conflict Job insecurity Cavanaugh et al. (2000) X X X X X X X First, three of the authors conducted a Q sort of it ems taken from several popular measures of work related stressors, categorizing the stressors into those that were challenging versus those that were hindering. A sorting procedure was then conducted on the job demands identified as challenging or hinderi ng by four individuals not associated with the project. This procedure provided general, independent support for the categorization of job demands into challenges and hindrances. The authors conducted a CFA which provided support for the challenge hindran ce stressors framework. Boswell et al. (2004) X X X X X X X These authors conducted a confirmatory factor analysis on the items used by Cavanaugh et al. (2000), finding support for separate dimensions of challenge and hindrance stressors LePine et a l. (2004) X X X X X X a Critical incidents of stressful events that were perceived to either promote or hinder growth and accomplishment were gathered from participants to develop challenge and hindrance stressor items. These items were then sorted by two of the authors into categories of challenge and hindrance related stressors. A second set of participants coded the stressor items into challenges and hindrances. The participant categorizations were reported to be very similar to the categorizations done by the authors. These authors conducted a CFA and reported that the challenge and hindrance stressors constructs were distinct from each another and that their measures were valid. Bingham et al. (2005) X X A CFA conducted by the s e authors f ound that challenge and hindrance stressors were distinct from each another and that their measures were valid.

PAGE 139

139 Table 3 1. Continued Studies Challenge stressors Hindrance stressors Evidence of construct validity Work load Work pace Job complexity J ob responsibility Role ambiguity Role conflict Resource inadequacies Organizational politics Administrative hassles Interpersonal conflict Job insecurity LePine et al. (2005) X X X X X X X 43 part time MBAs assessed the extent to which the job demand s used in this study acted as challenge stressors and/or hindrance stressors. The job demands that were identified as hindrance stressors in the study were rated as significantly more hindering than the job demands identified as challenge stressors; and th at the job demands that were identified as challenge stressors were rated as significantly more challenging than were the job demands identified as hindrance stressors. This study also found that both challenge and hindrance stressors positively related t o strains, indicating that both these forms of job demands are stressful. T his study provides meta analytic evidence for the criterion related validity of the challenge and hindrance stressors scales in that these stressors had differential relationships w ith employee motivation and performance. Podsakoff et al. ( 2007 ) X X X X X X X This study finds that challenge and hindrance stressors are positively related to strains, indicating that both these forms of job demands are stressful. T his study provid es meta analytic evidence for the criterion related validity of the challenge and hindrance stressors in that these stressors had differential relationships with employee job satisfaction, commitment, intentions to leave, turnover , and withdrawal behaviors . Zapata et al. (Under Review) X X X X X X X X X A CFA conducted by these authors found that their challenge and hindrance stressors were distinct from each another and that the measures were valid. Podsakoff et al. ( Under review ) X X X X X X X X X A CFA conducted by these authors found that their challenge and hindrance stressors were distinct from each another and that the measures were valid. Note: a T his study included measures of hassles that were not necess arily administrative in nature.

PAGE 140

140 T a ble 3 2. Construct conceptual domains for challenge and hindrance stressors. Construct/Dimensions Definition Dimensionality and Nature of Relationships Challenge Stressors Job demands that tend to be perceived as promoting the accomplishment of personal goal s and facilitate growth and development in the work context. Multi dimensional: Second order construct with formative indicators represented by first order dimensions. Workload The perception of the number or quantity of job tasks that one is assigned t o complete. Unidimensional: First order latent variable with reflective indicators. Work pace The perception of the speed with which one has to complete assigned tasks Unidimensional: First order latent variable with reflective indicators. Job complexi ty The breadth or variety of skills required to complete assigned job tasks. Unidimensional: First order latent variable with reflective indicators. Job responsibility The perceived accountability that employee feel for their own performance and the p erformance of others in the organization. Unidimensional: First order latent variable with reflective indicators. Hindrance Stressors Job demands that tend to be perceived as hindrances to the accomplishment of personal goals and impede growth and devel opment in the work context. Multi dimensional: Second order variable with formative indicators represented by first order dimensions. Role ambiguity The perception that there is uncertainty regarding what actions are appropriate to take to fulfill a rol e (Kahn et al., 1964) Unidimensional: First order latent variable with reflective indicators. Role conflict The perception that there is incompatibility between the expectations of parties or between aspects of a single role. (Kahn et al., 1964) Unidimens ional: First order latent variable with reflective indicators. Resource inadequacies The perception that tools, equipment, materials, and/or supplies are insufficient to adequately complete job tasks. Unidimensional: First order latent variable with refl ective indicators. Organizational politics The perception of unsanctioned influence attempts that seek to promote self interest at the expense of organizational goals. Unidimensional: First order latent variable with reflective indicators. Administrati ve hassles The perception that the organization has excessive or unnecessary must deal with in the completion of their work. Unidimensional: First order latent variable with refl ective indicators. Interpersonal conflict The perceived tension and frustration about personal differences concerning interpersonal style, preferences, attitudes, and personality. Unidimensional: First order latent variable with reflective indicators. Job insecurity The perception that one is powerless to maintain desired continuity in a threatened job situati on. (Greenhalgh and Rosenblatt, 1984 ) Unidimensional: First order latent variable with reflective indicators.

PAGE 141

141 Table 3 3. Demographic characte ristics of the first undergraduate content validation sample. Characteristic M (SD) N % of sample Age (in years) 20.47 (0.76) 44 100 Work experience (years) 3.18 (2.16) 44 100 Gender Female 27 61 Male 17 39 English 1 st Language Yes 38 86 No 6 14 Note: N = 44.

PAGE 142

142 Table 3 4. Means and standard deviations from the first content validation of the challenge stressor items. Items Work load Job responsibility Job complexity Work pace 1. Tasks on my job use a va riety of different skills and abilities. 2.34 (1.33) 2.02 (1.45) 4.34 (1.01) 2.16 (1.29) 2. I have to complete a great deal of work on this job. 4.00 (1.28) 2.41 (1.30) 2.40 (1.29) 3.05 (1.56) 3. My job requires me to be accountable for my work a nd the work of others. 2.14 (1.32) 4.45 (1.13) 2.21 (1.36) 2.07 (1.28) 4. My job requires a complex set of behaviors and skills. 2.05 (1.22) 2.18 (1.45) 4.39 (1.06) 2.02 (1.28) 5. To complete my work on time, I must work quickly. 2.98 (1.58) 2.18 (1.32) 2.21 (1.44) 4.57 (1.00) 6. I have to work at a rapid pace to complete my job. 2.66 (1.54) 2.27 (1.42) 2.00 (1.21) 4.43 (1.07) 7. My job requires me to use a broad set of skills and abilities. 2.23 (1.39) 2.12 (1.28) 4.49 (1.03) 1.98 (1.08) 8 . My work is characterized by tight time deadlines. 2.55 (1.47) 2.27 (1.45) 2.09 (1.31) 4.16 (1.38) 9. This job requires that I be responsible for the productivity of myself and others. 2.32 (1.36) 4.39 (1.06) 2.20 (1.34) 2.07 (1.32) 10. My job requ ires that I work at a fast tempo. 2.89 (1.56) 2.18 (1.35) 2.02 (1.30) 4.48 (1.09) 11. I am responsible for the performance of myself and my coworkers. 2.09 (1.43) 4.48 (0.95) 2.16 (1.35) 1.80 (1.13) 12. The work I do calls for several different talents and skills. 2.55 (1.34) 2.41 (1.47) 4.48 (1.02) 2.05 (1.26) 13. I experience time pressure to complete work on my job. 2.89 (1.48) 2.43 (1.47) 2.16 (1.45) 4.32 (1.03) 14. This job requires that I complete a lot of work. 4.30 (1.15) 2.14 (1.30) 2.23 ( 1.26) 2.77 (1.44) 15. In this job, I am responsible for the work of a lot of other people. 2.30 (1.36) 4.66 (0.68) 1.84 (1.10) 2.09 (1.39) 16. I am required to use a variety of different skills on my job. 2.57 (1.53) 2.20 (1.30) 4.43 (0.97) 1.89 (1.17) 17. My work load on this job is large. 4.14 (1.27) 2.36 (1.42) 2.14 (1.25) 2.59 (1.56) 18. I have to complete a lot of tasks each day at work. 4.41 (0.95) 2.14 (1.25) 2.30 (1.15) 3.09 (1.57) 19. I am responsible for making sure that work is completed on my job. 2.43 (1.40) 4.23 (1.20) 2.07 (1.26) 2.68 (1.58) 20. I have a lot of assignments to complete on my job. 4.50 (0.76) 2.34 (1.40) 2.30 (1.23) 2.75 (1.51) Note: Scale anchors ranged from 1 ( definition ) to 5 ( ). The first number in each cell represents the mean; the number in parentheses is the standard deviation. Intended factors for each item are indicated in bold. All t tests comparing ratings on the intended factor versus all other factors were significantly different at p < .01.

PAGE 143

143 Table 3 5. Means and standard deviations from the first content validation of the hindrance stressor items. Item Role ambiguity Role conflict Resource inadequacies Organizational politics Administ rative hassles Interpersonal conflict Job insecurity 1. I cannot complete my tasks because I am not given adequate equipment, tools, materials, or supplies. 2.23 (1.44) 2.11 (1.35) 4.23 (1.33) 1.66 (1.08) 1.70 (1.00) 1.77 (1.20) 1.77 (1.20) 2. I of ten receive conflicting requests from a supervisor (or supervisors). 3.09 (1.46) 4.16 (1.24) 1.66 (1.01) 2.23 (1.40) 2.14 (1.30) 2.61 (1.56) 1.91 (1.27) 3. complete my job. 2.52 (1.50) 2.61 (1.47) 1.91 (1.24) 2.36 (1.57) 4.16 (1.14) 1.86 (1.17) 1.77 (1.17) 4. I may be permanently laid off from my job in the near future. 2.49 (1.55) 2.09 (1.23) 1.70 (1.12) 2.02 (1.22) 1.93 (1.34) 2.32 (1.46) 4.23 (1.24) 5. I know that my job is not compl etely secure from termination. 2.64 (1.40) 2.25 (1.43) 1.64 (1.16) 1.93 (1.22) 2.02 (1.36) 2.05 (1.10) 4.05 (1.45) 6. Supervisors in this organization often carry out the policies in a self serving manner. 2.02 (1.17) 2.80 (1.47) 1.70 (0.93) 4.32 (1 .14) 2.25 (1.53) 2.52 (1.32) 1.95 (1.36) 7. I often do not have the resources that I need to adequately complete tasks. 2.32 (1.46) 2.05 (1.14) 4.11 (1.37) 1.75 (1.04) 1.95 (1.29) 1.65 (1.00) 1.79 (1.21) 8. I believe that I may be at risk of being t erminated from my job. 2.36 (1.43) 2.07 (1.23) 1.84 (1.27) 1.70 (1.00) 1.86 (1.29) 2.07 (1.23) 4.20 (1.29) 9. I work with two or more groups who ask very different things of me. 3.10 (1.66) a 3.62 (1.53) a 1.83 (1.23) 2.02 (1.24) 2.00 (1.29) 2.73 (1.53 ) 1.80 (1.27) 10. I am often not able to do work because I do not have the proper equipment, tools, materials, or supplies. 2.16 (1.36) 2.14 (1.41) 4.18 (1.32) 1.77 (1.23) 1.91 (1.36) 1.64 (1.08) 1.80 (1.15) 11. This job requires the completion of exces sive paperwork or computerwork. 2.07 (1.25) 1.93 (1.26) 1.86 (1.29) 1.95 (1.35) 3.80 (1.47) 2.00 (1.31) 1.66 (1.10) 12. There are many overly restrictive rules and regulations in this job. 2.23 (1.27) 2.26 (1.40) 1.77 (1.11) 1.91 (1.19) 4.18 (1.28) 2.02 (1.30) 1.82 (1.23) 13. In this job, it is often favoritism and not merit that gets employees ahead. 2.00 (1.22) 2.30 (1.30) 1.67 (1.07) 3.76 (1.27) 1.86 (1.30) 3.07 (1.42) 2.09 (1.38) 14. People in this company are more concerned with promoting themsel ves than the organization. 2.05 (1.22) 2.14 (1.23) 1.63 (0.93) 4.36 (1.04) 2.05 (1.41) 2.50 (1.32) 2.02 (1.39) 15. In my organization, workers get ahead by engaging in political activities, not by performing work. 2.43 (1.45) 2.18 (1.42) 1.68 (1.14) 3.68 (1.46) 1.89 (1.32) 2.98 (1.50) 1.86 (1.13) 16. My manager does not provide a clear idea of what I should be doing at work. 4.00 (1.31) 3.18 (1.56) 1.93 (1.15) 1.91 (1.34) 2.00 (1.41) 2.11 (1.20) 1.89 (1.26) 17. I often have to complete unnecessary form s during my work. 2.14 (1.41) 2.02 (1.35) 1.62 (0.88) 1.72 (1.12) 4.14 (1.34) 1.70 (1.15) 1.68 (1.20) 18. I am often given incompatible requests by supervisors. 2.93 (1.55) 4.00 (1.36) 1.67 (1.07) 1.74 (1.20) 1.98 (1.16) 2.21 (1.30) 1.98 (1.37) 19. My s upervisor does not provide clear expectations of what I should be doing in my job. 3.79 (1.58) 3.09 (1.72) 1.95 (1.20) 2.18 (1.50) 1.98 (1.47) 2.02 (1.17) 2.02 (1.34) 20. I am often given conflicting assignments. 3.25 (1.63) 4.05 (1.35) 1.84 (1.24) 1.86 (1.32) 2.05 (1.29) 2.18 (1.42) 2.00 (1.35) 21. I believe that I may be at risk of involuntarily losing my job. 1.98 (1.15) 1.77 (1.01) 1.89 (1.32) 2.11 (1.40) 1.91 (1.39) 1.98 (1.25) 4.27 (1.21)

PAGE 144

144 Table 3 5. Continued. Item Role ambiguity Role confl ict Resource inadequacies Organizational politics Administrative hassles Interpersonal conflict Job insecurity 22. I have conflicts with other coworkers regarding political views, personality, and/or interpersonal style. 2.02 (1.37) 1.93 (1.28) 1.49 (0.86 ) 2.00 (1.28) 1.75 (1.12) 4.32 (1.07) 1.86 (1.23) 23. It is likely that I will be laid off or fired from this position sometime in the near future. 2.14 (1.44) 2.00 (1.26) 1.64 (1.12) 1.82 (1.11) 1.93 (1.28) 2.20 (1.44) 4.41 (1.13) 24. I often experience administrative hassles while trying to complete work. 2.39 (1.45) 2.36 (1.33) 1.84 (1.12) 2.05 (1.33) 3.75 (1.45) 2.11 (1.37) 1.77 (1.12) 25. I am often asked to do things that are likely to be accepted by one supervisor and not accepted by another sup ervisor. 2.95 (1.61) 4.34 (1.06) 1.70 (0.98) 2.05 (1.16) 2.14 (1.30) 2.14 (1.36) 2.09 (1.38) 26. I do not know how I will be evaluated for a raise or promotion in my position. 3.73 (1.17) 2.27 (1.28) 1.58 (0.91) 2.05 (1.24) 2.16 (1.48) 2.30 (1.46) 2.07 (1.34) 27. I clash with people at my work over personal differences. 2.09 (1.34) 2.05 (1.36) 1.80 (1.21) 2.14 (1.36) 1.86 (1.30) 4.20 (1.27) 1.80 (1.21) 28. I have not been provided with clear, planned goals or objectives in my job. 3.91 (1.44) 2.89 (1. 50) 1.93 (1.26) 1.89 (1.17) 2.07 (1.30) 1.84 (1.22) 2.02 (1.32) 29. I disagree with people at work because we have different backgrounds, personalities, or political views. 2.00 (1.33) 2.05 (1.38) 1.73 (1.21) 2.00 (1.28) 1.77 (1.26) 4.30 (1.25) 1.77 (1.2 4) 30. My workplace does not provide the resources necessary to do my job. 2.39 (1.43) 2.20 (1.25) 4.11 (1.35) 1.65 (1.04) 2.16 (1.52) 1.73 (1.15) 1.70 (1.17) 31. I often get into conflicts with people at work over personal differences. 1.98 (1.27) 1.6 7 (1.06) 1.77 (1.21) 2.02 (1.34) 1.70 (1.17) 4.12 (1.40) 1.61 (1.04) 32. It is difficult or impossible to complete my job because we do not have the proper equipment, tools, materials, or supplies. 2.27 (1.39) 1.82 (0.97) 4.14 (1.42) 1.47 (0.85) 2.09 (1. 44) 1.75 (1.18) 1.91 (1.24) 33. I have arguments with people at my workplace about personal issues. 1.77 (1.12) 1.82 (1.33) 1.82 (1.35) 2.00 (1.29) 1.75 (1.22) 4.30 (1.15) 2.00 (1.40) 34. This organization is characterized by self serving behaviors, not by organization policies. 2.02 (1.25) 2.07 (1.19) 1.68 (1.09) 4.23 (1.26) 1.82 (1.13) 2.11 (1.33) 1.77 (1.12) 35. It is unclear what my supervisor expects of me in this job. 4.20 (1.15) 3.23 (1.61) 1.86 (1.23) 1.89 (1.22) 1.93 (1.30) 2.05 (1.29) 2.14 (1 .29) Note: Note: Scale anchors ranged from 1 ( definition ) to 5 ( ). The first number in each cell represents the mean; the number in parentheses is the standard deviation. Intended factors for each item are indicated in bold. All t tests comparing ratings on the intended factor versus all other factors were signi ficantly different at p < .01 , unless otherwise noted with the common superscript a (p > .10).

PAGE 145

145 Table 3 6. Demographic characterist ics of the second undergraduate content validation sample. Characteristic M (SD) N % of sample Subsample A Age (years) 21.18 (1.23) 86 100 Work experience (years) 4.01 (2.32) 86 100 Gender Female 35 41 Male 49 57 Non resp ondent 2 2 English 1 st Language Yes 73 85 No 11 13 Non respondent 2 2 Subsample B Age (years) 21.13 (1.11) 84 100 Work experience (years) 3.44 (2.14) 84 100 Gender Female 38 45 Male 46 55 Non respondent 0 0 English 1 st Language Yes 76 90 No 8 10 Non respondent 0 0 Subsample C Age (years) 21.16 (1.07) 86 100 Work experience (years) 3.68 (2.22) 86 100 Gender Female 39 45 Male 47 54 Non respondent 1 1 English 1 st Language Yes 75 86 No 11 13 Non respondent 1 1 Note: Total n = 256

PAGE 146

146 Table 3 7. Means and standard deviations from the second content validation of the challe nge stressor items. Item Workload Job responsibility Job complexity Work pace Subsample A 1. I am responsible for the performance of myself and my coworkers. 1.66 (1.23) a 4.59 (0.86) b 1.78 (1.06) c 1.66 (1.11) d 2. Tasks on my job use a variety of different skills and abilities. 2.45 (1.49) 1.67 (1.14) 4.63 (0.90) 1.90 (1.24) 3. I have to work at a rapid pace to complete my job. 2.57 (1.51) 1.68 (1.09) 1.78 (1.13) 4.70 (0.91) 4. My job requires a complex set of behaviors and skills. 2.15 (1 .35) 1.94 (1.29) 4.38 (1.12) 1.65 (1.04) 5. To complete my work on time, I must work quickly. 2.79 (1.47) 1.80 (1.56) 1.80 (1.15) 4.65 (0.89) 6. My job requires me to be accountable for my work and the work of others. 1.69 (1.19) 4.59 (1.01) 1.95 (1. 35) 1.49 (1.00) Subsample B 1. I am responsible for the performance of myself and my coworkers. 2.05 (1.17) a 4.49 (0.95) b 1.87 (1.07) c 1.83 (1.16) d 2. I experience time pressure to complete work on my job. 3.04 (1.56) 2.36 (1.39) 2.25 (1.30) 4.44 (1 .06) 3. In this job, I am accountable for the work of other people. 1.90 (1.35) 4.37 (0.98) 2.01 (1.19) 1.81 (1.19) 4. This job requires that I complete a lot of work. 4.56 (0.92) 2.24 (1.43) 2.37 (1.29) 3.04 (1.43) 5. I am required to use a varie ty of different skills on my job. 2.29 (1.47) 1.87 (1.25) 4.63 (0.92) 1.82 (1.26) 6. My work load on this job is large. 4.51 (1.00) 2.12 (1.34) 2.14 (1.33) 2.46 (1.49)

PAGE 147

147 Table 3 7. Continued. Item Workload Job responsibility Job c omplexity Work pace Subsample C 1. I am responsible for the performance of myself and my coworkers. 1.78 (1.11) a 4.54 (0.98) b 1.82 (1.20) c 1.98 (1.31) d 2. My job requires me to use a broad set of skills and abilities. 2.53 (1.52) 2.16 (1.40) 4.52 (1.00) 1.81 (1 .04) 3. My job requires that I work at a fast tempo. 2.76 (1.62) 1.92 (1.36) 1.72 (1.00) 4.63 (0.90) 4. This job requires that I be responsible for the work of myself and others. 2.85 (1.24) 4.56 (0.99) 2.01 (1.25) 1.63 (1.07) 5. I have a lot of assignments to complete on my job. 4.67 (0.90) 2.14 (1.28) 2.22 (1.45) 2.87 (1.57) 6. I have to complete a lot of tasks each day at work. 4.29 (1.22) 2.12 (1.21) 2.42 (1.41) 3.24 (1.62) Note: Scale anchors ranged from 1 ( e definition ) to 5 ( ). The first number in each cell represents the mean; the number in parentheses is the standard deviation. Intended factors for each item are indicated in bold. All t tests comparing ratings on th e intended factor versus all other factors were sig nif icantly different at p < .01. Similar letter superscripts a through d indicate that t tests comparing ratings for items across sub samples did not differ from each other (p > .10).

PAGE 148

148 Table 3 8. Summar y of means and standard deviations from the sec ond content validation of the h indrance stressor items. Item Role ambiguity Role conflict Inadequate resources Organizational politics Administrative hassles Interpersonal conflict Job I nsecurity Subsample A 1. I believe that I may be at risk of involuntarily losing my job. 2.83 (1.47) c 2.66 (1.46) d 1.81 (1.23) e 1.83 (1.10) f 2.13 (1.38) g 2.14 (1.32) h 4.53 (1.04 ) i 2. People in this company are more concerned with promoting themselves than the organizati on. 1.61 (1.16) 1.99 (1.25) 1.28 (0.68) 4.50 (1.11) 1.54 (0.93) 2.42 (1.46) 1.69 (1.14) 3. I often have to complete unnecessary forms during my work. 1.95 (1.30) 2.04 (1.31) 2.09 (1.44) 1.43 (0.86) 4.35 (1.24) 1.69 (1.21) 1.49 (0.96) 4. My manager do es not provide a clear idea of what I should be doing at work. 4.12 (1.29) a 3.73 (1.50) a 1.84 (1.32) 1.66 (1.19) 1.85 (1.28) 1.96 (1.37) 1.77 (1.28) 5. In my organization, workers get ahead by engaging in political activities, not by performing work. 2 .00 (1.28) 2.02 (1.28) 1.52 (1.14) 3.96 (1.40) 1.69 (1.13) 2.76 (1.48) 1.65 (1.20) 6. I am often given incompatible requests by supervisors. 2.40 (1.37) 4.25 (1.31) 1.84 (1.20) 1.55 (1.05) 2.13 (1.47) 2.05 (1.28) 1.49 (0.86) 7. I believe that I may b e at risk of involuntarily losing my job. 2.70 (1.55) 2.42 (1.40) 1.66 (1.01) 1.67 (1.06) 1.85 (1.21) 1.89 (1.24) 4.16 (1.40) 8. I often get into conflicts with people at work over personal differences. 1.59 (1.10) 1.87 (1.20) 1.35 (0.82) 2.19 (1.29) 1. 51 (0.99) 4.49 (0.96) 1.73 (1.19) 9. This organization is characterized by self serving behaviors, not by organization policies. 1.80 (1.16) 2.11 (1.33) 1.60 (1.16) 4.08 (1.41) 1.77 (1.26) 2.17 (1.32) 1.67 (1.07) 10. I have arguments with people at my workplace about personal issues. 1.49 (0.92) 1.75 (1.17) 1.33 (0.83) 2.04 (1.30) 1.53 (1.07) 4.24 (1.27) 1.56 (0.96) 11. I am often unclear about what is expected of me on this job. 4.19 (1.22) 3.34 (1.53) 1.60 (1.10) 1.64 (1.14) 2.14 (1.43) 1.70 (1.21) 1 .73 (1.19) Subsample B 1. I believe that I may be at risk of involuntarily losing my job. 2.81 (1.62) c 2.54 (1.36) d 1.78 (1.16) e 1.90 (1.27) f 2.07 (1.30) g 2.25 (1.41) h 4.57 (1.06) i 2. I am often given conflicting assignments on my job. 3.48 (1.55) b 3.48 (1.53) b 1.65 (1.12) 1.71 (1.08) 2.41 (1.47) 1.90 (1.30) 1.86 (1.36) 3. I am often asked to do things that are likely to be accepted by one supervisor and not accepted by another supervisor. 3.13 (1.55) 4.33 (1.17) 1.47 (0.90) 2.11 (1.46) 1.91 (1.30) 2.60 (1.49) 1.99 (1.37) 4. It is likely that I will be laid off or fired from this position sometime in the near future. 2.79 (1.60) 2.23 (1.36) 1.83 (1.20) 2.08 (1.36) 1.90 (1.25) 2.02 (1.33) 4.07 (1.37) 5. I often experience administrative hassles while tr ying to complete work. 2.20 (1.36) 2.92 (1.62) 2.19 (1.37) 1.76 (1.12) 3.94 (1.42) 1.91 (1.10) 1.54 (1.03) 6. I have conflicts with other coworkers regarding political views, personality, and/or interpersonal style. 1.82 (1.25) 2.18 (1.48) 1.43 (0.97) 2 .59 (1.53) 1.73 (1.20) 4.39 (1.24) 1.66 (1.19) 7. I do not know how I will be evaluated for a raise or promotion in my position. 3.34 (1.56) 2.72 (1.53) 1.60 (1.04) 2.12 (1.43) 1.83 (1.19) 1.83 (1.22) 1.84 (1.32) 8. I have not been provided with clear, planned goals or objectives in my job. 4.04 (1.35) 2.84 (1.60) 2.06 (1.33) 2.02 (1.41) 1.96 (1.28) 1.60 (1.11) 1.98 (1.29)

PAGE 149

149 Table 3 8. Continued. Item Role ambiguity Role conflict Inadequate resources Organizational P olitics Administrative hassles Interp ersonal c onflict Job insecurity 9. I disagree with people at work because we have different backgrounds, personalities, or political views. 1.76 (1.20) 2.22 (1.50) 1.54 (1.12) 2.40 (1.40) 1.63 (1.09) 4.24 (1.38) 1.56 (1.01) 10. My workplace does not pro vide the resources necessary to do my job. 2.12 (1.39) 1.74 (1.18) 4.53 (1.15) 1.74 (1.14) 1.86 (1.26) 1.52 (1.06) 1.57 (1.10) Subsample C 1. I believe that I may be at risk of involuntarily losing my job. 2.94 (1.51) c 2.76 (1.28) d 1.94 (1.24) e 2.18 (1.42) f 2.01 (1.35) g 2.36 (1.44) h 4.43 (1.25) i 2. I cannot complete some of my tasks because I am not given adequate equipment, tools, materials, or supplies. 2.00 (1.25) 2.30 (1.42) 4.52 (1.12)) 1.47 (1.01) 1.73 (1.09) 1.45 (0.90) 1.78 (1.31) 3. I often receive conflicting requests from a supervisor (or supervisors). 3.12 (1.55) 4.42 (1.10) 1.55 (1.07) 2.07 (1.38) 1.99 (1.23) 2.61 (1.52) 1.93 (1.27) 4. order to complete my job. 2.61 (1.48) 2.13 (1.25) 1.75 (1.15) 2.12 (1.43) 4.04 (1.47) 1.91 (1.15) 1.85 (1.22) 5. I may be permanently laid off from my job in the near future. 2.53 (1.52) 2.24 (1.35) 1.61 (1.03) 1.99 (1.38) 1.92 (1.28) 1.96 (1.30) 4.33 (1.82) 6. Supervisors in this organi zation often carry out the policies in a self serving manner. 1.88 (1.20) 2.84 (1.49) 1.62 (1.13) 4.16 (1.34) 1.95 (1.37) 2.44 (1.49) 1.83 (1.33) 7. I often do not have the resources that I need to adequately complete tasks. 2.19 (1.40) 1.97 (1.33) 4.47 (1.09) 1.57 (1.15) 1.84 (1.31) 1.56 (1.02) 1.89 (1.34) 8. I believe that I may be at risk of being terminated from my job. 2.85 (1.45) 2.38 (1.41) 1.76 (1.29) 1.86 (1.22) 1.61 (1.08) 2.08 (1.20) 4.37 (1.27) 9. I am often not able to do work because I do not have the proper equipment, tools, materials, or supplies. 1.95 (1.23) 1.99 (1.38) 4.41 (1.20) 1.55 (1.09) 1.79 (1.19) 1.70 (1.23) 1.98 (1.24) 10. This job requires the completion of unnecessary paperwork or computer work. 2.08 (1.30) 2.01 (1.27) 1.97 (1.36) 1.55 (1.33) 4.13 (1.37) 1.85 (1.24) 1.61 (1.03) 11. There are many overly restrictive rules and regulations in this job. 2.34 (1.44) 2.38 (1.39) 1.55 (1.03) 1.72 (1.16) 4.36 (1.25) 1.95 (1.39) 1.87 (1.36) Note: Scale anchors ranged from 1 ( definition ) to 5 ( ). The first number in each cell represents the mean; the number in parentheses is the standard deviation. Intended factors for each item are indicated in bold. All t tes ts comparing ratings on the intended factor versus all other factors were sig nificantly different at p < .01, unless otherwise noted with the common superscript a or b (p > .10). Similar letter superscripts c through i i n dicate that t tests comparing rat ings for items across sub samples did not differ from each other (p > .10).

PAGE 150

150 Table 3 9. Demographic characteristics of working MBA sample (Study 1) Variable M (SD) N % of sample Age (years) 28.96 (4.46) Gender Female 49 27.5 Male 128 71.9 Organizational tenure (years) 3.70 (2.56) Job tenure (years) 2.92 (2.60) Tenure with supervisor (years) 1.94 (1.75) Cognitive ability (GMAT Score) 623.27 (56.87) Race Caucasian 139 78.1 African Amer ican 8 4.5 Hispanic 3 1.7 Asian/Pacific Islander 18 10.1 Other 6 3.4 Marital status Single 59 33.1 Living with partner 12 6.7 Married 103 57.9 Separated/Divorced 2 1.1 Occupation Manager/Supervisor 41 23.0 Engineer/Scientist 41 23.0 Banker/Accountant/Financier 29 16.3 Professional 16 9.0 Information Technician 10 5.6 Sales/Marketing 5 2.8 Oth er 12 6.7 Note: Percentages are out of total valid sample of 178 participants; some totals do not sum to 100% because of missing data.

PAGE 151

151 Table 3 10. Summary of model fit for the challenge and hindrance stressor construct CFAs. Model 2 (df) CFI TLI S RMR RMSEA A. First order challenge stressor CFA 182.31** (98) .96 .96 .05 .07 B. Revised first order challenge stressor CFA 51.38 (48) 1.00 1.00 .04 .02 C. First order hindrance stressor CFA 540.04** (303) .94 .93 .06 .07 D. Revised firs t order hindrance stressor CFA 233.04** (168) .98 .97 .04 .05 E. First order challenge/hindrance stressor CFA 549.15** (440) .98 .97 .04 .04 F. Second order formative challenge/hindrance stressor CFA 744.09** (563) .96 .95 .07 .05 G. Seco nd order reflective challenge/hindrance stressor CFA 619.28** (483) .97 .97 .06 .04 Note: CFI = comparative fit index; TLI = Tucker Lewis Index; SRMR = root mean standardized residual; and RMSEA = root mean squared error of approximation. N = 178 for all statistics in all models. ** p < .01.

PAGE 152

152 Table 3 11. Results of first order CFA of the challenge stressor construct. Item M (SD) Unstandardized factor loading t value Standardized factor loading Workload My job requires that I complete a lot of work. 5.93 (1.15) 1.00 -.94 My workload on this job is large. 5.72 (1.28) 1.13 22.33 .95 I have a lot of assignments to complete on my job. 5.58 (1.36) 0.93 13.15 .74 Job responsibility This job requires that I be responsible for the productivity of myself and others. 5.62 (1.41) 1.00 -.84 I am responsible for the performance of myself and others. 5.44 (1.54) 1.14 11.44 .87 In this job, I am accountable for the work of other people. 4.57 (2.04) 1.26 10.14 .73 Job complexit y Tasks on my job use a variety of different skills and abilities. 5.95 (1.23) 1.00 -.79 My job requires me to use a broad set of skills and abilities. 5.90 (1.13) 1.06 13.82 .91 I am required to use a variety of different skills on my j ob. 5.83 (1.11) 1.08 14.08 .94 Work pace To complete my work on time, I must work quickly. 5.35 (1.42) 1.00 -.87 I have to work at a rapid pace to complete my job. 5.16 (1.38) 1.06 17.96 .94 My job requires that I work at a fast tempo. 5.16 (1.41) 1.05 17.10 .91 Note: N = 178. All t values are significant at p < .01.

PAGE 153

153 Table 3 12. Results of first order CFA of the hindrance stressor construct. Item M (SD) Unstandardized Factor Loading t value Standardized Factor Loading Role ambig uity My manager does not provide a clear idea of what I should be doing at work. 3.45 (1.86) 1.00 -.75 I have not been provided with clear, planned goals or objectives in my job. 3.28 (1.92) 1.17 11.15 .85 I am often unclear about what is exp ected of me on this job. 3.03 (1.75) 1.09 11.31 .87 Resource inadequacies I cannot complete my tasks because I am not given adequate equipment, tools, materials, or supplies. 2.53 (1.64) 1.00 -.90 I am often not able to do my work because I do not have the proper equipment, tools, materials, or supplies. 2.51 (1.51) .93 17.32 .91 My workplace does not provide the resources necessary to do my job. 2.74 (1.69) .93 14.28 .82 Organizational politics People in this company are more concerned with promoting themselves than the organization. 3.92 (1.92) 1.00 -.81 In my organization, workers get ahead by engaging in political activities, not by performing work. 4.02 (1.99) 1.17 15.00 .92 This organization is characterized by self serving and/or political behaviors, not organizational policies. 3.88 (1.96) 1.19 15.44 .94 Administrative hassles This job requires the completion of unnecessary paper or computer work. 4.23 (1.92) 1.00 -.88 There are many overly restricti ve rules and regulations in this job. 4.00 (2.05) 1.03 15.29 .85 I often have to complete unnecessary forms during my work. 3.78 (1.97) 1.11 18.12 .95 Interpersonal conflict I disagree with people at work because we have different backgrounds, pe rsonalities, or political views. 2.54 (1.64) 1.00 -.74 I often get into conflicts with people at work over personal differences. 1.94 (1.30) 1.02 11.69 .95

PAGE 154

154 Table 3 12. Continued Item M (SD) Unstandardized Factor Loading t value Standardized Fa ctor Loading I have arguments with people at work about personal issues. 1.68 (1.05) .72 11.13 .83 Job insecurity I may be permanently laid off from my job in the near future. 1.88 (1.50) 1.00 -.84 I believe that I may be at risk of inv oluntarily losing my job. 2.03 (1.50) 1.07 16.68 .92 It is likely that I will be laid or fired from this position sometime in the near future. 1.89 (1.41) 1.06 16.99 .95 Role conflict I often receive conflicting requests from a supervisor (or supervisors). 3.24 (2.07) 1.00 -.93 I am often given incompatible requests by supervisors. 3.08 (1.98) .94 18.63 .91 I am often asked to do things that are likely to be accepted by one supervisor and not accepted by another supervisor. 3.2 1 (2.09) .87 14.45 .80 Note: N = 178. All t values are significant at p < .01.

PAGE 155

155 Table 3 13. Results of CFAs modeling challenge and hindrance stressors as second order constructs. Measures Formative indicator m odel Reflective indictor m odel Unstand ardized parameter est. t value Standardized parameter est. Unstandardized parameter est. t value Standardized parameter est. Challenge stressor construct Work load .03 .22 .02 2.38 4.04** .86 Job responsibility .09 1.02 .09 1.04 3.00** .34 Job complexity .40 3.50** .31 1.00 -.40 Work pace .22 2.03* .22 2.31 4.22** .74 Reflective challenge stressor indicator #1 1.00 -.84 ---Reflective challenge stressor indicator #2 .98 7.30** .88 --Hindrance stressor construct Role ambiguity .06 .51 .04 .86 6.17** .67 Role conflict .19 2.36* .20 1.26 7.10** .72 Interpersonal conflict .10 .91 .06 .56 5.06** .51 Resource inadequacies .33 3. 18 ** .27 1.00 -.74 Administrative hassles .02 .21 .02 .71 4.95** .46 Organizational politics .37 3.77** .32 .95 6.45** .66 Job insecurity .19 2.10* .13 .23 2.22* .20 Reflective hindrance stressor indicator #1 1.00 -.95 ---Reflective hindrance stressor indicator #2 .94 17.25** .92 ---Note: N = 178 for all models. ** p < .01. * p < .05.

PAGE 156

156 Table 3 14. Means, standard deviations, reliabilities, and intercorrelations of the dimensions of the chal lenge and hindrance stressor constructs. Construct M SD vc WL JR JC WP RA IC AH RI OP JI RC Work load 5.74 1.16 .78 .90 Job responsibility 5.21 1.46 .67 .24 .83 Job complexity 5.89 1.06 .78 .38 .16 .91 Work pace 5.22 1.31 .82 .63 .35 .22 .93 Role ambiguity 3.25 1.64 .68 .01 .03 .05 .19 .87 Interpersonal conflict 2.05 1.19 .71 .03 .07 .12 .12 . 38 .85 Administrative hassles 4.00 1.84 .80 .01 .01 .07 .06 .17 .27 .92 Resource inadequacies 2.59 1.48 .77 .02 .08 .12 .10 .5 0 .41 .32 .91 Organizational politics 3.94 1.82 .80 .01 .00 .01 .00 .43 .28 .51 .46 .92 Job insecurity 1.93 1.38 .82 .01 .07 .14 .03 .08 .09 .01 .13 .13 .93 Role conflict 3.18 1.89 .78 .09 .02 .01 .04 .52 .33 .27 .54 .45 . 24 .91 = work pace; RA = role ambiguity; IC = interpersonal conflict; AH = administrative hassles; RI = resource inadequacies; OP = organizational politics; JI = job insecurity; RC = role conflict. r significant at p < .05 .

PAGE 157

157 Table 3 15. Correlations between the dimensions of the challenge and hindrance stressor construct and the global reflective indicator s rk stress measures. Stressor Dimension Global challenge stressor indicator #1 Global challenge stressor indicator #2 Global hindrance stressor indicator #1 Global hindrance stressor indicator #2 Challenge related work stress scale Hindrance related w ork stress scale Challenge stressors Work load .25 ** .27 ** .07 .07 .58 ** .00 Job responsibility .23 ** .14 .01 .02 .11 .07 Job complexity .27 ** .35 ** .08 .12 .12 .06 Work pace .29 ** .27 ** .11 .05 .59 ** .02 Hindrance stressors Role ambiguity .21 ** .29 ** .33 ** .32 ** .13 .38 ** Role conflict .16 * .24 ** .50 ** .41 ** .12 .52 ** Resource inadequacies .13 .24 ** .51 ** .49 ** .16 * .36 ** Interpersonal conflict .09 .20 ** .28 ** .30 ** .04 .29 ** Organizational politics .15 .26 ** .52 ** .53 ** .04 .60 ** Administrative hassles .26 ** .33 ** .30 ** .34 ** .03 .44 ** Job insecurity .01 .07 .2 5 ** .23 ** .01 .30 ** Note: N = 178. * * p < .01. * p < .05.

PAGE 158

158 Table 3 16. Correlations among the dimensions of the challenge and hindrance stressor constructs and employee strains, emotions, job satisfaction, and motivation. Stressor dimension EE FR CE BE TE HE JS MO Challenge stressors Work load .47 ** .30 ** .21 * .10 .35 ** .13 .02 .25 ** Job responsibility .25 * .18 * .21 * .05 .12 .09 .03 .17 Job complexity .01 .01 .32 ** .26 ** .03 .11 .27 ** .28 ** Work pace .38 ** .30 ** .17 .09 .29 ** .14 .01 .27 ** Hindrance stressors Role ambiguity .31 ** .41 ** .40 ** .32 ** .26 ** .39 ** .42 ** .23 ** Role conflict .37 ** .48 ** .30 ** .31 ** .27 ** .43 ** .48 ** .22 * Resource inadequacies .38 ** .38 ** .27 ** .41 ** .28 ** .48 ** .55 ** .28 ** Interpersonal conflict .31 ** .32 ** .17 .27 ** .27 ** .40 ** .23 ** .11 Organizational politics .40 ** .39 ** .38 ** .43 ** .20 * .54 ** .51 ** .30 ** Administrative hassles .37 ** .32 ** .20 * .35 ** .13 .33 ** .31 ** .24 ** Job insecurity .18 * .09 .00 .10 .27 ** .15 .24 ** .24 ** Note: EE = emotional exhaustion; FR =frustration; CE = challenge emotion; BE = benefit emotion; TE = threat emotion; HE = harm emotion; JS = job satisfaction; MO = motivation. N = 178. ** p < .01. * p < .05.

PAGE 159

159 Table 3 17. Results of structural models with challenge and hindrance stressors modeled as second order constructs (control ling for negative affectivity). Construct EE FR CE BE TE HE JS MO TP OCB CWB Formative model Challenge stressors .20** .14 .73** .54** .34** .06 .43** .58** .12 .20 .14 Controlling for NA .13 .16* .84** .79** .24** .20** .71** .68** .15 .23 .18 Hindrance stressors .75** .65** .33** .58** .46** .75** .80** .40** .05 .17 .13 Controlling for NA .38** .23** .14* .21** .02 .26** .43** .25** .04 .10 .01 R 2 (model without NA) .60 .46 .63 .63 .32 .57 .82 .48 .02 .07 .03 Reflective model Challenge stressors .39** .30** .55** .37** .44** .12 .31** .55** .01 .09 .14 Controlling for NA .39* * .31** .48** .34** .38** .05 .31** .53** .01 .07 .20 Hindrance stressors .67** .62** .68** .82** .29** .72** .94** .64** .12 .26* .40** Controlling for NA .65** .58** .73** .83** .03 .59** .95** .65** .04 .23 .35* R 2 (model without NA) .70 .54 .62 .69 .33 .56 .87 .57 .01 .07 .17 Note: EE = emotional exhaustion; FR =frustration; CE = challenge emotion; BE = benefit emotion; TE = threat emotion; HE = harm emotion; JS = j o b satisfaction; MO = motivation; TP = task performance; OCB = organizational citizenship behavior. CWB = counterproductive work behavior. NA = negative affectivity. N = 178 for models with frustration, emotional exhaustion, all em otion constructs, job sat isfaction, and motivation; N = 101 for models with task performance, OCB, and CWB. All parameters estimates reported in this table are standardized; significance values were computed from the unstandardized parameter estimate for eac h relationship. ** p < .01. * p < .05.

PAGE 160

160 Table 3 18. Results of structural models examining relationships between the dimensions of the challenge and hindrance stressor constructs and employee strains, emotions, job satisfaction, motivation, and performance criteria. Stressor d imension EE FR CE BE TE HE JS MO TP OCB CWB Challenge stressors Workload .55** .29** .10 .19 .35** .24* .20* .06 .05 .05 .15 Job responsibility .13 .10 .10 .00 .00 .05 .02 .07 .02 .12 .13 J ob complexity .20* .14 .31** .28** .04 .17* .31** .21** .05 .06 .04 Work pace .02 .05 .22* .23* .03 .05 .13 .30** .04 .02 .32 Hindrance stressors Role ambiguity .16 .19 .44** .27** .07 .13 .26** .29** .02 .26 .14 Role conflict .01 .22* .01 .10 .02 .03 .05 .11 .03 .02 .09 Resource inadequacies .10 .02 .01 .19* .10 .14 .29** .16 .13 .10 .01 Interpersonal conflict .08 .07 .06 .01 .16 .15 .11 .11 .06 .07 .11 Organizational politics . 15 .12 .25* .23* .04 .37** .27** .14 .28 .42** .27 Administrative hassles .21 .15** .03 .13 .03 .02 .05 .11 .19 .23 .04 Job insecurity .10 .02 .11 .02 .22* .04 .10 .19* .17 .35** .12 R 2 .58 .46 .43 .41 .29 .47 .58 .36 .08 .27 .23 Note: EE = emotional exhaustion; FR =frustration; CE = challenge emotion; BE = benefit emotion; TE = thr eat emotion; HE = harm emotion; JS = job satisfaction; MO = motivation; TP = task performance; OCB = organizational citizenship behavior. CWB = counterproductive work behavior. NA = negative affectivity. Fit indices for model with non performance criteria: 2 (2067, N = 178) 2 (1864, N = 178) = 3120.23; CFI = .75, TLI = .73; SRMR = .11; RMSEA = .08. All parameter estimates are standardized; the s ignificance values were computed from the unstandardized parameter estimates. ** p < .01. * p < .05.

PAGE 161

161 CHAPTER 4 DISCUSSION OF STUDY 1 Research indicates that work related stress is an important concern for researchers and managers alike. Unfortunately, o ur ability to accurately assess work stressors has been hindered by the conceptual and methodological limitations of previously reported measures of work stressors. In addition, recent research in this domain (Boswell et al. 2004; Cavanaugh et al., 2000 ; L ePine et al., 2005; Podsakoff et al., 2007 ) indicates that it is important to distinguish work stressors that tend to be appraised as obstacles to employee development and task accomplishment (i.e., hindrance stressors) from those that tend to be appraised as promoting employee growth and task achievement (i.e., challenge stressors). Therefore, t he purpose of Study 1 was to develop and validate new measures of challenge and hindrance stressors that addre ss l imitations of previous measures used in the work s tress literature. Assessment of Challenge and Hindrance Stressor Measure Construct Validity Consistent with this purpose, t he results of Study 1 provide d preliminary support for the construct validity of the newly developed measures in four ways . First, data gathered from two undergraduate samples provides support for the content validity of the measures. Specifically, items developed to assess the stressor dimensions were generally rated as tapping the content domain of their intended construct , and were not found to be contaminated by content tapping the domains of other constructs. Second, the findings of Study 1 indicate that the stressor measures demonstrated good convergent and discriminant validity. For example, results from the confirmatory facto r analyses indicated that all of the stressor items loaded strongly on their intended dimensions, and did not cross load on other stressor dimensions. These results were bolstered by the findings from correlational analyses which indicated that: (a) the di mensions of the challenge stressor construct

PAGE 162

162 were moderately correlated with each other; (b) the dimensions of the hindrance stressor construct were m oderately correlated with each other; and (c) that the dimensions of the challenge and hindrance stressor constructs did not correlate strongly with each other. In addition , with a few exceptions, the challenge stressor dimensions were positively related to the g lobal reflective related work s tress scale, and were not related to the global hindrance stressor measures or Cavanaugh et related work stress scale. Conversely, all of the hindrance stressor dimensions wer e positively related to the global reflective hindrance stressor measures and the hindrance related work stress scale ; and were negatively related to the global reflective challenge stressor measures and the challenge related work stress scale. T he only exceptions to these findings were that neither the job complexity nor the job responsibility dimensions were found to correlate positively with the challenge related work stress scale. It is perhaps not surprising that the job complexity dimension was not related to the (2000) hindrance related work st ress scale, because this work stress scale does not contain items measuring job complexity. However, another potential explanation for the lack of relationships between these measures is that they assess somewhat different aspects of the stress process. Sp ecifically, the measures developed by Cavanaugh et al. assess the amount of stress produced from work demands whereas the newly developed scales assess the extent to which employees experience the demands. Nevertheless, when taken together, the results fro m this study provide relatively strong support for the convergent and discriminant validity of the challenge an d hindrance stressor dimensions. Third, all of the dimensions of the newly developed scales demonstrated relatively high reliabilities (ranging

PAGE 163

163 newly developed scales. In addition, the vc indices for the stressor dimensions indicated that the vast majority of the variance in these items was accounted for by their underlying dimensions (average vc = .76) , and not measurement error. Thus, although additional information should be g athered on the reliability of these scales over time (e.g., test retest reliability), the preliminary evidence suggests that the stressor dimensions possess good internal consistency reliability. Finally, the results of Study 1 also provide preliminary support for the nomological (or criterion related ) validity of both the challenge and hindrance stressor dimensions. With the possible exception of the job insecurity dimension, virtually all of the expected relationships between the hindrance stressor dimensions and the criterion variables were supp orted. Specificall y, the dimensions of the hindrance stressor construct were generally found to be positively related to frustration, emotional exhaustion, and thr eat and harm emotions, and negatively related to challenge and benefit emotions, job satisfaction , and motivati on. Although the relationships between the dimensions of the challenge stressors and t he criteria were not quite as strong , the re sults indicated that these forms of stressors were positively related to frustration, emotional exhaustion, challenge related emotion, and motivation, and were not related to harm related emotion. S omewhat unexpectedly, both the work load and work pace dimensions were also found to be positi vely related to threat related emotion . These findings would suggest that relationships bet ween these work stressors and em ployee emotions may be more complex than for the other challenge stressor dimensions. However, the pattern of relationships reported in the present study between the challenge and hindrance stressor dimensions and the employ ee criteria is generally consistent with expectations.

PAGE 164

164 In summary, the data gathered from three samples reported in Study 1 provides strong evidence for the construct validity of the newly developed measures of the challenge and hindrance stressor constru cts, in that the measures were shown to tap their intended underlying construct, be reliable, and to possess good convergent, discriminant and nomological validity. Challenge and Hindrance Stressors as Higher Order Factors C onsistent with my conceptual mod el, Study 1 also provided support for the second order factor structure of the stressor constructs. The notion that challenge and hindrance stressors are higher order constructs made up of their first order dimensions was supported in three ways. First, th is conceptualization was supported by theory linking the stressor dimensions through the stress appraisal process described by Lazarus and Folkman (1984). Generally, this theory was used to argue that employees will appraise specific work demands as either : (a) challenging, when they are perceived to facilitate the expression of commitments to meet goals, complete tasks, and develop themselves; or (b) hindering, when they are perceived as creating an obstacle to task accomplishment and personal growth. Sec ond, the higher order conceptualization was also supported by empirical evidence which indicated that: (a) the dimensions of the challenge stressor construct tended to have moderate, positive correlations with each other; (b) the dimensions of the hindranc e stressor construct tended to have moderate, positive correlations with each other; but (c) the dimensions of the challenge stressor construct were virtually unrelated to the dimensions of the hindrance stressor construct. Additional analyses indicated th at although dimensions of the challenge and hindrance stressors exhibited similar relationships with emotional exhaustion and frustration, they generally had differential relationships with the other employee criteria. Specifically, dimensions of the chall enge stressor construct tended to have functional relationships with employee challenge and threat emotions, job satisfaction, and motivation; whereas the

PAGE 165

165 dimensions of the hindrance stressor construct tended to have dysfunctional relationships with these criteria. Therefore, the dimensions of the respective stressor constructs appear to be related to each other and exhibit some similar relationships with employee criteria. Finally, although results from the second order and first order challenge and hindr ance stressor CFAs indicated similar levels of fit, structural models examining the stressor criterion relationships using these two approaches produced different results. Specifically, the second order stressor models tended to explain as much, and in man y cases more variance in employee strains, emotions, satisfaction, and motivation than did models examining the unique effects of the first order dimensions. The exception to this trend was the relationships between the challenge and hindrance stressors an d the employee performance criteria. However, these results may have been a function of the smaller sample used in this analysis. Therefore, this study provided theoretical and empirical support for modeling the challenge and hindrance stressors as second order constructs comprised of their respective first order dimensions. The Nature of the Relationships between the Stressor Constructs and Their Dimensions Although the results of this study generally supported the second order factor structure of the stre ssor constructs, some questions may still exist over the appropriate modeling of the relationships between the constructs and their dimensions. I have provided strong theoretical arguments that the second order challenge and hindrance constructs should be modeled as having formative first order dimensions. This perspective is consistent with conceptualizations of employee perceptions of stress as being formed by environmental events and conditions as opposed to being an underlying variable that causes emplo yees to experience work d emands. This study supported this perspective in that the dimensions of the challenge and hindrance stressor constructs were not correlated at a high level with each other, and were not found to make eq ual contributions to the seco nd order stressor constructs. Specifically, in this study, work

PAGE 166

166 pace and job complexity were found to have significant effects on the second order challenge stressor construct, and role conflict, resource inadequacies, organizational politics, and job inse curity were found to have significant effects on the second order hindrance stressor construct. This may lead readers to ask why the stressor dimensions that had non significant relationships with the higher order constructs were retained in the analysis. Indeed, some may argue that dimensions of the challenge and hindrance stressor construct that have non significant empirical relationships (i.e., workload and job responsibility, and role ambiguity, administrative hassles, and organizational politics, res pectively) with the second order construct should be removed because they do not contribute to the creation of these construct. However, it is important to remember that formative indicators are not subject to the same guidelines as reflective indicators w hen assessing their relationships with the focal construct. Formative indicators are developed to represent a census of the dimensions that make up the focal construct and therefore, must all be included when modeling the construct. If they are not, constr uct deficiency will result. Because the relationships between the formative indicators and the second order constructs are based on theory, the removal of dimensions having non significant paths to the higher order constructs will result in an incomplete a ssessment of the challenge and hindrance stressors. Of course, I am not suggesting that researchers should ignore the fact that some dimensions did not seem to make significant empirical contributions to the challenge and hindrance stressor constructs. Obv iously, it is necessary to assess the validity of our theories and conceptual frameworks with empirical data, and constructs modeled with formative indicators are no exception. However, we do not reject a theory (or elements of a theory) with the results o f

PAGE 167

167 a single empirical test. The results provided by the 178 working MBAs in this sample do provide preliminary evidence indicating that some of the dimensions may not be as important to the measurement of the stressor constructs as are others. However, at t his stage in the development and validation of the challenge and hindrance stressor constructs, it would be premature to remove dimensions of the challenge and hindrance stressors without additional data. Therefore, these dimensions were retained to allow for the complete specification of the second order constructs, and to provide more accurate comparisons of the formative and reflective indicator models. Others may argue that all the dimensions are unnecessary once the second order construct has been ide ntified with two global reflective indicators. In other words, the global reflective indicators may be seen as sufficient to model the challenge and hindrance stressor constructs, making the dimensions unnecessary. However, there are at least two reasons t hat the individual dimen sions should be retained when modeling the challenge and hindrance stressor constructs as having formative indicators. First, the dimensions of the challenge and hindrance stressor constructs have been identified as critical to the conceptual domains of these constructs. Each dimension has been argued to contribute to the creation of the challenge and hindrance stressor constructs, respectively. Because each of these stressor constructs has been specified using formative indicators, each of the dimensions is necessary to estimate a complete census of the order dimensions from models estimating these relationships does not assess the complete construct of interest. I would argue tha t the challenge and hindrance stressors, when assessed only by their respective global reflective focal constructs.

PAGE 168

168 Second, it is also important to retain the first order dimensions for empirical reasons. Including the paths between the first order dimensions and the second order stressor constructs allows researchers to assess the relative weights of the dimensions. These weights provide an empirical assessmen challenge and hindrance stressors in the workplace. Retaining these dimensions will allow for comparisons across samples, work context s , and time; thus providing researchers with a m ore complete picture of which challenge and/or hindrance stressors are most important to a given group of individuals , or within a specific organization or industry, and how these patters fluctuate over time or with environment events. Limitations As is th e case with a ny study, th ere are some limitations to my r esearch. One limitation is that only one sample was used to examine the factor structure, reliability, and nomological validity of the newly developed scales. The development and validation of a cons truct is an ongoing process that requires data from several samples to determine the validity and generalizability of the measures. Therefore, additional research designed to replicate Study 1 and extend the nomological network of the challenge and hindran ce stressors to include other employee outcomes such as organizational commitment, voice behavior, or creative performance, will be important. This information is critical to a better understanding of the nature of the relationships between the stressor co nstructs and their respective dimensions, and the relationships between the stressor constructs and employee outcomes. Another limitation is that many of the criteria variables used in the present study were obtained from the same source as the challenge obtained from the same source (task performance, OCB, and CWB) tended to provide equivocal relationships with the stressor constructs. This raises the possibil ity that common method biases

PAGE 169

169 (Podsakoff et al., 2003a) played a role in the stressor criterion relationships observed in this study. However, it is worth noting that it was more appropriate to obtain measures of many of the criteria examined in this study from the respondents themselves. It is difficul t to believe that other sources could provide more accurate, or less biased, assessments of employee strains, emotions, or job attitudes than could self report ratings. In addition, this study explicitly controlled for employee negative affectivity in the relationships between the stressor constructs and the employee dimensions. Employee NA has been identified as a rater characteristic that may be an important confound in the stress literature (Podsakoff et al. , 2003a). Therefore, this study did control for one important form of method variance and generally found stressor criterion relationships to be robust against the effects of employee NA. However, future research may be better able to address this limitation using procedural and/or analytic technique s. For example, researchers may limit the effects of method biases by temporally separating the employee responses regarding perceived work stressors from the data collection of other employee criteria. Separating the survey responses in time would also be consistent with recommendations in the stress literature for more longitudinal research (Kahn & Byosiere, 1992; Sonnentag & Frese, 2003). In addition, researchers could also attempt to use statistical procedures in latent variable modeling to help control for method biases. To my knowledge, no one has attempted to do this using constructs modeled as having formative indicators, and therefore, this may be an interesting avenue for future research. Finally, the lack of support for the relationships between the stressors and performance criteria may have been limited by range restriction in the performance measures. T he majority of the MBA s who participated in Study 1 were sponsored for this program by the companies for which they work. This means that the co

PAGE 170

170 student agrees to continue to work for the company for a specified period of time after they have received their degree (typically for a period of two to four years). Given the investment these compan ies are making, it would make sense for them to only sponsor employees who have been identified as high performers. Consistent with this explanation, an examination of the means and standard deviations for the supervisor ratings of task performance (M = 4. 38/5.00, SD = 0.59), OCB (M = 3.90/5.00; SD = 0.54), and CWB (M = 1.68/5.00, SD = 0.92) suggests that Therefore, additional research using samples with greater variability in performance will be necessary to provide a more adequate test of the relationships between the newly developed challenge and hindrance stressor scales and job performance. Conclusions Despite these limitations, the evidence provided in Study 1 is encour aging in that the newly developed challenge and hindrance stressor scales were found to possess content validity, be reliable and consistent with the hypothesized factor structure, and to demonstrate c onvergent, discriminant , and nomological validity. In a ddition, this study provides support for the use of higher order constructs when modeling challenge and hindrance stressors. Drawing more specific conclusions will require data to be gathered from multiple additional samples. Nevertheless, the support prov ided in this study justifies the use of these scales to as sess work stressors in Study 2.

PAGE 171

171 CHAPTER 5 STUDY 2 Generally speaking, the purpose of Study 1 was to develop and validate new measures of c hallenge and hindrance stressor s. The purpose of Study 2 is to use these new measures to test the potential linear and curvilinear effects of these stressors on employee responses in the workplace and to examine the potential moderating effect s of some e mployee individual differences on these relationships. To tes t the hypotheses described in Chapter 2 , I use d an experience sampling method (ESM) design and multi level modeling techniques to examine the relationships between daily levels of challenge and hindrance stressors and trains, jo b satisfaction, negative affectivity, ability, and locus of control moderate d these relationships. Method Participants The sample for this study was comprised of 164 employees of a national insura nce branch office located in the southeastern United States. T able 5 1 reports the demographic characteristics of this sample. As indicated in this table, the sample was primarily comprised of female participants (79.9 %), who were either married (4 2.7 %) or single (37.8%) , and who were on average, approximately 34 years (SD = 10.9) of age. In addition, participants in this study had an average of approximately two and one half years (SD = 3.73) of tenure in their current jobs and over five years (SD = 5.73) of tenure in the organization. Procedure Participants first completed an online survey which included measures of demographic variables, negative affect ivity , ability, and locus of control. Following this, participants were asked to fill out a daily quest ionnaire on each day they worked over the course of a three week

PAGE 172

172 period. The daily surveys were comprised of scales measuring challenge and hindrance stressors, employee strains, job satisfaction, task performance, OCB, and CWB. Participants were sent emai ls each day at noon which reminded them to complete their daily survey and provided a link to the online survey website. They were instructed to complete the daily surveys toward the end of their work day, before they left their workspace. Although the vas t majority of participants used the online response format, about 10% of the sample did not have access to the internet at work. These individuals completed paper versions of the surveys, which were submitted to a collection box daily and retrieved from th e branch office each night. Each daily questionnaire included measures of challenge and hindrance stressors, emotional exhaustion, frus tration, job satisfaction, task performance , OCB, and CWB. Employee participants were paid $30.00 for completing the one time survey and at least eig ht of the daily surveys, and a bonus of $20.00 if they completed 15 daily surveys. U sable data was gathered from 164 participa nts who provided a total of 2134 daily surveys, for an average of 13 daily surveys (SD = 3.12) per pa rticipant. Measures The r epeated measu res design of this study lead to concerns regarding survey length and participant fatigue. Specif i cally, I was worried that participants would become weary of completing 15 daily surveys if these surveys included too many questions . If this occurred, the honesty or accuracy of the responses may have been diminished, and/ or re sponse rates may have substantially declined . Therefore, I used shortened versions of several scales so that the daily surveys could typically be completed in less than five minutes, and so that the likelihood of these biases occurring was decreased.

PAGE 173

173 Challenge and hindrance stressors E versions of the scales devel oped and validated in Study 1. Each challenge and hindrance stressor dimension was measured with one item. Each i tem wa s selected because it: (1) best represented the construct, (2) could be modified to measure daily work demands, and (3) was expected to b est capture potential daily fluctuations over the three week survey period. Unfortunately, due to concerns over participant reactivity to this measure, the organization would not allow me to include measures asse ss ing employee job insecurity. Therefore, t he challenge stressor dimensions of work load, job responsibility, job complexity, and work pace and the hindrance stressor dimensions of role ambiguity, role conflict, resource inadequacies, administrative hassles, interpersonal conflict, and organization al politics were assessed with one item scales on each daily survey. In the present study, t he average internal co nsistency reliabilities for the challenge and hindrance s tressor scales were .85 and .84, respectively . Job related strains Job related stra ins were measured with a 6 item version of the emotional exhaustion scale develope d by Maslach and Jackson (1986), and a 3 item frustration scale developed by Peters et al. (1980). The average internal consistency reliabilities for these scales in the pres ent study were .93 and .81, respectively. Job satisfaction The 3 item measure develo ped by Cammann et al. (1983) was used t o assess general employee job satisfaction. The average internal consistency reliability for this scale was .86 in the present study .

PAGE 174

174 Task performance Employee task performance was measured using a 3 item version of Williams and daily surveys in the present study was .92. Organizational ci tizenship behavior Employee OCB was measured using a 7 scales. The average internal consistency reliability for this scale was .87 in the present study. Counterproductive work behavior Counterproductive work beha vior was measured using 7 items taken from the workplace deviance scale developed by Bennett and Robinson (2000). The average internal consistency reliability for this scale was .75 in the present study. Negative affectivity Employee trait based negative affectivity was measured using 10 items taken from the Positive and Negative Affect Schedule ( PANAS) scale reported by Watson et al. (1988). The internal consistency reliability for this scale in the present study was .76. Employee ability Employee ability was assessed using a self report measure of employee ability was gathered using the 7 item Ability, Experience, Training, and Knowledge (AETK) scale reported by Podsakoff and MacKenzie (1996). exper ience, training, or job knowledge to act independent of my immediate supervisor in Previous research reported by these authors has found this scale to be reliable ( Podsakoff & MacKenzie, 1996) . The reliability for this scale in the p resent study was .80.

PAGE 175

175 Locus of control Employee locus of control was assessed using five items taken from the I nternational Personality Item Pool (IPIP, 2006). An I believe that my success depends on my ability rather th an luck in the present study was .61. Items reflecting the challenge and hindrance stressor dimensions and the emotional exhaustion, frustration, job satisfaction, task performance , CWB , negative affec tivity, (self reported) ability, and locus of control constructs were rated on 7 point Likert scales with anchors complete versions of the scales identified above were included in at least one of the daily surveys to ensure that the selected items were valid representations of the constructs they assessed. For example, Daily Survey #1 included all four items used to measure the workload dimension (taken from Step #3 of S tudy 1) and Daily Survey #2 included all four items used to measure the job responsibility dimension (taken from Step #3 of Study 1). These additional items were used to assess the validity of the single item scales (see below). In addition, the full set o f items for each of the following scales was included in one of the daily surveys: emotional exhaustion (Daily Survey # 15), task performance (Daily Survey # 4), OCB (Daily Survey # 7), and CWB (Daily Survey # 13). Analysis I used multilevel analysis to t est hypotheses regarding the linear and curvilinear relationships between challenge and hindrance stressors and employee strains, job satisfaction, task performance, OCB, and CWB, and the potential moderating effects of employee negative affect ivity , abili ty, and locus of control on these relationships. All hypotheses were tested using the HLM 5.04 program developed by Bryk, Raudenbush, and Congdon (2001).

PAGE 176

176 In the first step of this process, I estimated null models for the dependent variables by entering eac h of these variables into a multi level model by itself and examining the variance components for these models. Second, I estimated Level 1 regression models designed to test the linear effects of the stressor constructs on employee strains, job satisfacti on, and performance by entering both challenge and hindrance stressors as Level 1 predictors of these criteria. In the third step, I tested the hypothesized curvilinear stressor relationships by adding the squared challenge and hindrance stressor product terms to the Level 1 regression models in HLM. Finally, to test the cross level mod erating hypotheses, I entered negative affectivity , ability, and locus of control as Level 2 predictors of the intercepts and slopes of the Level 1 regressions. To interpre t the estimates from these models as strictly representing within individual relationships, I centered the predictor variables at the individual means for each participant (Hoffman, Griffin, & Gavin, 2000). Centering the variables in this manner removes be tween individual variables are Judge et al., 2006, p. 130). Results Creating Scale Scores for the Challenge and Hindrance St ressor Constructs Consistent with the arguments made in Chapter 3, I believe that it is more appropriate to model the challenge and hindrance stressors as constructs having formative indicators. Briefly, this is the case because the dimensions of these con structs are: (a) conceptualized as combining together to form the construct rather than representing reflections of the underlying constructs, (b) not expected to covary at a high level with each other, and (c) not necessarily interchangeable with each oth er. Therefore, for the purposes of the HLM analysis, I thought it more appropriate to create scale scores for the challenge and hindrance stressors that was consistent with this conceptualization.

PAGE 177

177 Unfortunately, the previous literature discussing formativ e indicators (Boll en, 1989; MacKenzie et al., 2005 ; Podsakoff et al., 2006) has focused on using latent variable modeling techniques and does not provide clear procedures for creating scales scores using formative indicators. As a result, researchers who h ave examined c onstructs having formative indicators have typically used traditional reflective indicator modeling techniques when creating scale scores for these types of constructs. For example, in their development of the job embeddedness construct, Lee, Mitchell, and their colleagues (Lee, Mitchell, Sablyniski, Burton, & Holtom, 2004; Mitchell, Holtom, Lee , Sablynski, & Erez, 2001) argued that this construct is comprised of multiple, distinct dimensions (community and work fit, links, and sacrifice) that are not necessarily highly correlated, are not interchangeable, and combine to form the higher order construct. However, despite the recognition by these authors that job embededdeness is most appropriately modeled as having formative indicators, Mitchell et al. (2001) created a composite 111). Although this technique is often used for indicators of a reflective constructs, it is not appropriate when the indicators of a con struct are not expected to contribute equally to the structure of a higher order construct. Therefore, I attempted to create scale scores for the challenge and hindrance stressor constructs that better addressed the formative nature of their indicators, and that were not subject to the limitations of traditional techniques. In order to obtain the relative weights of the dimensions of the challenge and hindrance stressor constructs, I conducted a CFA using data taken from the sample of insurance branch em ployees. I did not use the weights provided in Study 1 because: (1) the relative weights for different types of work stressors may fluctuate across samples, occupations, or organizational contexts, and (2) I was unable to gather data

PAGE 178

178 regarding the job inse curity dimension of the hindrance stressor construct in the current sample, which would likely affect the relative weights of the stressors . Therefore, I conducted the CFA by modeling the challenge and hindrance stressor constructs as having formative ind icators using the data gathered from Daily Survey #1. I selected the first day of the ESM study because: (1) almost all of the participants provided data on this day and (2) employee responses to this survey would not be primed by questions from any previo us daily surveys. I used AMOS 6.0 to estimate this model and obtained factor score weights for each of the dimensions of the stressor constructs. Factor score wei ghts are the regression coefficients used to predict the latent variables from the observed it ems. In other words, they provide the unique empirical contribution of each factor to the overall composition of the latent variable. Therefore, I created the scale scores for the challenge and hindrance rating of the stressor items based on the factor score weight provided by the CFA. I also created scales scores for the stressor constructs using a traditional unit weighting approach that was more consistent with reflective indicator modeling. This allow ed me to compare and contrast the results from both approaches and identify patterns. In the following sections, I will first present the results of the analyses using the scale scores created using the formative indicator approach , and then present analys es using the scale scores developed using the reflective indicator approach. Construct V alidity The reliabilities and intercorrelations for the variables used in this study are reported in Table 5 2. This table reports the average of the n weight ed reliabi lities for each scale across the daily (Level 1) surveys. As indicated in this table, the average reliability coefficients were fairly high and generally Indeed, with the exception o , the measures used in this study

PAGE 179

179 tended to be reliable in both the pre study and daily surveys. Table 5 2 also reports the intercorrelations between these variables at the between individual and within individual leve ls. The between individual correlations were calculated by aggregating the L evel 1 data across individuals. Within individual correlations were computed by standardizing the results of a series of simple regressions conducted in HLM. In order to ass ess the extent to which the one item measures of the dimensions of the challenge and hindrance stressor constructs used in this study were representative of the complete constructs , I examined the relationship between the one item measures and the additional scal e items included in the individual daily surveys. Generally sp eaking, I found the correlation s between these two set of items to be quite high for the measures of the challenge stressors. For example, the average correlation between the one item measure of workload used in this study and the other three items added to Daily Survey #1 to measure this construct was .88. Similar results were obtained for the average correlations calculated for the one item measures of job responsibility (average r = .68), job complexity (average r = .79), and work pace (average r = .83). The correlations between the one item measures of the dimensions of the hindrance stressor construct and the additional scale items were also found to be quite high. Indeed, t he average correl ation s between t he on e item measure s and the additional items used to assess resource inadequacies (average r = .83 ), role ambiguity (average r = .53), role conflict (average r = .78), organizational politics (average r = .68), administrative hassles (aver age r = .71), and interpersonal conflict (average r = .67) were all reasonably high . Thus, there was evidence indicating that the one item measures of the challenge and hindrance stressor constructs were representative of the complete , validated measures o f these constructs.

PAGE 180

180 Partitioning of Variance C omponents Before testing the hypothesized relationships, I estimated null models for each of the endogenous variables to determine the systematic within versus between individual variance in each variable . A n ull model is estimated in HLM by enter ing a L evel 1 variable as dependent variable into the model by itself. The results of the null model est imates are reported in Table 5 3 . For each model, this table reports the intercept ( 00 ; the average level of that variable across individuals), within 2 ), t value, between individual variance 00 ), and the percen tage of variance that is within individual. This is an important step as it determines whether multilevel analysis is appropriate; if no within individual variation exists in the endogenous variables, then multilevel analysis should not be used. As indicated in Table 5 3 , all the endogenous variables examined in this study had significant between individual variance (all p values < .01). In addition, a substantial amount of the variance in each of the constructs was found to be within individual. Specifically, a pproximately one fifth to one quarter of the varian ce in emotional exhaustion (24.2%), job satisfaction (21. 5 %), and OCB (2 5.9%); one third of the variance in frustration (3 4.8%); and nearly a half of the variance of in task performance (54.2 %) and CWB (45.9 %) was within person. These results suggest ed that multilevel modeling was appropriate and that there was a substantial amount of within person variability to explain in each of the endogenous variables. Tests of the L inear Effects H ypotheses I tested hypotheses regarding the main effects of the c hallenge and hindrance stressors on the employee outcomes by r egressing employee strains, job satisfaction, and the performance criteria on the challenge and hindrance stressors . T he results of these regression analyses are reported in Step 1 in Table 5 4 . This table reports the unstandardized parameter estimates for models estimating the relationships between the challenge and hindrance stressors when modeled

PAGE 181

181 as having formative indicators (B f ; located in the left hand column under each criterion ) and when modeled as having reflective indicators (B r ; located in the r ight hand column under each criterion ). Linear effects when challenge and hindrance stressors are modeled as having formative indicators Generally speaking, the findings reported in Table 5 4 provide mixed support for the hypothesized liner relationships between challenge and hindrance stressors (when modeled as having formative indicators) and employee strains , job attitudes, and performance . Specifically, the results reported in Step 1 in this table support Hypotheses 1 and 11 in that both challenge and hindrance stressors were found to have positive relationships with employee emotional exhaustion (B f values = .07 and .08, respectively, both p values < .05) and frustration (B f values = .17 and .15, respectively, both p values < .01). Therefore, both chal lenge and hindrance stressors were found to be have positive relationships with employee psychological and emotional strains. T he results reported in Step 1 in Table 5 4 provided mixed support for the hypothesized differential linear stressor relationships with job satisfaction. Specifically, t he data reported in that table provides s upport for Hypothesis 12 in that hindrance stressors (when modeled as having formative indicators) were found to have a negative relationship with job satisfaction (B f = .05, p < .05). However, Hypothesis 2 was not supported by the results reported in this table , because challenge stressors (when modeled as having formative indicators) were not found to have a significant positive relationship with this criterion (B f = .03, p > .05). Generally speaking, the data presented in Step 1 in Table 5 4 also provides limited support for hypotheses regarding the differential linear relationships between challenge and hindrance stressors and the three employee performance criteria. First, the results reported in this table do

PAGE 182

182 not provide support for Hypothesis 3 or Hypothesis 13 in that when modeled as having formative indicators, neither challenge stressors (B f = .03, p > .05) nor hindrance stressors (B f = .01, p > .05) were found to hav e significant relationships with task performance. Second, the results reported in Table 5 4 indicate that Hypothesis 4 was supported in that challenge stressors had a positive relationship with employee OCB (B f = .11, p < .01). However, the results repor ted in this table did not support Hypothesis 14 in that hindrance stressors were not found to have a significant negative relationship with this criterion (B f = .00, p > .05). Finally, the results reported in Table 5 4 provide no support for hypotheses reg arding the linear relationships between the stressors and employee CWBs. Specifically, this table does not provide support for Hypothesis 5 in that challenge stressors had a non significant relationship with CWB (B f = .01, p > .01), and Hypothesis 15 was not supported in that hindrance stressors had a non significant relationship with t his criterion (B f = .02, p > .01). In summary, analyses testing the linear relationships between challenge and hindrance stressors found that both of these types of work dem ands had positive relations hips with employee strains, but fo und limited support for th e differential relationships these stressors were expected to have with employee job satisfaction and the performance criteri a . Linear effects when challenge and hindran ce s tressors are modeled as having reflective indicators When comparing the above results with those obtained from analyses in which the challenge and hindrance stressor scales modeled as having reflective indicator techniques, two patterns are important t o note. First, the findings were consistent in that the challenge and hindrance stressors (when modeled as having reflective indicators ) both had positive relationships with emotional exhaustion (B r = .29 and B r = .28, both p values < .01) and frustration (B r = .22, and B r = .36, both p values < .01) , and hindran ce stressors also predicted

PAGE 183

183 variance in employee job satisfaction (B r = .22, p < .01). Second, the results from these analyses provided support for the hypothesized relationships between the stress ors and the employee performance criteria. Indee d, although challenge stressors were found to have a positive relationship with OCB in both sets of analyses, only when the stressor constructs were modeled as having reflective indicators were challenge and hindrance stressors found to have their hypothesized differential relationships with task performance (B r = .12, p < .01 and B r = .10, p < .05, respectively ) and CWB (B r = .07 , p < .01 and B r = .05 , p < .01 , respectively ) . Therefore, the primary differen ce between these two sets of analyses is that the linear stressor performance relationships tended to be more consistent wit h the hypotheses when the challenge and hindrance stressor constructs were modeled using traditional reflective indicator techniques than when they were modeled using a formative indicator approach. Tests of the C urvilinear Effects H ypotheses To test for the hypothesized curvilinear stressor effects , I calculated squared product terms for the challenge an d hindrance stressor construct s. Before creating the product terms, the stressor variables were centered to reduce nonessential multicollinearity in the higher order polynomial terms and to facilitate interpretability (Cohen, Cohen, West, & Aiken, 2003). Following this, the squared cha llenge stressor product term was added to regressions that controlled for the linear effects of the challenge and hindrance stressors constructs. The results of these analyse s are reported in Step 2 of Table 5 4 . This table reports the unstandardized para meter estimates for models estimating the relationships between the challenge and hindrance stressors when modeled as having formative indicators (B f ; located in the left hand column under each criterion ) and when modeled as having reflective indicators (B r ; located in the right hand column under each criterion ).

PAGE 184

184 Curvilinear effects when challenge and hindrance s tressors are modeled as having formative indicators Generally speaking, the results reported in Step 2 in Table 5 4 provide limited support for th e hypothesized curvilinear relationships between challenge stressors (when modeled using formative indicator techniques) and the employee criteria. Specifically, the results indicate that H ypotheses 6, 7, 8, and 10 were not supported in that challenge str essors did not have significant curvilinear relationships with emotional exhaustion (B f = .02, p > .05 ) , frustration (B f = .00, p > .05), task performance (B f = .01, p > .05), or CWB (B f = .01, p > .05). Indeed, Table 5 4 indicate s that the curvilinear cha llenge stressor product term was only found to be significant in the prediction of OCB (B f = .04, p < .01). Figure 5 1 provides an illustration of this curvilinear relationship. Unfortunately, this figure indicates that the curvilinear relationship does no t support Hypothesis 7 in that OCB tended to increase at an accelerated rate as the level of challenge stressors experienced by employees exceeded their average daily levels. Therefore, it appears that the data gathered in this study do not support the exp ectation that high levels of challenge stressors lead to accelerating rates of employee strains and counterproductive work behavior , and decelerating rates of job satisfaction, task performance, and OCB. In this the next step of my analysis, I entered th e squared hindrance stressor product term into the regressions equations reporte d in Step 2 of Table 5 4 . Although I did not develop specific hypotheses regarding these relationships, this analysis provided a more complete assessment of the curvilinear str essor relationships. The results of the regression analyses including the squared hindrance stressor product term are reported in Step 3 of Table 5 4 . This table indicates that the curvilinear hindrance stressor term did not account for significant varianc e in any of the employ ee criteria (all p values >.05). Therefore, this study provides little

PAGE 185

185 evidence suggesting that challenge or hindrance stressors had curvilinear relationships with strains, job satisfaction, or employee performance . Curvilinear effec ts when challenge and hindrance s tressors are modeled as having reflective indicators Table 5 4 also reports the results from the curvilinear stressor effects analyses with the challenge and hindrance stressor constructs modeled as having reflective indica tors. When comparing these results to those obtained when the stressor constructs were modeled as having formative indicators, there are three important things to note. First, consistent with the results when challenge stressors where modeled as having for mative indicators, the results i n this table indicate that when these stressors were modeled as having reflective indicators, none of the curvilinear relationships with any of the employee performance criteria were significant. Specifically, when challenge stressors were modeled as having reflective indicators, the curvilinear challenge stressor product term was not significant in regressions with task performance ( B r = .01 , p > .05) , OCB ( B r = .01, p > .05) , and CWB ( B r = .00 , p > .05). H owever, the result s reported in Table 5 4 indicate that when challenge stressors were modeled as having reflective indicators, they did exhibit significant curvilinear relationships with other employee outcomes . For example, Table 5 4 indicates that challenge stressors had a significant curvilinear relationship with emotional exhaustion (B r = .05, p < .01). Figure 5 2 provides a n illustration of this curvilinear relationship. Consistent with expectations, this graph indicates employees perceived levels of emotional exhaustio n increased at an accelerating rate as the level of experienced challenge stressors increases. The results reported in this table also indicate that the squared challenge stressor product term predicted variance in employee job satisfaction (B r = .04, p < .01). An illustration of this relationship is provided in Figure 5 3. C onsistent with expectations, this figure indicates that

PAGE 186

186 employee job satisfaction decreases at an accelerated rate as c hallenge stressors exceed an normative daily level s . Therefore, although when modeled as having reflective indicators, challenge stressors do not have significant curvilinear relationships with any of the employee performance criteria, they do have curvilinear relationships with employee strains and satisf action. Finally, the results of the regression analyses for hindrance stressors reported in Table 5 4 are fairly consistent regardless of the method by which the se stressor constructs were modeled. Specifically, the hindrance stressor product term was fou nd to be non significant in nearly all of the regressions. The only exception to this was that the curvilinear relationships between hindrance stressors and task performance was found to be significant (B r = .07, p > .05) , when these stressors were modele d as having reflective indicators. A graph of this relationship is provided in Figure 5 4. This figure indicates that employee task performance became increasingly negative as employee s experienced hindrance stressors that exceeded their normative daily le vels. In summary, c ompar ing the results of the curvilinear stressor effects analyses using these two approaches to construct modeling suggests that, when modeling the stressor constructs as having reflective indicators : (1) the challenge stressors had cur vilinear relationships with emotional exhaustion and job satisfaction , but not with OCB; and (2) hindrance stressors had an unexpected curvilinear relationship with employee task performance. Tests of the Cross L evel M oderating Effects H ypotheses To test hypotheses regarding cross level moderation effects, I entered a series of Level 2 variables into the multi level regression equations to estimate their effect on the Level 1 relationships reported in Step 2 of Table 5 4 . Specifically, I entered negative a ffectivity, ability, and locus of control as predictors of the variability in the Level 1 intercepts and slopes of the

PAGE 187

187 relationships between challenge and hindrance stressors and the other employee criteria. The results from these ana lyses are reported in Tables 5 5, 5 6, 5 7, 5 8, 5 9, and 5 10 . These tables report the unstandardized parameter estimates for models estimating the relationships between the challenge and hindrance stressors when modeled as having formative indicators (B f ; located in the left hand column under each criteria) and when modeled as having reflective indicators (B r ; located in the right hand column under each criteria). Moderator effects when challenge and hindrance s tressors are modeled as having formative indicators Generally spe aking, the results from Tables 5 5 to 5 10 provide very limited support for the hypothesized moderator effects of employee individual differences on the relationships between challenge and hindrance stressors and employee strains, satisfaction, and perform ance. Table 5 5 reports analyses testing the potential moderating effects of negative affectivity, ability, and locus of control on the relationships between challenge and hindrance stressors and emotional exhaustion. As indicted in this table, the hypothe sized moderating effects of these individual differences were not supporte d. Specifically, Hypotheses 16a, 17a, 18a, 19a, 20a, and 21a did not receive any support in this study in that negative affectivity, ability, and locus of control did not moderate an y of the linear or curvilinear relationships between the challenge and hindrance stressors (when modeled using formative indicator techniques) and emotional exhaustion (all p values > .05). However, the findings in this table indicate that negative affecti vity 01 = 1.27, p < .01); and that locus of control ( 01 = .29, p < .01) had a significant negative main effect on this criterion. Therefore, these findings suggest that individuals higher in N A typically experience higher average levels of emotional exhaustion than do individuals lower in NA; whereas

PAGE 188

188 individuals with a more internal locus of control experience lower average levels of emotional exhaustion than individuals with an external locus of control. Table 5 6 reports data from analyses examining the moderating effects of negative affectivity, ability, and locus of control on the relationships between challenge and hindrance stressor s and employee frustration. As indicated in this table, lo cus of control was found to moderate the linear relationship between hindrance stressors (when modeled using formative indicator techniques) and frustration ( 01 = .07, p < .05). Contrary to expectations, Figure 5 5 indicates that the relationship between challenge stressors and employee frustration was stronge r for employees with a more internal locus of control than for employees with a more external l ocus of control. Therefore, these findings do not provide support for Hypothesis 21a. However, none of t he other interactive effects reported in Table 5 5 were found to be significant. When taken together , these findings provide mixed support for Hypothesis 21b and a general lack of support for Hypotheses 16a, 17a, 18a, 19a, and 20a i n that the individual di fferences measured in this study (n egative af fectivity, ability, and locus of control) did not moderate the linear or curvilinear relationships between c hallenge or hindrance st ressors (when modeled as having formative indicators) and employee strains. How ever, it should be noted that NA had a p 01 = 1.14 , p < .01) , whereas locus of control had a negative main 01 = .30, p < .01). Similar to the findings for emotional exhaustion, these results suggest that individuals higher in NA typ ically experience higher average levels of frustration than do individuals lower in NA; whereas individuals with a more internal locus of control experience lower average levels of frustration than individuals with a more external locus of control.

PAGE 189

189 Data f rom analyses examining the moderating e ffects of negative affectivity, ability, and locus of control on the relationships between challenge and hindrance stressors (when modeled as having formative indicators) and employee job satisf action are reported in Table 5 7 . As indicated in this table, the data did not provide much support for the hypothesized moderating relationships. Specifically, Hypotheses 16b, 17b, 20b, and 21b were not supported in that neither negative affectivity nor locus of control were fo und to moderate any of the linear or curvilinear relationships between challenge and hindrance stressors (when modeled as having formative indicators) and employee job satisfaction (all p values > .05). However, Hypothesis 18b did receive partial support i n that ability was found to moderate the curvilinear relationship between challenge stressors and employee job satisfaction ( 31 = .02, p < .01 ) . Consiste nt with expectations, Figure 5 6 indicates that employees with lower levels of perceived ability were more sensitive to challenge stressors such that their levels of job satisfaction increased at an accelerated rate when these stressors are experienced in low to moderate levels, but then decreased at an accelerated rate when these stressor s exceeded daily norms for the employee. However, Hypothesis 19b received no support as ability was not found to moderate the linear relationship between hindrance stressors (when modeled as having formative indicators) and employee job satisfaction ( 21 = .01, p > .05). Finally, Table 5 7 indicates that NA had a 01 = .72 , p < .01); and locus of control had a positive main effect on this job attitude 00 = .29 , p < .01). These findings indicate that employees who had higher levels of NA experienced lower average daily le vels of job satisfaction than did individuals with lower NA; whereas employees who had a more internal locus of control experienced higher average daily levels of jo b satisfaction than employee s who had a more external locus of control.

PAGE 190

190 Table 5 8 reports the results of analyses examining the moderating effects of negative affectivity, ability, and locus of control on the relationships between challenge and hindrance stressor s and employee task performance. T he results reported in this table do not provide support for any of the se po tential moderating effects. Specifically, Hypotheses 16c, 17c, 18c, 19c, 20c, and 21c were not supported in that none of the individual differences were found to mode rate of any of the linear or curvilinear relationships between the stressors (when modeled as having formative indicators) and task performance. In addition, the results in this table indicate that none of the individual differences had main effects on emp loyee task performance. The results of the analyses examining the moderating effects of the individual differences on the relationships between challenge and hindrance stressor s (when modeled using formative indicators techniques) and employee OCB are rep orted in Tab le 5 9. The data reported in this table provide limited support for the hypothesized moderating relationships. Indeed, negative affectivity and locus of control were not found to be significant predictors of the linear or curvilinear relationsh ips between challenge and hindrance stressors and employee OCB. Therefore, Hypotheses 16d, 17d, 20d, and 21d received no support from the data analyzed in this study. However, employee ability was found to have a significant moderating effect on the linear relationship between hindrance stressors and employee OCB ( 21 = .03 , p < .05 ). Figure 5 7 provides an illustration of this relationship. Contrary to expectations, this figure indicates that the negative relationship between hindrance stressors and OCB was weaker f or individuals with low levels of ability than fo r individuals with high levels of ability. Given that ability was not found to moderate the linear or curvilinear relationships between challenge stressors and OCB (both 11 and 31 = .01, both p values > .05), the data gathered for this study do not provi de evidence supporting Hypotheses 18d or 19d. Finally, the results reported in this table indicate that none of

PAGE 191

191 the individual difference variables (i.e., negative affectivity, ability, and locus of control) had a main effect on OCB. Table 5 10 reports th e results of analyses examining the moderating effects of negative affectivity, ability, and locus of control on the relationships between challenge and hindrance stressors and employee counterproductive work behaviors. Generally speaking, the results repo rted in this table provide no support regarding the hypothesized moderating relationships. S pecifically, negative affectivity, ability and locus of control were not found to account for variance in the slopes of any of the linear or curvilinear relationshi ps between the challenge and hindrance stressors (when modeled as having formative indicators) and employee CWB. Therefore, these results provide no support for Hypotheses 16e, 17e, 18e, 19e, 20e, or 21e. However, this table does report that employee locus of control had a negative main effect on employee CWB ( 01 = .07 , p < .05 ) such that employees who had a more internal locus of control engaged in lower levels of CWB than those who had a more external locus of control. Moderator effects when challenge and hindrance s tressors are modeled as having reflective indicators Tables 5 5 through 5 10 also provide the results of analyses testing the moderated effects of employee individual differences on the stressor criterion relationships when the stressor con structs were modeled as having reflective indicators. A comparison of these findings with results when the stressor constructs were modeled as having formative indicators produced three interesting patterns. First, when the stressor constructs were modeled as having reflective indicators, the individual differences tended to moderate more of the relationships between the challenge stressors and the employee criteria. For example, when the stressor construct s were modeled as having reflective indicators: (1) negative affectivity was found to moderate the linear relationship between the challenge stressors and emotional exhaustion 11 = .27, p < .01); (2)

PAGE 192

192 locus of control was found to moderate the linear relationship between challenge stressors and employee job 1 1 = .06, p < .05); ( 3) ability was found to moderate the curvilinear relationship between challenge stress ors and task performance 31 = .02, p < .05) ; (4) locus of control was found to moderate the linear relationship between challenge stressors and OCB 1 1 = .07, p < .05) ; and (5) negative affectivity was found to moderate the linear relationship between challenge stressors and CWB ( 1 1 = .06, p < .05). Second, plots of these cross level moderating effects showed that the majority of them were supportive of the hypothesized relationships between challenge stressors and the employee outcomes. For example, consistent with my expectations, Figure 5 8 indicates that the positive relationship between challenge stressors and emotional exhaustion was stronger for employees who exhibited higher levels of NA than for those with had lower levels of this trait. As an other example, Figure 5 9 provides a graph of the moderating effect of locus of control on the linear relationship between challenge stressors and employee job satisfaction . This graph also support s in that the relationship between challenge stressors and job satisfaction was somewhat more positive for individuals who had a more internal locus of control than for individuals who had a more external locus of control. Similarly, the plot of the moderating effect of locus of control on the relationship between challenge stressors and employee OCB presented in Figure 5 10 is also consistent with expectations. As indicated in this figure, the relationship between challenge stressors and OCB was more positive for individuals who had a more internal locus of control than for individuals who had a more external locus of control. Fina lly, as indicated in Figure 5 11, negative affectivity moderated the relationship between challenge stressors and CWB such that this relationship was somewhat more negative for individuals who had l ower levels of NA than for individuals who had higher levels of this trait.

PAGE 193

193 However, the significant interaction between challenge stressors and ability on employee task performance did not support m y hypotheses. Indeed, contrary to my expectations, Figure 5 12 indicates that employees who perceiv ed themselves to have lower levels of ability tended to experience an accelerated increase in task performance when faced with challenge stressors that were at level s higher than their daily norms; whereas employees who perceived themselves to have high er levels of ability experienced initial increases in task performance that decelerated as challenge stressors increased beyond normative daily levels. Third, the curvilinear relat ionships between hindrance stressors (when modeled as having reflective indicators) and task performance were moderated by some of the individual differences examined. Although I did not develop hypotheses for these specific relationships, I tested the pot ential moderating effects of negative affectivity, ability, and locus of control on the curvilinear relationship between hindrance stressors and task performance as this relationship was found to be significant in my previous analysis (see Step 3 of Table 5 4). Interestingly, both ability ( 41 = .06 , p < .05) , and locus of control ( 41 = .10 , p < .01 ) were found to moderate the curvilinear relationship between these variables. Illustrations of these relationsh ips are provided in Figures 5 13 and 5 14 , res pectively. As shown in Figure 5 13 , the curvilinear relationship between hindrance stressors and task performance indicated that these stressors had no relationship with task performa nce for low ability individuals; whereas the task performance of high ab ility individuals decreased at an accelerating rate as hin drance stressors exceeded normative daily levels. Similarly, Figure 5 14 depicts a somewhat similar relationship for the moderating effect of employee locus of control on the curvilinear relationshi ps between hindrance stressors and employee task performance. Specifically, this figure shows that employees with a more internal locus of control seem to

PAGE 194

194 initially increase their performance when experiencing low levels of hindrance stressors, but that pe rformance for these individuals decre ases at an accelerating rate when hindrance stressors exceed normative daily levels for the individual; whereas individuals who had a more external locus of control tended to experience a decelerating decline in task pe rformance as daily levels of challenge stressor s increased. In summary, the main differences between these two sets of analysis analyses were that when modeling the stressor constructs using reflective indicator techniques: (1) several additional challeng e stressor individual difference interaction s were identified, (2) many of these were consistent with the hypothesized relationships, and (3) the curvilinear relationship between hindrance stressors and task performance was moderated by ability and locus o f control. Tests of the Interactive S tressor Effects H ypotheses To test the potential interactive effects of challenge and hindrance stressors on the employee criteria , I calculated a challenge stress or X hindrance stressor product variable (when modeled using formative indicator techniques) and entered it into regressions that controlled for the linear effects of both the stressor variables. This product term was calculated from group mean centered versions of the challenge and hindrance stressor variable s. The results of these ana lyses are reported in Table 5 11 . As indicated in this table, these results do not provide support for Hypotheses 22a, 22b, 22c, 22d, or 22e in that the stressor product variable did not account for significant variance in any of the regressions involving emotional exhaustion, frustration, job satisfaction, task performance, OCB, or CWB. However, a comparison of these findings with results of the interactive stressor analyses when the stressor constructs were modeled using reflec tive indicator techniques identified two important differences. More specifically, the results across the two modeling techniques differ to the extent that the interactive stressor product term significantly predict ed variance in employee

PAGE 195

195 emotional exhaust ion (B r = .15, p < .05) and frustration (B r = .12, p < .05) when the challenge and hindrance stressors were modeled as having reflective indicators . Graphs depicting these relationsh ips are provided in Figures 5 15 an d 5 16 , respectively. These figures are consistent with expectations in that the relationships between challenge stressors and strains (in the form of emotional exhaustion or frustration) were more positive when hindrance stressors were experienced at higher as compared to lower levels. Therefo re, Table 5 11 indicates that similar results were obtained for the interactive effect of the stressors on employee satisfaction and performance regardless of which modeling technique was employed, but that significant interactive stressor effects on emplo yee job related strains were found only when modeling the stressor s as having reflective indicators.

PAGE 196

196 Figure 5 1. Plot of the curvilinear relationship between challenge stressors (modeled as having formative indicators) and organizational citizensh ip behaviors.

PAGE 197

197 Figure 5 2. Plot of the curvilinear relationship between challenge stressors (modeled as having reflective indicators) and emotional exhaustion.

PAGE 198

198 Figure 5 3. Plot of the curvilinear relationship between challenge stressors (modeled as having reflective indicators) and job satisfaction.

PAGE 199

199 Figure 5 4. Plot of the curvilinear relationship between hindrance stressors (modeled as having reflective indicators) and task performance.

PAGE 200

200 Figure 5 5. Plot of the m oderating effect of lo cus of control on the linear relationship between hindrance s tressors (modeled as having formative indicators) and frustration. External Locus Internal Locus

PAGE 201

201 Figure 5 6 . Plot of the moderating effect of ability on the curvilinear relationship between challenge stressors (modele d as having formative indicators) and job satisfaction . High Ability Low Ability

PAGE 202

202 Figure 5 7 . Plot of the moderating effect of ability on the linear relationship between hindrance stressors (modeled as having formativ e indicators) and organizational citizenship behaviors. Low Ability High Ability

PAGE 203

203 Figure 5 8 . Plot of the moderating effect of negative affectivity on the linear relationship between challenge stressors (modeled as having reflective indicators) and emotional exhaustion. High NA Low NA

PAGE 204

204 Figure 5 9 . Plot of the moderating effect of locus o f control on the linear relationship between challenge stressors (modeled as having reflective indicators) and job satisfaction. Internal Locus External Locus

PAGE 205

205 Figure 5 10 . Plot of the moderating effect of locus of control on the linear relationship between challenge stressors (mo deled as having reflective indicators) and organizational citizenship behavior. Internal Locus Ext ernal Locus

PAGE 206

206 Figure 5 11 . Plot of the moderating effect of negative affectivity on the linear relationship between challenge stressors (modeled as having reflective indicators) and c ounterproductive work behavior. High NA Low NA

PAGE 207

207 Figure 5 12 . Plot of the moderating effect of ability on the curvilinear relationship between challenge stressors (modeled as having reflective indicators) and task performance . Low Ability High Ability

PAGE 208

208 Figure 5 13 . Plot of the moderat ing effect of ability on the curvilinear relationship between hindrance stressors (modeled as having reflective indicators) and task performance. Low A bility High A bility

PAGE 209

209 Figure 5 14 . Plot of the moderating effect of locus of control on the curvilinear relationship between h indrance stressors (modeled as having reflective indicators) and task performance. External Locus Internal Locus

PAGE 210

210 Figure 5 15 . Plot of the moderating effect of hindrance stressors on the linear relationship between challenge stressors (modeled as having reflective indicators) and emotional exhaustion. High Hindrance Stressors Low Hindrance Stressors

PAGE 211

211 Figure 5 16 . Plot of the moderating effect of hindrance stressors on the linear relationship between challenge stressors (modeled as having reflective indicators) and frustration. High Hindrance Stressors Low Hindrance Stressors

PAGE 212

212 Table 5 1 . Demographic characteristics of insurance employee s ample (Study 2) Variable M (SD) N % of sample Age (years) 33.7 (10.9) Gender Female 131 79.9 Male 33 20.1 Job tenure (years) 2.61 (3.90) Organizational tenure (years) 5. 17 (5.73) Cognitive a bility (Wonderlic IQ) 108.95 (8.25) Marital status Single 62 37.8 Living with partner 18 10.8 Married 70 42.7 Separated/Divorced 14 8.5 Education level High school/GED 17 9.6 Associates d egree 18 10.8 Some college 59 36.0 College degree 57 34.8 Some graduate school 7 4.3 MBA/Ph.D. 6 3.7 No te: Percentages calculated from a total valid n of 164.

PAGE 213

213 Table 5 2. Variable means, standard deviation s, relia bilities and intercorrelations Variable CS HS EE FR JS TP OCB CWB NA AB LOC Challenge stressors .85 .01 .04 .03 .19* .36** .53** .30** .08 .14 .03 Hindrance stressors .06 .84 .54** .56** .53** .24** .01 .20* .15 .00 .27** Emotional exhaustion .26** .26** .93 .84** .60** .21** .09 .21** .39** .12 .26** Frustration .19** .27** .60** .81 .61** .24** .08 .16* .40** .07 .23** Job satisfaction .02 .23** .31 ** .24** .86 .29** .28** .32** .25** .06 .24 Task performance .20** .09** .07* .06 .25** .92 .46** .41** .00 .02 .03 OCB .26** .02 .00 .04 .16** .32** .87 .29** .07 .06 .08 CWB .19** .09 .07* .01 .10** .20** .13** .75 .08 .03 .16* Negative affectivity .76 .03 .26** Ability .80 .01 Locus of control .61 Note: CS = Challenge stressors; HS = Hindrance stressors; EE = Emotional exhau stion; FR = frustration; JS = Job satisfaction; TP = Task performance; OCB = Organizational citizenship behavior; CWB = counterproductive work behavior; NA = Negative affectivity; AB = Ability; LOC = Locus of control. Correlations above the diagonal repres ent between individ ual (aggregated) scores (n = 164 ). Correlations below the diagonal were estimated by standardizing the regression coefficient obtained from conducting a simple regression in hierarchical linear analysis using each pair of variables (N = 2134). Reliabilities are reported in parentheses on the diagonal. For challenge stressors, hindrance stressors, emotional exhaustion, frustration, job satisfaction, tas k performance, OCB, and CWB these values represent the n weighted average internal consi stency reliability estimates from the Level 1 (daily) surveys. ** p < .01. * p < .05.

PAGE 214

214 Table 5 3. Parameter estimates and variance components of null models for strains, job satisfaction, and performance criteria Variable 00 ) Within individua l 2 ) Between individual 00 ) % of variability within individual Emotional exhaustion 3.39** .72 2.25** 24.2 Frustration 3.47** . 96 1.80** 34.8 Job satisfaction 5.60** .48 1.66** 21.5 Task performance 6.19** .45 .38** 5 4.2 OCB 4.95** .43 1.23** 25.9 CWB 1.30** .17 .20** 45.9 Note: OCB = organizational citizenship behavior. CWB = counterproductive work behavior. 00 = pooled intercept representing the average level of the 2 = within 00 = between individual variance in the variable. Percentage of variability within 2 2 00 ). ** p < .01. Table 5 4. Results of HLM analyses testing linear stressor relationships Variab le Emotional exhaustion Frustration Job satisfaction Task performance OCB CWB B f B r B r B f B f B r B f B r B f B r B f B r Step 1 Intercept 3.39** 3.40** 3.47** 3.48** 5.60** 5.60** 6.20** 6.20** 4.95** 4.94** 1.30** 1.30** Challeng e stressors .07* .29** .08* .22** .03 .03 .05 .12** .11** .17** .01 .07** Hindrance stressors .17** .28** .15** .36** .05* .22** .01 .10* .00 .01 .02 .05* Step 2 Challenge stressors 2 .02 .05** .00 .03 .02 .04** .01 .00 .04* .01 .01 .00 Step 3 Hindrance stressors 2 .02 .05 .03 .05 .01 .05 .01 .07* .00 .04 .00 .01 Note: OCB = organizational citizenship behavior. CWB = counterprod uctive work behavior. B f = unstandardized parameter estimate with stressor constructs modeled as having formative indicators . B r = unstandardized parameter estimate with stressor constructs modeled as having reflective indicators . All predictor scores were The parameter estimates are reported for the step from which they were obtained. ** p < .01. * p < .05.

PAGE 215

215 Table 5 5 . Results of HLM analyses testing interindividual and cross level interactive relationships with emotiona l exhaustion. Parameter Negative affectivity Ability Locus of control B f B r B f B r B f B r Level 1 0 0 (Intercept) 1.52** 1.57** 2.64** 2.56** 5.24** 5.15** 1 0 (Challenge stressors) .07 .07 .13 .37* .08 .55** 2 0 (Hindrance stressors) .10 .24 .02 .05 .02 .33 3 0 (Challenge stressors 2 ) .06 .03 .04 .01 .08 .07 Level 2 01 1 .27** 1.28** .15 .16 .29** .37** 1 1 (Challenge stressors) .00 .27** .04 .01 .00 .04 21 (Hindrance stressors) .04 .03 .04 .01 .04 .01 31 (Challenge stressors 2 ) .03 .05 .01 .05 .02 .01 Note: B f = unstandardized parameter estimate with str essor constructs modeled as having formative indicators . B r = unstandardized parameter estimate with stressor constructs modeled as having reflective indicators .01. * p < .05. Negative affectivity, ability, and locus of control were not centered in these analyses. Table 5 6 . Results of HLM analyses testing interindividual and cross level interactive relationships with frustration. Parameter Negative affectivity Ability Locus of contro l B f B r B f B r B f B r Level 1 0 0 (Intercept) 1.79** 1.74** 3.04** 3.00** 4.89** 4.84** 1 0 (Challenge stressors) .10 .01 .04 .05 .20 .42** 2 0 (Hindrance stressors) .05 .28 .14 .29 .16 .07 3 0 (Challenge stressors 2 ) .03 . 04 .07 .06 .02 .01 Level 2 01 1.14** 1.17** .09 .09 .30** .29** 1 1 (Challenge stressors) .12 .18 .03 .04 .02 .04 21 (Hindrance stressors) .07 .04 .00 .01 .07* .09 31 (Challenge stressors 2 ) .02 .01 .01 .02 .01 .01 Note: B f = unstandardized parameter estimate with stressor constructs modeled as having formative indicators .B r = unstandardized parameter estimate with stressor constructs modeled as having reflective indicators . All predictor scores were centered at the individua .01. * p < .05. Negative affectivity, ability, and locus of control were not centered in these analyses.

PAGE 216

216 Table 5 7 . Results of HLM analyses testing interindividual and cross level interactive relationships with job satisfaction. Paramet er Negative affectivity Ability Locus of control B f B r B f B r B f B r Level 1 0 0 (Intercept) 6.67** 6.62** 6.14** 6.03** 4.21** 4.22** 1 0 (Challenge stressors) .09 .13 .05 .19 .02 .33** 2 0 (Hindrance stressors) .07 .24 .11 . 36* .05 .04 3 0 (Challenge stressors 2 ) .03 .00 .12 .01 .03 .00 Level 2 01 .72* .69** .11 .08 .29** .29** 1 1 (Challenge stressors) .05 .12 .01 .03 .00 .06 * 21 (Hindrance stressors) .01 .01 .01 .03 .02 .04 31 (Challen ge stressors 2 ) .00 .02 .02** .01 .01 .00 Note: B f = unstandardized parameter estimate with stressor constructs modeled as having formative indicators . B r = unstandardized parameter estimate with stressor constructs modeled as having reflective indicato rs .01. * p < .05. Negative affectivity, ability, and locus of control were not centered in these analyses. Table 5 8 . Results of HLM analyses testing interindividual and cross level in teractive relationships with task performance. Parameter Negative affectivity Ability Locus of control B f B r B f B r B f B r Level 1 0 0 (Intercept) 6.18** 6.23** 6.28** 6.16** 5.98** 5.85** 1 0 (Challenge stressors) .01 .22 .11 .08 .03 .08 2 0 (Hindrance stressors) .12 .06 .05 .01* .19 .12 3 0 (Challenge stressors 2 ) .05 .00 .03 .10* .03 .05 40 ( Hindrance stressors 2 ) -.07 -.27* -.41** Level 2 01 .01* .01 .02 .01 .04 .08 1 1 (Challenge stressors) .04 .07 .01 .01 .00 .04 21 (Hindrance stressors) .07 .00 .01 .01 .04 .01 31 (Challenge stressors 2 ) .02 .00 .01 .02* .01 .01 41 ( Hindrance s tressors 2 ) -.01 -.06* -.10** Note: B f = unstandardized parameter estimate with stressor constructs modeled as having formative indicators . B r = unstandardized parameter estimate with stressor constructs modeled as having reflective indicators .01. * p < .05. Negative affec tivity, ability, and locus of control were not centered in these analyses.

PAGE 217

217 T able 5 9 . Results of HLM analyses testing interindividual and cross level interactive relationships with organizational citizenship behavior. Parameter Negative affectivity Abil ity Locus of control B f B r B f B r B f B r Level 1 0 0 (Intercept) 5.20** 5.21** 5.28** 5.25** 4.68** 4.74** 1 0 (Challenge stressors) .05 .20* .07 .04 .03 .18 2 0 (Hindrance stressors) .02 .12 .17* .06 .09 .07 3 0 (Challeng e stressors 2 ) .01 .04 .03 .01 .11* .01 Level 2 01 .19 .19 .07 .07 .05 .04 1 1 (Challenge stressors) .05 .02 .01 .03 .03 .07* 21 (Hindrance stressors) .01 .10 .03 * .02 .02 .01 31 (Challenge stressors 2 ) .03 .04 .01 .01 .0 2 .00 Note: B f = unstandardized parameter estimate with stressor constructs modeled as having formative indicators . B r = unstandardized parameter estimate with stressor constructs modeled as having reflective indicators . All predictor scores were centered .01. * p < .05. Negative affectivity, ability, and locus of control were not centered in these analyses. Table 5 1 0 . Results of HLM analyses testing interindividual and cross level interactive relationships with counterp roductive work behavior. Parameter Negative affectivity Ability Locus of control B f B r B f B r B f B r Level 1 0 0 (Intercept) 1.19** 1.22** 1.19** 1.25** 1.62** 1.62** 1 0 (Challenge stressors) .02 .16** .01 .01 .02 .12 2 0 (Hi ndrance stressors) .01 .06 .05 .11 .11* .13 3 0 (Challenge stressors 2 ) .00 .06* .01 .04 .00 .01 Level 2 01 .07 .05 .02 .01 .07 .07* 1 1 (Challenge stressors) .00 .06* .00 .02 .00 .01 21 (Hindrance stressors) .02 .06 .01 .03 .02 .02 31 (Challenge stressors 2 ) .01 .04 .00 .01 .00 .00 Note: B f = unstandardized parameter estimate with stressor constructs modeled as having formative indicators . B r = unstandardized parameter estimate with stressor constructs modeled as having reflective indicators .01. * p < .05. Negative affectivity, ability, and locus of control were not centered in these analyses.

PAGE 218

218 Table 5 11 . Results of HLM analyses testing interactive cha llenge stressor hindrance stressor relationships. Variable Emotional exhaustion Frustration Job satisfaction Task performance OCB CWB B f B r B f B r B f B r B f B r B f B r B f B r Intercept 3.39** 3.39** 3.47** 3.48** 5.60** 5.60** 6.20** 6.20** 4. 94** 4.93** 1.30** 1.30** Challenge stressors .07* .30** .08* .22** .03 .03 .05 .12** .11** .17** .01 .06** Hindrance stressors .16** .29** .14** .35** .06* .23** .01 .07* .01 .02 .02 .04 Challenge stressors X Hindrance stressors .01 . 15* .01 .12* .01 .04 .01 .00 .03 .03 .01 .03 Note: OCB = organizational citizenship behavior. CWB = counterproductive work behavior. B s = unstandardized parameter estimate with stressor constructs modeled as having formative indicators . B r = unstandardized parameter estimate with stressor constructs modeled as having reflective indicators

PAGE 219

219 CHAPTER 6 DISCUSSION OF STUDY 2 The hypothesis that employee stress h as a curvilinear relationship with outcomes such as job performance is popular in OB theory and textbooks. However, empirical studies testing this hypothesis have provided very limite d support (Muse et al., 2003), leading some researchers to recommend that field tests of this relationship be aba ndoned (Westman & Eden, 1996). My review of this literature suggested that the challenge hindrance stressor framework could be used to explain inconsistent findings regarding the curvilinear stressor effects hypothes is. Specifically, I found that distinguishing between these two forms of work stressors in previous research indicated that studies examining challenge stressors (Champoux, 1980, 1981; Edw ards & Harrison, 1993) were more likely to detect curvilinear relati onships with employee criteria than were studies that focused on the effects of hindrance stressors (Jamal, 1984, 1985 ). However, this review also identified several limitations in the previous research related to the measurement of stressor constructs and research design . Therefore, I conducted a study designed to more appropriately test the curvilinear stressor hypotheses using the newly developed challenge and hindrance stressor measures from Study 1 , a repeated measures design that assessed the reactio ns of employees to daily work stressors over a three week period , and multi level analytic techniques which facilitate the examination of within and between person differences in reactions to work stressors. Specifically, Study 2 tested hypotheses that: ( 1) hindrance stressors have positive, linear relationships with employee strains, and negative, linear relationships with employee job satisfaction and performance; (2) challenge stressors have positive, linear relationships with strains, job satisfaction, and performance; and (3) challenge stressor have curvilinear relationships with these employee criteria; (4) employee negative affectivity, perceived ability, and locus of control have

PAGE 220

220 moderating effects on the linear and curvilinear relationships between stressors and employee outcomes. In the following sections, I discuss the results from analyses testing each set of hypotheses examined in this s tudy. I first discuss the findings from analyses modeling the stressor constructs as having formative indicato rs, and then briefly discuss the results from analysis modeling the stressor constructs as having reflective indicators. Assessment of Hypothesized Linear Effects Generally, the results of Study 2 provide mixed support for the linear relationships between work stressors and employee outcomes. Consistent with previous research (Cavanaugh et al., 2000; LePine et al., 2005; Podsakoff et al., 2007), both challenge and hindrance stressors (when modeled as having formative indicators) were found to be positively related to employee strains in the form of emotional exhaustion and frustration. Therefore, this study generalizes the findings of Study 1 and other cross sectional research ( Boswell et al., 2004; Cavanaugh et al., 2000) to a daily context by providing ad ditional evidence that both challenge and hindrance stressor s are associated employee strains. Also consistent with previous research (Cavanaugh et al., 2000; Podsakoff et al., 2007), hindrance stressors (when modeled as having formative indicators) were found to have a negative relationship with e mployee job satisfaction. Study 2 reported that e mployees who experienced higher levels of role ambiguity and role conflict, perceived higher levels of political decision making , felt they had inadequate resource s to complete their tasks, perceived higher levels of interpersonal conflict, and/or experienced administrative hassles when completing their daily work expressed lower levels of satisfaction with their job on a daily basis. However, contrary to the result s of Study 1 and previous meta analysis (Podsakoff et al., 2007), this study reported no linear relationship between challenge stressors (when modeled as having formative indicators) and employee job satisfaction. That is, the extent to which employees exp erienced

PAGE 221

221 dai /or the level of accountability employees felt at work had no relationship with daily levels of satisfaction experienced by employees. The lack of support for this relationship may be a function of the time frame used in this study. As reported in Study 1, challenge stressors have opposite effects on e mployee strains and emotions which may make it di fficult for employees to evaluate the influence of these countervail ing forces when assessing their daily level of job satisfaction . For example, employees may feel the strain associated with a high workload or job responsibility on a daily basis, but may not simultaneously recognize the growth potentia l associated with ch allenge stressors such as work pace or job responsibility on a daily basis. Alternatively, although employees may recognize that having several different types of tasks to complete on a given day provides them with increased knowledge of the jo b and potent ial development opportunities, they may not immediately experience the exhaustion associated with the accumulation of these types of demands . Taken together, this s uggests that employees may differentially weight the cognitive, affective, and behavioral in puts resulting from challenge stressors experienced at a daily level , versus those experienced over prolonged periods of time . If this is the case, it would explain why positive relationships between challenge stressors and job attitudes reported in previo us cross sectional field research (Boswell et al., 2004; Cavanaugh et al., 2000 ) were not supported in the present study, which focuses on more fine grained daily assessments. Generally speaking, the results presented in Study 2 also do not provide much support for hypotheses regarding the linear relationships between the ch allenge and hindrance stressors and employee performance. C ontrary to expectations, both challenge and hindrance stressors were generally not found to predict the employee performance criteria. The exception to th is finding

PAGE 222

222 was that challenge stressors (when modeled using formative indicator techniques) were found to have a positive, linear relations hip with employee OCB. This suggests that employees exhibited higher levels of OCB on wo rk days in which they experienced higher levels workload, more accountability, higher levels of job complexity, and faster work pace. However, when taken together, the findings provide little support for the expectation that daily challenge and hindrance s tressors (when modeled as having formative indicators) have differential relationships with daily levels of employee job performance. Although it is not clear why the challenge and hindrance stressors were generally unrelated to the performance criteria i n this study, it is possible that the design of this study may provide explanations for these findings. For example, the daily nature of the study may not have capture d P articipants had , on average, several ye ars of tenure with the company, and may have learned to adapt to their work environment fairly well . Thus, large variations in daily work stressors may not be likely over the course of a three week period. Similarly, employees having yea rs of tenure in a particular o rganization probably have a good conception of what behaviors are expected of them on a daily basis. If this is the case, range restriction in both the work stressors and performance measures may have contributed to the lack o f significant findings in this study. Assessment of the Hypothesized Curvilinear Stressor Effects The results of Study 2 also provided very limited support for the curvilinear stressor effects hypotheses. For example, chall enge stressors (when modeled usi ng formative indicator techniques ) were not found to have their hypothesized curvilinear effects on employee strains, job satisfaction, task performance, or CWB. Moreover, although I did find challenge stressors experienced by employees on a daily basis to have a curvilinear relationship with daily levels of employee OCB, the plot of this relationship showed it to be contrary to my expectations (see

PAGE 223

223 Figure 5 1). That is, employees exhibited increasingly higher levels of OCB as their challenge stressors exce eded normative daily levels . One potential explanation for this counterintuitive finding may be that employees experiencing high levels of challenge stressors increasingly engage in extra role behaviors as a form of withdrawal from task related activities. For example, employees who are overloaded, have excessive levels of time pressure, and feel that their jobs are too complex or require too much responsibility may avoid the excessive pressure of their job by assisting others in the organization or engagin g in psychological or physical withdrawal behaviors. Ho wever, the non significant relationships reported in this study between challenge stressors and task performance and challenge stressors and CWB make this explanation unlikely. P erhaps a better explan ation for this finding is that jobs associated with g reater amounts of challenge stressors may require more significant in teraction with other coworke rs and departments in the organization, and therefore , offer more opportunity for employees to help others . J obs that have high workloads, high levels of complexity, and press ure to complete work on time may be more likely to expose individuals to situations or people requiring their help , and increased perceptions of responsibility may facilitate more extra r ole behaviors bec ause employees feel more accountable for organizational outcomes. Whereas OCB research tends to focus on motiva tional or trait based perspectives (Organ , Podsakoff, & MacKenzie , 2006), this expl anation suggests that the performance of OCB by employees maybe a function, at least in part, of the relative opportunity for employees to exhibit these behaviors. O f course, future research w ill have to examine the validity of this explanation in greater detail . In summary, the results of the pres ent study indicate that challenge and hindrance stressors (when modeled as having formative indicators) provided limited support for the hypothesized curvilinear relationships with employee criteria. Some readers may interpret these findings as

PAGE 224

224 additional support for the moratorium on field research testing curvilinear stressor effects recommended by Westman and Eden (1996). However, there may be alterative explanations for the lack of support for the hypothesized curvilinear stressor effects fou nd in this study. For example, it is possible that the homogenous nature of the sample used in this study limited the potential for cur vilinear relationships. T he data for Study 2 were gathered from one branch of an insurance company where the majority o f participant s were employed in one of only a few occupations . P revious research using large heterogenous samples comprised of dozens of different occupations (Edwards & Harrison, 1996 ; Xie & Johns, 1995 ) have been more successful at identifying significant curvilinear stressor effects than research using smaller, more homogenous samples (Jamal, 1984, 1985 ). D ifferent occupations subject employees to different varieties of work related demands . E mployees who work in different occupations, within different industries, an d in different units/departments of an organization are more likely to experience different types of work demands than those who work in the same occupation in the same company. If this is the case, replicating this study using samples that are more hetero genous with respect to their industries, occupations, and locations may provide a better context for identifying significant curvilinear stressor effects. Another explanation for the lack of significant curvilinear stressor criterion relationships is tha t the time period used in this stu dy was too short to capture extreme fluctuations in employee stressors and emotional, psychological, attitudinal, and behavioral responses. Measuring extreme levels of high and low stressors is critical to identifying pote ntial curvilinear effects (Muse et al., 2003) . Typically, ESM studies conducted in OB research take multiple assessments either at the daily (e.g., one survey per day) or intradaily level (e.g., three to eight surveys per day) for a period of two to three weeks (Beal & Weiss, 2003). However, this assessment period may be too

PAGE 225

225 short to capture extreme s in experienced stressors. Therefore, stress researchers need to consider assessing stressors and employee responses at longer intervals, as this ma y allow res earchers to capture a larger variety of fluctuations over a longer period of time. Related to this, it may also be important to identify situations in which extreme levels of stressors and subsequent employee reactions are likely to be prevalent. This is o ften difficult in organizational settings and may be a reason why significant curvilinear stressor eff ects are not typically found in field studies (Muse et al., 2003). Indeed, capturing a full spectrum of extremely high and low levels of employee job dema nds may require identifying specific contexts in which to conduct studies testing these relationships. For example, researchers might increase the potential for identifying curvilinear stressor effects by studying employees in retail or customer service po sitions immediately before, during, and after the holiday shopping season. As another example, r esearchers could also assess stressors levels and reactions of tax agents before, during, and after national tax day. Although not every occupation will have ob vious periods of extreme stressors, being able to identify and capture representative examples may provide a more accurate test of the curvilinear stressor hypothesis. Assessment of the Hypothesized Moderator Effects Generally speaking, neither i nteracti ons between employ ee individual differences (negative affectivity, ability, and locus of control) and challenge and hindrance stressors (when modeled using formative indicator techniques) nor interactions between the challenge and hindrance stressors thems elves (when modeled using formative indicator techniques) were found to significantly influence empl oyee outcomes. However, there were two notable exceptions. First, the relationships between hindrance stressors and employee OCB was found to be more negati ve for employees with higher levels of ability than for employees with lower levels ability.

PAGE 226

226 This finding was counterintuitive, but the plot presented in Figure 5 6 indicates that this effect may be relatively negligible. More importantly , the results re ported in Study 2 indicated that employees who perceived themselves to have lower levels of ability were more likely to display curvilinear trends in the relationship between challenge stressors (when modeled as having formative indicators) and job satisfa ction, than were employees who perceived themselves to have higher levels of ability. This is particularly interesting because it suggests that some individual differences may influence the curvilinear relationships between stressors and employee outcomes. If this is the case , it may offer another potential explanation for the lack of supportive findings for the curvilinear stressor effects hypothesis reported in this study. Specifically, some individual differences m ay make individuals more or less suscept ible to the effects of extremely high or low levels of challenge or hindrance stressors. That is to say that curvilinear relationships between stressors and outcomes may exist, but are a function of specific individual differences. Because this study did n ot find that negative affectivity or locus of control moderated any of the curvilinear relationships between challenge stressors and employee strains, job satisfaction, or job performance, it may be advantageous to examine other personality traits in futur e research. Assessment of the Results Obtained from Formative versus Reflective Indictor Modeling Techniques It may be important to note that the results of Study 2 indicate that a somewhat different set of results were obtained from analyses modeling the stressor constructs using formative versus reflective indicators techniques. Generally speaking, a comparison of the analyses using these two modeling techniques indicated that when the stressor constructs were modeled as having refle ctive indicators: (1) there was stronger support for the linear effects hypotheses, (2)

PAGE 227

227 more curvilinear stressor criterion relationships were identified , and (3) there tended to be more interactive effects between the individual differences and the stressor criterion relation ships. Although the additional findings were not entirely supportive the results may be interpreted by some readers as providing support for the use of reflective indicator techniques when modeling c hallenge and hindrance . Howev er, it is important to note that this interpretation is based on purely empirical, rather than theoretical, arguments. The number of hypotheses supported or the strength of the linear, curvilinear, and moder ating effects cannot be the primary basis for dec iding the nature of the relationships between a construct and its indicators. Instead, this decision must be based on theoretical arguments, which indicate that the dimensions of the challenge and hindrance stressors are more appropriately modeled as formi ng the stressor constructs, and not as reflections of the underlying constructs. Nevertheless, the differential effects observed in this study when the stress ors were modeled as having reflective indicators as opposed to when they are modeled as having for mative indicators is evidence that canno t be ignored. Therefore, I discuss this issue in the following section. Limitations Study 2 has several potential l limitations. The first potential limitation relates to the weighting procedure used to create scale scores for the formative indicator of the challenge and hindrance stressor constructs. In contrast to reflective indicators modeling techniques, which typically sum or average the reflective indicators of a construct, a formative indicator approach should take into account the relative contribution that each indicator makes to the higher order construct. Unfortunately, previous research has not provided guidelines for creating scales scores for constructs having f ormative indicators. In the present study, this was accomplished by obtaining the factor score weights from a CFA conducted on the data gathered from the first daily survey, and then creating stressor scale scores b y wei ghting the indicators using the se

PAGE 228

228 scores. Although this technique helped ensure that data represented a large port ion of the sample and that responses were not biased by previous daily surveys, limitations to t his approach may or immedi ately before this daily survey was administered. In addition, this daily survey may not capture the typical distribution of challenge and hindrance stressors experienced by employees in this organization. Finally, stressor scale scores created in this mann er may also be viewed as post hoc assessments to the extent that they are fo rmed by ratings provided by employees already engaged in the study. However, there are other techniques that could be used to create challenge and hindrance stressor scale scores which are also consistent with f ormative indicator modeling approa ches. For example, the one time (Level 2) survey could be designed to assess the general or characteristic levels of the challenge and hindrance stressors experienced by employees in their j ob. Obviously, this approach may prime participant responses on later daily surveys. Researchers could also att empt to obtain characteristic levels of work stressors for different jobs or organizations using focus groups, interviews, or by surveying employ ees working in contexts similar to that of the focal sample (i.e., same department, office, industry, etc.). For example, it may be possible to use benchmarked data from another insurance office in the southeastern Unites States to obtain the relative weig hts of the dimensions of the challenge and hindrance stressor constructs. T he weights obtained from focus group, interview, survey, or benchmarking data could be used to ponses. However, given that there are no guidelines for how this sho uld be done in our field and that misspecification is prevalent with constructs having formative indicators (Mackenzie et al.,

PAGE 229

229 2005; Podsakoff et al. 2007) , future research should compare these different techniques and provide recommendations on the appropriateness of each one of them. The second limitation has to do with considerations made for the design of this study. As with most ESM designs, the results of this study are potentially s ubject to common method variance from single source biases. This is a concern in most cross sectional studies in which a single rater provides assessments of the independent and dependent variables (Podsakoff et al., 2003a) . It may be particularly p roblema tic to address this concern in ESM designs because of the difficult y involved in obtaining r epeated daily ratings of employee attitudes or behaviors from m anagers or coworkers. However, method biases may be less of a concern in this study for two reasons. First, as noted by Podsakoff et al. (2003), method variance may be somewhat unavoidable to the extent that certain dependent variables are most appropriately measured through self reports. For example, to accurately assess employee strains or satisfaction at work, researchers should gather this information by surveying the individual who directly experience s these psychological or emotional states. Fu ture research may attempt to limit the effect of single source bias in ESM designs by temporally separating the ratings of the stressors from employee responses. Me thod variance due to single source biases may be less of concern in this study because I directly assessed the effects of rater negative affectivity . Rater negative affectivity has been identified by some researchers (Bolger & Zuckerman, 1989; Brief et al., 1988) as an important rater characteristic that may bias the relationships between stressors and strains. However , Spector and his colleagues ( Chen & Spector, 1991; Spector, 1994; Spector et al., 20 00) have argued that NA should not be considered a source of bias, but rather a theoretical ly important construct when assessing the relationship s between work stressors and strains. In the current

PAGE 230

230 study, I have assessed the main effect s of NA on employee strains, satisfaction, and job performance criteria, and the potential moderating effect s this trait on the relationships between employee stressors and outcomes. Therefore, whether one argues that NA is a source of bias or a theoretically important constr uct for understanding stressor criterion relationships, I have assessed and controlled for the effects of this construct in my study. However, given the general l ack of support for the hypotheses in Study 2, this may not be a very comfort ing notion . The t hird potential limitation of this study relates to the considerations made for survey length and participant fatigue. Study 2 included shortened versions of the challenge and hindrance stressor scales and employee emotional exhaustion, task performance, OC B, and CWB because participants were completing similar versions of this survey each day over the course of a three week period . Although, there may be a concern regarding the construct validity of these shortened scales, my analyses indicated that the rep resentative items had strong correlations with the items in the full scales that were not included in all of the daily surveys. Therefore, the modified scales seem to be representative of the more complete scales they were designed to assess. Including add itional items in some of the daily surveys allowed me assess the validity and representativeness of the shortened scales. This may be a method other researchers using ESM designs can employ to help deal with issues related to survey length . Unfortunately, I was unable to assess employee job insecurity in this study. This limitation may be particularly problematic because the hindrance stressor construct is conceptualized as having formative indicators and thus, should be represent ed by a census of its item s . Because job insecurity was argued to be part of the hindrance stressor construct and was not assessed in Study 2, this stressor construct is content deficient. This concern may be compounded by the fact that this particular work stressor was found to ha ve a significant relationship with the second order

PAGE 231

231 hindrance stressor construct in Study 1. Therefore, excluding employee job insecurity in Study 2 may be an important omission that leads to different relative weights for the dimensions of the hindrance s tressors and to an inaccurate assessment of this construct. As a result, future research should include a complete set of the stressor dimensions to determine the replicability of these results and the changes in the weights for the stressor dimensions acr oss studies. Finally, the findings of Study 2 may also be limited by the theory used to develop the differential curvilinear stressor hypotheses tested in this study. I have argued that activation theory explains the differential curvilinear relationships between work stressor and employee criteria such that challenge stressors have curvilinear relationships with outcomes and hindrance stressors have linear relationships. Unfortunately, measuring the intervening mechanism ( i.e., neural arousal in the retic ular activation system) on which this theory is based would require expensive brain scan measures that would undoubtedly results. Therefore, although brain activation represented an unmeasured variable in my study, it may be nearly impossible to obtain measures of this mediator in future research testing the potential curvilinear effects of work stressors . Fortunately, there may be more practical alternatives to activa tion theory as an explanation for curvilinear wor k stressor effects. For example, LePine et al. (2005) have argued a motivational approach based on expectancy theory can be used to explain the differential relationships between challenge and hindrance stressors and employee outcomes. Although this argume nt was used to explain differential linear effects of these work stressors on performance, this research may also provide support for differential curvilinear stressor criterion relationships. LePine et al. have reported that employee strains and motivatio n have offsetting effects on challenge stressor s such that these types of stressors decrease performance through their

PAGE 232

232 relationship with the strains that result from work related demands, but that challenge stressors increase performance through the positi ve relationships that these stressors share with employee motivation. Although their meta analytic data did not allow LePine et al . to assess potential curvilinear relationships, it is possible that employee strains and motivation have countervailing effec ts, such that employee performance decreases at an accelerating rate as challenge stressors increase because, (a) emotional, physical, and psychological strains begin to accumulate as a result of trying to cope with these demands and (b) employee motivatio n decreases because they are less likely to make strong links between successfully coping with demands and receipt of rewards that are valued. This indicates that employee expectancies regarding their ability to meet demands with increased effort, their be lief that meeting demands will result in outcomes, and the v alence associated with the potential outcomes of successful coping (i.e., positive and negative emotions, psychical and psychological strains) could represent the mediating mechanisms that drive t he curvilinear relationships between challenge stressors and employee outcomes. Therefore, despite the fact that the hypotheses of this study re ceived limit ed support in this study, the above discussion suggests that limitation s in the sample and study de sign may contribute to this lack of support. Furthermore, future research using samples in which inter and intra indi vidual differences are magnified and potentially more practical theoretical frameworks may provide support for the notion that challenge s tressors have curvilinear relationships with employee strains, job satisfaction, and job performance, and hindrance stressors have linear relationships with these criteria. Conclusions Study 2 provided evidence that both challenge and hindrance stressors (when modeled using formative indicator techniques) had positive, linear relationships with employee strains, and differential linear relationships with job satisfaction and OCB, but not task performance or

PAGE 233

233 CWB. Although this study provides little support for the curvilinear stressor effects hypotheses, there are several explanations for the lack of support for these relationships. Although the potential moderating effects of employee negative affectivity, ability, and locus of control received only limited support, the moderating effect of ability on the curvilinear relationship between challenge stressors and job satisfaction suggested that these curvilinear relationships may exist, but are at least somewhat a function of individual differences. Finally, a lthough the results of Study 2 varied considerably depending on whether the str essor constructs were modeled using formative or reflective indicator techniques , the formative indicator model is argued to be more appropriate for theoretical reasons.

PAGE 234

234 CHAPTE R 7 IMPLICATIONS AND FUTURE RESEARCH The research presented in this dissertation had two focal purposes. First, Study 1 was designed to d evelop and validate scales of challenge and hindrance stressors in the workplace . This study used a six step construct validation procedure that was designed to assist in the development and assess ment of constructs having formative indicators. The results of this study generally provided support for the construct validity of challenge and hindrance stressor measures , in t hat these scales were found to be content valid, reliable, and to demonstrate convergent, discriminant, and criterion related validity. In addition, this study supported the higher order structure of the stressor constructs , and the modeling of these const ructs as having formative indicators. The second purpose of my research was to ex amine the potential linear, curvilinear, and moderated relationships between these stressors and employee strains, satisfaction, and job performance. Study 2 tested t hese rel ationships using an ESM design in which daily surveys from 164 employees of an insurance branch office were gathered over the course of a three week period. Generally, this study provide s limited support for the hypothesized linear stressor criterion relat ionships, the curvilinear challenge stressor employee criterion relationships, and for the moderating effects of employee negative affectivity, ability, and locus of control on these relationships . Taken together, the results of these two studies have seve ral important theoretical implications . In addition, my research also provides a foundation for future research. In the following sections, I discuss the theoretical implications of my studies, as well as my recommendations for future research.

PAGE 235

235 Theoretica l Contributions Generally, th ere are at least four important theoretical contributions of my research. First, Study 1 provide s theoretical support for the challenge/ hindrance stressor framework. From a conceptual perspective, I have provided clear concept ual definitions for the challenge and hindrance stressor constructs and their respective dimensions . In addition, I have described the promoting or hindering task accomplish ment and personal development. This facilitated the development of stressor measures that are not deficient or contaminat ed by the content of other constructs . Therefore, my research provides a stronger theoretical foundation on which to group the challeng e/hindrance stressor framework, and valid measures of these constructs to assess stressor criterion relationships. Second, m y research also supports the findings of pr evious empirical studies examining the challenge/hindrance stressor framework. Study 1 r eported that although both challenge and hindrance stressors had positive relationships with employee strains, challenge stressors also had positive relationships with job satisfaction and employee motivation, whereas hindrance stressors had negative relat ionships with these criteria. In addition, Study 1 extended the nomolog i cal network of these stressor constructs to include employee emotions. On the one hand, hindrance stressors were found to have positive relationships with employee threat and harm emot ions and negative relationships with challenge and benefit emotions. On the other hand, challenge stressors demonstrated positive relationships with challenge and benefit emotions, and no relationship with harm related emotions. Interestingly, challenge st ressors were found to have a positive relationship with threat related emotions. Taken together these results suggest that employees may tend to perceive the potential for demands such as workload and work pace to result in negative outcomes, but that they do not tend to perceive negative outcomes actually

PAGE 236

236 result from these demands . T herefore, this study provides another step toward understanding the complex relationships between challenge and hindrance stressors, individual appraisal, and subsequent behavi oral reactions. Third, my research contributes to a better understanding of the potential curvilinear relationships between stressors and employee outcomes. Specifically, the results of Study 2 suggest that curvilinear relationships between challenge stre ssors and some employee criterion (OCB) do exist. However, this study also indicated that curvilinear relationships may exist between these work stressors and other employee outcomes, but that they are primarily a function of individual differences. Specif ically, the curvilinear relationship between challenge stressors and j ob satisfaction was found to be significant only when employee ability interacted with the challenge stressors. The plot of this relationship indicated that the daily job satisfaction of employees with lower ability was more susceptible to extreme high and low levels of challenge stressors, than of employees with higher ability. Although, Champoux (1981) and Janssen (2001) have reported that supervisory satisfaction and employee fairness perceptions, respectively, did moderate the curvilinear relationships between job demands and employee outcomes, Study 2 is the first to examine these interactive stressor effects in the context of an ESM study. This is important because a within person re peated measures design: (a) provides for a more appropriate test of the curvilinear stressor effects hypotheses, (b) facilitates the examination of individual differences as moderators of the curvilinear stressor criterion relationships, and (c) has the po tential power to detect both curvilinear effects and moderated curvilinear effects. Therefore, this study provides evidence that curvilinear stressors effects may be a function of employee characteristics and that ESM studies may be an effective way of app ropriately identifying these relationships.

PAGE 237

237 Finally, although not necessarily a theoretical contribution, I have provided a detailed six step program that can be us ed by researchers to develop, refine, and/or validate othe r constructs important to organiz ation research . The process I have described incorporates classic scale validation techniques (cf. Churchill, 1979; Schwab, 1980) and contemporary advancements in this domain, including content validation techniques (Hinkin & Tracy , 1999), confirmatory fac tor analysis (Bollen, 1989), and formative indicators modeling in structural equations (Bollen, 1989; Macken zie et al. , 2005 ). Although the steps in this process can be generalized to constructs having reflective indicators, I have discussed some important considerations for developing and validating constructs that have formative indicators. To my knowledge, this is the first research study that attempts to do this and therefore may serve as a general guide for future researchers who wish to validate or m odel other constructs as having formative indicators. For example, contemporary conceptualizations of job performance (Campbell , 1990; Rotundo & Sackett, 2002) indicate that this construct is comprised of at least three distinct dimensions including task p erformance, extra role activities (e.g., OCBs or contextual performance) , and deviant behaviors (e.g., CWBs). Because these dimensions are performance and are not necessa rily interchangeable with each other, it is more appropriate to model them as formative indicators of a higher order individual performance construct. Similarly, researchers who take a stakeholder approach to strategic management may model measures of fina ncial and social performance as formative indicators of firm performance (Podsakoff et al., 2007). Therefore, this study has important methodological implications for researchers who are interesting in the development, validatio n, and modeling of construct s within (and outside) the OB field.

PAGE 238

238 Future Research The studies presented in this manuscript lay the foundation for several avenues of future research. Foremost, additional research needs to be conducted to better assess the nature of the challenge and h indrance stressor scales. Construct validation is an ongoing process that requires multiple samples taken from varied work contexts. The validation data in Study 1 was gathered from one sample of working MBA students. Thus, al though the preliminary data su pports the validity of the dimensions of both stressor constructs, I recommend that at least two large r samples of data be gathered to test the construct validity of these measures. One of these samples should be comprised of a heter o gen e ous assortment of different occupations which will help limit range restriction in the ratings of the stressors and criteria. However, it may also be important to examine the context specific relationships between the dimensions of the stressor constructs , the constructs t hey form, and the relationships with other criteria. One might expect that the relative relationships between the work stressor dimensio ns with their respective second order constructs to differ in settings emphasizing emotional labor and interpersonal con tact (e.g., position s in health care organizations) versus those emphasizing data analysis and empirical research (e.g., positions in the financial sector). Examination of these relationships across different settings may help illuminate the potential adva ntages of formative versus reflective indicator model ing of the stressor constructs, and test the generalizability of the findings of Study 1. In addition to gathering samples that replicate the findings of Study 1, future research could also address the limitations of Study 2. Bec ause of the nature of data collection in these designs , few ESM studies have been able to avoid issues related to method variance due to single source biases. Futu re research sh ould extend the findings of Study 2 by examining the se relationships in a context that allows for the objective measurement of performance criteria on a

PAGE 239

239 daily basis. For example, call centers or other service settings may provide the opportunity for employee performance, such as customer satisfaction ratings, retail dollars sold, or product quality metrics. Although this data may be more difficult to obtain, researchers who tap this setting will contribute significantly to understanding behavioral res ponses to different forms of stress in the workplace. Future research may also benefit from the examination of additional personality traits as moderators of stressor criterion relationships. This may be important for two reasons. First, the examination of other traits may help researchers identify indiv idual differences that moderate curvilinear stressor criterion relationships and to understand under what conditions these effects are more likely to exist. Second, research examining the role of individual differences in the stress process has been somewhat stagnant over the past decades. Reviews of the organizationa l stress literature conducted by Kahn and Byosiere in 1992 and b y Sonnentage and Frese in 2003 indicate that very similar sets of individual dif ferences ha ve been used in stress research during decade that occurred between these studies. U nfortunately, the conclusions of these r eviews are similar to the findings of this s moderator analyses : mixed and fairly inconclusive. This suggests the need for research testing new theoretical moderators of the relationships between work stressors and strains. To this end, Judge et al. (2000) have suggested that OB researchers examine the role of positive biases in the stress process. Specifically, the se authors argue that individuals high in NA p. 102). There fore, the higher levels of stress and strain experienced by NA individuals may result from a lack of self deception by these individuals. Using data gathered from 214 university

PAGE 240

240 employees, Judge et al. (2000) found support for their arguments and continued by suggesting that researchers should also examine the role of positive affectivity in the stress process. Because challenge stressors are work demands that tend to be appraised at providing the opportunity for growth and personal accomplishment, it may b e more beneficial for researchers to examine the role of positive affectivity (PA) in relationships between these demands and other outcomes. Finally, future research should attempt to identify other individual differences that moderate the relationships b etween hindrance stressors and employee criteria. The results of the HLM analysis indicated that these stressor criterion relationships generally had significant within individual variability, but that NA, ability, and locus of control did not consistently explain differences. Therefore, researchers might examine the potential moderating effect of other traits s uch as employee agreeableness, conscientiousness, or other Big Five traits. Finally, although my research focused on individual differences as mode rators of stressor criterion relationships, it may also be important to examine the moderating effect of environmental characteristics. For example, social support is a characteristic of the work environment that may influence the positive or negative effe cts of challenge and hindrance stressors. E mployees who perceive that their coworker s and/or supervisors support their thoughts, feelings, and behaviors may be more likely to experience the positive effects of challenge stressors , and less likely to experi ence the negative effects of hindrance stressors than employees who perceive less social or supervisory support. Examining this and other characteristics of the work environment using repeated measures designs will allow researchers to better understand th e individuals and environmental factors that influence the employee stress process .

PAGE 241

2 41 LIST OF REFERENCES Abramis, D. J. (1994). Work role ambiguity, job satisfaction, and job performance: Meta analysis and review. Psychological Reports , 75, 1411 1433. All en, V.I. & Greenberger, D.B. (1980). Destruction and perceived control. In A. Baum & J.E. Singer (Eds.), Applications of Personal Control. Hillsdale, NJ: Erlbaum. Alliger, G.N. & Williams, K.G. (1993). Using signal contingent sampling methodology to stud y work in the field: A discussion and illustration task perceptions and mood. Personnel Psychology, 46, 525 549. Amabile, T.M., Conti, R., Coon, H., Lazenby, J., & Herron, M. (1996). Assessing the work environment for creativity. Academy of Management Jo urnal, 39, 1154 1184. Andrew, F.M. & Farris, G.H. (1972). Time pressure and performance of scientists and engineers: five year panel study . Organizational Behavi or and Human Performance 8, 185 200. Bagozzi, R.P. & Yi, (1991). Multitrait multimethod matrices in consumer research . Journal of Consumer Research, 17, 426 439 . Barsky, A., Thoresen, C.J., Warren, D.R., & Caplan, S.A. (2004). Modeling negative affectivity and job stress: A contingency based approach. Journal of Organizational Behavior, 25, 915 936. Baumeister, R.F., Bratslovsky, E., Finkenauer, C., & Vohs, K. D. (2001). Bad is stronger than good. Review of General Psychology, 5, 323 370. Beal, D.J. & Weiss, H.M. (2003). Methods of ecological momentary assessment in organizational research. Organizational Research Methods, 6, 440 464. Beaudoin, L.E., & Edg Nursing Economics, 21, 113. Belton, B. (1999, February 17 th ). Fed chief: Tech advances raise job insecurity. USA Today, p. B2. Bennett, R.J. & Robinson, S.L. (2000). Developmen t of a measure of workplace deviance. Journal of Applied Psychology, 3, 349 360. Bentler, P.M. (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107, 238 246. Bentler, P.M. (1995). EQS structural equations program manual. En cino, CA: Multivariate Software.

PAGE 242

242 Bhagat, R.S., McQuaid, S.J., Lindholm, H., & Segovis, J. (1985). Total life stress: A multimethod validation of the construct and its effects on organizationally valued outcomes and withdrawal behaviors. Journal of Applied Psychology, 70 , 202 214. Biddle, B.J. (1979). Role theory: Expectations, identities, and behaviors. London, England; Academic Press. Bingham, J.B., Boswell, W.R., & Boudreau, J.W. (2005). Job demands and job search among high level managers in the Unit ed States and Europe. Group & Organizational Management, 30, 653 681. Bolger, N. & Zuckerman, A. (1995). A framework for studying personality in the stress process. Journal of Personality and Social Psychology, 69, 890 902. Bollen, K.A. (1984). Multipl e indicators: Internal consistency or no necessary relationship? Quality and Quantity, 18, 377 385. Bollen, K.A. (1989). Structural equations with latent variables. New York, NY: Wiley. Bollen, K.A., & Lennox, R. (1991). Conventional wisdom on measur ement: A structural equation perspective. Psychological Bulletin , 110 , 305 31. Boswell, W.R., Olson Buchanan, J.B., & LePine, M.A. (2004). The relationship between work related stress and work outcomes: The role of felt challenge and psychological strain . Journal of Vocational Behavior, 64 , 165 181. Brannick, M.T. & Levine, E.L. (2002). Job analysis: Methods, research, and applications for human resource management in the new millennium. Thousand Oaks, CA; Sage. Brief, A.P., Burke, M.J., George, J.M ., Robinson, B.S., & Webster, J. (1988). Should negative affectivity remain an unmeasured variable in the study of job stress? Journal of Applied Psychology, 73, 193 198. Brissett, M. & Nowicki, S. Jr. (1973). Internal versus external locus of control of reinforcement and reaction to frustration. Journal of Personality and Social Psychology, 25, 135 144. Brough, P. (2005). A comparative investigation of the predictors of work related psychological well being within police, fire and ambulance workers. Ne w Zealand Journal of Psychology, 34, 127 134. Bryant, F.B. (2000). Assessing the validity of measurement. In L.G. Grimm and P.R. Yarnold (Eds.), Reading and understanding more multivariate statistics (pp. 99 146). Washington, D.C.: American Psychologica l Association. Bryk, A.W., Raudenbush, S.W., & Congdon, R.T. (2001). Hierarchical linear and nonlinear modeling 5.04. Lincolnwood, IL: SSI.

PAGE 243

243 Burke, M.J., Brief, A.P., & George, J.M. (1993). The role of negative affectivity in understanding relations be tween self report stressors and strains: A comment on the applied psychology literature. Journal of Applied Psychology, 78, 402 412. Butterfield, E.C. (1964). Locus of control, test anxiety, reactions to frustrations, and achievement attitudes. Journal o f Personality, 32, 355 370. Cammann, C., Fichman, M., Jenkins, D., & Klesh, J. (1983). Assessing the attitudes and perceptions of organizational members. In S. Seashore, E. Lawler, P. Mirvis, & C. Cammann (Eds.), Assessing organizational change: A guide to methods, measures, and practices (pp. 71 138). New York: John Wiley. Campbell, D.P. (1990). Modeling the performance prediction problem in industrial and organizational psychology. In M.D. Dunnette and L.M. Hough (Eds.), Handbook of Industrial and Org anizational Psychology (Vol. 1, pp. 687 732). Palo Alto, CA: Consulting Psychologists Press. Campbell, D.T., & Fiske, D.W. (1959). Convergent and discriminant validation by the multitrait multimethod matrix. Psychological Bulletin, 56, 81 105. Caplan, R .D., Cobb, S., French, J.R.P., Jr., Harrison, R.V., & Pinneau, S.R., Jr. (1975). Job demands and worker health: Main effects and worker health (HEW publication No. (NIOSH) 75 160). Washington, D.C.: U.S. Government Printing Office. Cavanaugh, M.A., Boswe ll, W.R., Roehling, M.V., & Boudreau, J.W. (2000). An empirical examination of self reported work stress among U.S. managers. Journal of Applied Psychology, 85 , 65 74. Champoux, J.E. (1980). A three sample test of some extensions to the job characteristic s model of work motivation. Academy of Management Journal, 23, 466 478. Champoux, J.E. (1981). The moderating effect of work context satisfactions on the curvilinear relationship between job scope and affective response. Human Relations, 34, 505 515. Ch en, P.Y. & Spector, P.E. (1991). Negative affectivity as the underlying cause of correlations between stressors and strains. Journal of Applied Psychology, 76, 398 407. Chen, P.Y. & Spector, P.E. (1992). Relationships of work stressors with aggression, w ithdrawal, theft and substance abuse: An exploratory study. Journal of Occupational and Organizational Psychology, 65, 177 184. Churchill, G. A., Jr. (1979). A paradigm for developing better measures of marketing constructs. Journal of Marketing Resear ch , 16 , 64 73.

PAGE 244

244 Cohen, P., Cohen, J., Teresi, J., Marchi, M., & Velez, C.N. (1990). Problems in the measurement of latent variables in structural equations causal models. Applied Psychological Measurement, 14, 183 196. Cohen, J., Cohen, P., West, S.G., & Aiken, L.S. (2003). Applied multiple regression/correlation analysis for the behavioral sciences. Mahwah, NJ; Lawrence Erlbaum. Colquitt, J.A., Conlon, D.E., Wesson, M.J., Porter, C.O.L.H., & Ng, K.Y. (2001). Justice at the millennium: A meta analytic r eview of the 25 years of organizational justice research. Journal of Applied Psychology, 86, 425 445. Crocker, L. & Algina, J. (1986). Introduction to classical and modern test theory. New York, NY : Hardcourt Brace. Cronbach, L.J., & Meehl, P.E. (1955) . Construct validity in psychological tests. Psychological Bulletin, 52, 281 302. Cropanzano, R., Rupp, D.E., & Byrne, Z.S. (2003). The relationship of emotional exhaustion to work attitudes, job performance , and organizational citizenships behaviors. Jo urnal of Applied Psychology, 88, 160 169, Csikszentmihalyi, M. & Larson, R. (1987). Validity and reliability of the experience sampling method. Journal of Nervous and Mental Disease, 175, 526 536. Dalal, R.S. (2005). A meta analysis of the relationship between organizational citizenship behavior and counterproductive work behavior. Journal of Applied Psychology, 90, 1241 1255. De Dreu, C.K.W. & Van Vianen, A.E.M. (2001). Managing relationship conflict and the effectiveness of organizational teams. Jo urnal of Organizational Behavior, 22, 309 328. De Dreu, C.K.W. & Weingart, L.R. (2003). Task versus relationship conflict, team performance and team member satisfaction: A meta analysis. Journal of Applied Psychology, 88, 741 749. De Jonge, J. & Schauf eli, W.B. (1998). Job characteristics and employee well being: A test of Journal of Organizational Behavior, 19, 387 407. DeVellis, R.F. (1991). Scale development: Theory and applications . Newbury Park, CA: Sage. Diamantopoulos, A. & Winklhofer, H.M. (2001). Index construction with formative indicators: An alternative to scale development. Journal of Marketing Research, 36, 269 277.

PAGE 245

245 Dohrenwend, B.S., Askenasy, A.R. Krasnoff , L., & Dohrenwend, B.P. (1978). Exemplification of a method for scaling life events: The PERI life events scale. Journal of Health and Social Behavior, 19, 205 229. Donatelle, R. J. & Hawkins, M. J. (1989). Employee stress claims: Increasing implication s for health promotion programming. American Journal of Health Promotion, 3 , 19 25. Easterbrook, J.A. (1959). The effect of emotion on cue utilization and the organization of behavior. Psychological Review, 66, 187 201. Edwards, J.R. & Bagozzi, R.P. (2 000). On the nature and direction of relationships between constructs and measures. Psychological Methods, 5, 155 174. Edwards, J.R. & Cooper, C.R. (1990). The person environment fit approach to stress: recurring problems and some suggested solutions. Jo urnal of Organizational Behavior, 11, 293 307. Edwards, J.R. & Harrison, R.V. (1993). Job demands and worker health: Three dimensional reexamination of the relationships between person environment fit and strain. Journal of Applied Psychology, 78, 628 6 48. Ferris, G.R., & Judge, T.A. (1991). Personnel/human resource management: A political influence perspective. Journal of Management, 17, 447 488. Ferris, G.R. & Kacmar, K.M. (1992). Perceptions of organizational politics. Journal of Management, 18, 93 116. Fields, D.L. (2002). Taking the measure of work: A guide to validated scales for organizational research and diagnosis. Thousand Oaks, CA: Sage. Finkelstein, M.A., & Penner, L.A. (2004). Predicting organizational citizenship behavior: Integrating the functional and role identity approaches. Social Behavior and Personality, 32, 383 398. Fisher, C.D. & Gitelson, R. (1983). A meta analysis of the correlates of role conflict and ambiguity. Journal of Applied Psychology, 68, 320 333. Fiske, D.W. (198 2) Convergent discriminant validation in measurements and research strategies. In D. Brinberg & L.H. Kidder (Eds.), Forms of validity in research (pp. 77 92). San Francisco, CA: Jossey Bass. Fornell, C. & Larcker, D.F. (1981). Evaluating structural equat ion models with unobservable variables and measurement error. Journal of Marketing Research , 18 , 39 50. Folkman, S. & Lazarus, R.S. (1985). If it changes it must be a process: Study of emotion and coping during three stages of a college examination. Journ al of Personality and Social Psychology, 48, 150 170.

PAGE 246

246 Folkman, S. & Lazarus, R.S. (1986). Stress processes and depressive symptomatology. Journal of Abnormal Psychology, 95, 107 113. Folkman, S. & Lazarus, R.S. (1988). Coping as a mediator of emotions. Journal of Personality and Social Psychology, 54, 466 475. Fortunato, V.J., Jex, S.M., & Heinish, D.A. (1999). An examination of the discriminant validity of the strain free negative affectivity scale. Journal of Occupational and Organizational Psychology , 72, 503 522. Fox, S. & Spector, P.E. (1999). A model of work frustration aggression. Journal of Organizational Behavior, 20, 915 931. Fox, S., Spector, P.E., & Miles, D. (2001). Counterproductive work behavior (CWB) in response to job stressors and organizational justice: Some mediator and moderator tests for autonomy and emotions. Journal of Vocational Behavior, 59, 291 309. French, J.R.P., Jr., Caplan, R.D., & Harrison, R.V. (1982). The mechanisms of job stress and strain. New York, NY: Wiley. F ried, Y., & Ferris, G.R. (1987). The validity of the job characteristics model: A review and meta analysis. Personnel Psychology, 40 , 287 322. Friend, K.E. (1982). Stress and performance: Effects of subjective workload and time urgency. Personnel Psycholo gy, 35, 623 633. Fuller, J.A., Stanton, J.M., Fisher, G.G., Spitzmuller, C., Russell, S.S., & Smith, P.C. (2003). A lengthy look at the daily grind: Time series analysis of events, mood, stress, and satisfaction. Journal of Applied Psychology, 88, 1019 1 033. Funderburg, S.A., & Levy, P.E. (1997). The influence of individual and contextual variables on 360 degree feedback system attitudes. Group & Organization Management, 22, 210 235. Gardner, D.G. (1986). Activation theory and task design: An empirica l test of several new predictions. Journal of Applied Psychology, 71, 411 418. Gardner, D.G. & Cummings, L.L. (1988). Activation theory and job design. In B.W. Staw and L.L. Cummings, Research in Organizational Behavior (Vol. 10, pp. 81 122). Greenwich, C N; JAI. Goetzel, R.Z., Anderson, D.R., Whitmer, R.W., Ozminkowski, R.J., Dunn, R.L. & Wasserman, J. (1998). The relationship between modifiable health risks and health care expenditures. Journal of Occupational and Environmental Medicine, 40, 843 854.

PAGE 247

247 Gouldner, A.W. (1960). The norm of reciprocity: A preliminary statement. American Sociological Review, 25, 161 178. Graen, G.B. & Scandura, T.A. (1987). Toward a psychology of dyadic organizing. In B.M. Staw and L.L. Cummings (Eds.), Research in Organiza tional Behavior (Vol., 9, pp. 175 208). Greenberg, E.R. & Canzeroni, C. (1996). Organizational staffing and disability claims (American Management Association Report). New York, NY; American Management Association. Greenberg, J. (1990). Employee theft as a reaction to underpayment equity: The hidden costs of pay cuts. Journal of Applied Psychology, 75, 561 568. Greenhalgh, L. & Rosenblatt, Z. (1984). Job insecurity: Toward conceptual clarity. Academy of Management Review, 9, 438 448. Hackman, J.R., & Oldham, G.R. (1980). Work redesign. Reading, MA: Addison Wesley. Harris, K. & Kacmar, K.M. (2005). Organizational politics. In J. Barling,, E.K. Kelloway, and M.R. Frone (Eds.), Handbook of work stress (pp. 353 374). Thousand Oaks, CA; Sage. Hart, P. M. (1999). Predicting employee life satisfaction: A coherent model of personality, work and nonwork experiences, and domain satisfactions. Journal of Applied Psychology, 84, 564 584. Hart, P.M., Wearing, A.J., & Headey, B. (1994). Perceived quality of lif e, personality, and work experiences: Construct validation of the police daily hassles and uplifts scale. Criminal Justice and Behaviors, 214, 283 311. Hinkin, T.R., & Tracey, J.B. (1999). An analysis of variance approach to content validation. Organizat ional Research Methods, 2, 175 186. Hobfoll, S.E. (1989). Conservation of resources: A new attempt at conceptualizing stress. American Psychologist, 44, 513 524. Hockey, R. (1979). Stress and the cognitive components of skilled performance. In V. Hamilt on & D. Warburton (Eds.), Human stress and cognition (pp. 141 178). London, England; Wiley. Hoffman, S.A., Griffin, M.A., & Gavin, M.B. (2000). The application of hierarchical linear modeling to organizational research. In K.J. Klein & S.Kozlowski (Eds.) , Multilevel theory, research, and methods in organizations: Foundations, extensions, and new directions (pp. 467 511). San Francisco, CA: Jossey Bass. Hu, L. & Bentler, P.M. (1998). Fit indices in covariance structural modeling: Sensitivity to underpara metized model misspecification. Psychological Methods, 3, 424 453.

PAGE 248

248 Hui, C., Organ, D.W., Crooker, K. (1994). Time pressure, Type A syndrome, and organizational citizenship behavior: A laboratory experiment. Psychological Reports, 75, 199 208. Hulin, C. L. (2002). Lessons from industrial and organizational psychology. In J.M. Brett and F. Drasgow (Eds.), The psychology of work: Theoretically based empirical research (pp. 3 22). Mahwah, NJ; Lawrence Erlbaum. Humphreys, M.S. & Revelle, W. (1984). Personal ity, motivation, and performance. A theory of the relationship between individual differences and information processing. Psychological Review, 91, 153 184. Hunter, J.E. & Hunter, R.F. (1984). The validity and utility of alternative predictors of job per formance. Psychological Bulletin, 96, 72 98. Ilies R. & Judge, T.A. (2002). Understanding the dynamic relationships among personality, mood, and job satisfaction : A field experience sampling study . Organizational Behavior and Human Decision Processes, 89, 1119 1139. International Personality Item Pool: A Scientific Collaboratory for the Development of Advanced Measures of Personality Traits and Other Individual Differences (http://ipip.ori.org/). Internet Web Site. Ito, T.A., Larsen, J.T., Smith, N.K., & Cacioppo, J.T. (1998). Negative information weighs more heavily on the brain: The negativity bias evaluative categorization. Journal of Personality and Social Psychology, 75, 887 900. Ivancevich, J.M. & Matteson, M.T. (1983). Stress diagnostic survey. Houston, TX: Stress Research Systems. Jackson, C.L., Colquitt, J.A., Wesson, M.J., & Zapata Phelan, C.P. (2006). Psychological collectivism: A measurement vali dation and linkage to group member performance. Journal of Applied Psychology, 91, 884 899. Jackson, S.E. & Schuler. R.W.S. (1985). A meta analysis and conceptual critique of research on role ambiguity and role conflict in work settings. Organizational Be havior and Human Decision Processes, 36, 16 78. Jamal, M. (1984). Job stress and performance controversy: An empirical assessment. Organizational Behavior and Human Performance, 33, 1 21. Jamal, M. (1985). Relationships of job stress to job performance c ontroversy: A study of managers and blue collar workers. Human Relations, 38, 409 424. Janssen, O. (2001). Fairness perceptions as a moderator in the curvilinear relationships between job demands and job performance and job satisfaction. Academy of Manage ment Journal, 44, 1039 1050.

PAGE 249

249 Jehn, K.A. (1994). Enhancing effectiveness: An investigation of advantages and disadvantages of value based intragroup conflict. International Journal of Conflict Management, 5, 223 238. Jehn, K.A. (1995). A multimethod exa mination of the benefits and detriments of intragroup conflict. Administrative Science Quarterly, 40, 256 282. Jehn, K. (1997). Qualitative analysis of conflict types and dimensions in organizational groups. Administrative Science Quarterly, 42, 530 557. Jex, S.M. (1998). Stress and job performance: Theory, research, and implications for managerial practice. Thousand Oaks, CA: Sage. Jex, S.M. & Gudanowski, D.M. (1992). Efficacy beliefs and work stress: An exploratory study. Journal of Organizational Be havior, 13, 509 517. Jones & Bartlett (1995). Stress management (National Safety Council Report). Boston, MA; Authors. Judge, T.A. , Erez, A. & Thoresen, C.J. (2000). Why negative affectivity (and self deception) should be included in job stress researc h: bathing the baby with the bath water. Journal of Organizational Behavior, 22 , 101 111 . Judge, T.A. & Hulin, C.L. (1993). Job satisfaction as a reflection of disposition: A multiple source causal analysis. Organizational Behavior and Human Decision Proc esses, 56, 388 421. Judge, T.A. & Ilies, R. (2004). Affect and job satisfaction: A study of their relationship at work and at home . Journal of Applied Psychology , 89, 661 673. Judge, T.A., Scott, B.A., & Ilies, R. (2006). Hostility, job attitudes, and workplace deviance: Test of a multilevel model. Journal of Applied Psychology, 91, 126 138. Judge, T.A. , Thoresen, C.J. , Bono, J.E., & Patton, G.K. (2001). The job satisfaction job performance relationship: A qualitative and quantitative review . Psychological Bulletin, 127, 376 407. Kacmar, M. & Ferris, G.R. (1991). Perceptions of organizational politics scale (POPS): Development and construct validation . Educational and Psychological Measurement, 51, 193 205. Kahn, R.L. & Byo siere, P. (1992). Stress in organizations. In M.D. Dunnette and L.M. Hough (Eds.), Handbook of industrial and organizational psychology (Vol. 3, pp. 571 649). Palo Alto, CA: Consulting Psychologists Press, Inc.

PAGE 250

250 Kahn, R.L., Wolfe, D.M., Quinn, R.P., Snoek , J.D., & Rosenthal, R.A. (1964). Occupational stress: Studies in role conflict and ambiguity. New York, NY: Wiley. Kanner, A.D., Coyne, J.C., Schaefer, C., & Lazarus, R.S. (1981). Comparison of two modes of stress measurement: Daily hassles and uplifts v ersus major life events. Journal of Behavioral Medicine, 4, 1 39. Klapp, O.E. (1969). Collective search for identity. New York, NY; Holt. Koeske, G.F., Kirk, S.A., & Koeske, R.D. (1993). Coping with job stress: Which strategies work best. Journal of O ccupational and Organizational Psychology, 66, 319 355. Kuhn, A. (1974). The logic of social systems. San Francisco, CA: Jossey Bass. Latack, J.C. (1986). Coping with job stress: Measures and future directions for scale development. Journal of Applied P sychology, 71, 377 385. Law, K.S. , Wong, C.S., & Mobley, W.H. (1998). Toward a taxonomy of multidimensional constructs . Academy of Management Review, 23, 741 75 5. Lazarus, R.S. (1993). From psychological stress to the emotions: A history of changing outlooks. Annual Review of Psychology, 44, 1 21. Lazarus, R.S. & Folkman, S. (1984). Stress, appraisal, and coping. New York, NY: Springer Publishing. Lee, K. & A llen, N.J. (2002). Organizational citizenship behavior and workplace deviance: The role of affect and cognitions. Journal of Applied Psychology, 87, 131 142. Lee, T.W., Mitchell, T.R., Sablynski, C.J., Burton, J.P., & Holtom, B.C. (2004). The effects of job embeddedness on organizational citizenship, job performance, volitional absences, and voluntary turnover. Academy of Management, 47, 711 722. Leiter, M.P. & Maslach, C. (1988). The impact of interpersonal environment on burnout and o rgan i z ational com mitment. Journal of Organizational Behavior, 9, 297 308. LePine, M.A., LePine, J.A., & Jackson, C.L. (2004). Challenge and hindrance stress: Relationships with exhaustion, motivation to learn, and learning performance. Journal of Applied Psychology, 89, 883 891. LePine, J.A., Podsakoff, N.P., & LePine , M.A. (2005). A meta analytic test of the challenge stressor hindrance stressor framework: An explanation for inconsistent relationships among stressors and performance. Academy of Management Journal, 48, 7 64 775.

PAGE 251

251 Locke, E.A. (1976). The nature and causes of job satisfaction. In M.D. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 1297 1349). Chicago, IL: Rand McNally. MacCallum, R.C. & Browne, M.W. (1993). The use of causal indi cators in covariance structure models: Some practical issues. Psychological Bulletin , 114 , 533 541. MacKenzie, S.B., Podsakoff, P.M., & Jarvis, C. (2005). The problem of measurement model misspecification in behavioral and organizational research and some recommended solutions. Journal of Applied Psychology . 90, 710 730. Marino, S. (1997). The stress epidemic. Industry Week, 246, 14. Maslach, C. & Jackson, S.E. (1986). Maslach Burnout Inventory Manual. Palo Alto, CA; Consulting Psychologists Press. McC auley, C.D., Ruderman, M.N., Ohlott, P.J., & Morrow, J.E. (1994). Assessing the developmental components of managerial jobs. Journal of Applied Psychology, 79, 544 560. McCormick, E. J. (1979). Job analysis: Methods and applications . New York, NY: Amacon . McFarlin, D.B., & Rice, R.W. (1992). The role of facet importance as a moderator in the job satisfaction process. Journal of Organizational Behavior, 13 , 41 54. McLain, D.L. (1995). Responses to health and safety risk in the work environment. Academy of Management Journal, 38, 1726 1746. Mitchell, T.R., Holtom, B.C., Lee, T.W., Sablynski, C.J., & Erez, M. (2001). Why people stay: Using job embeddedness to predict voluntary turnover. Academy of Management Journal, 44, 1102 1121. Morrison, E.W. & Phel ps, C.C. (1999). Taking charge at work: Extra role efforts to initiate workplace change. Academy of Management Journal, 42, 403 419. Moyle, P. (1995). The role of negative affectivity in the stress process: Tests of alternative models. Journal of Organiz ational Behavior, 16, 647 668. Muse, L.A., Harris, S.G., & Feild, H.S. (2003). Has the inverted U theory of stress and job performance had a fair test? Human Performance, 16, 349 364. National Institute for Occupational Health and Safety (NIOSH). (1999) . Stress . . . at work. (DHHS Publication No. 99 101). Cincinnati, OH: Author. Neiss, R. (1988). Re conceptualizing arousal: Psychobiological states in motor performance. Psychological Bulletin, 103, 345 366.

PAGE 252

252 Nunnally, J.C. & Bernstein, I. (1994). Psych ometric theory (3 rd ed.). New York, NY: McGraw Hill. Oldham, G.R. & Cummings, A. (1996). Employee creativity: Personal and contextual factors at work. Academy of management Journal, 39, 607 634. Organ, D.W. Podsakoff, P.M. & MacKenzie, S.B. (2006). Organ izational citizenship behavior: Its nature, antecedents, and consequences. Thousand Oaks, CA; Sage. Parkes, J. (1990). Coping, negative affectivity, and the work environment: Additive and interactive predictors of mental health. Journal of Applied Psycho logy, 75, 399 409. Payne, R. (1988). A longitudinal study of the psychological well being of unemployed men and the mediating effect of neuroticism. Human relations, 41, 119 138. comes: The influences of a frequently overlooked construct. Academy of Management Review, 3, 391 397. situational variables relevant to performance settings . Organizational Behavior and Human Performance, 25, 76 96. Peterson, R.A. (2000). Constructing effective questionnaires. Thousand Oaks, CA: Sage. ] Piccolo, R.F., & Colquitt, J.A. (2006). Transformational leadership and job behaviors: The mediating rol e of core job characteristics. Academy of Management Journal, 49, 327 340. Podsakoff, N.P., LePine, J.A., & LePine, M.A. (2007). Differential challenge stressor hindrance stressor relationships with job attitudes, turnover intentions, turnover, and withdr awal behavior: A meta analysis. Journal of Applied Psychology, 92, 438 454. Podsakoff, N.P., Rich, B.L., Saul, J.R., & LePine, J.A. (Under review). Job demands, situational strength, and employee performance: The moderating effect of challenge and hindra nce stressors on agreeableness performance criteria relationships. Podsakoff, N.P. Shen, W., & Podsakoff, P.M. (2006). The role of formative measurement models in strategic management research: Review, critique, and implications for future research. In D .J. Ketchen and D.D. Bergh (Eds.), Research methods in strategy and management. (Vol. 3, pp. 201 256). Greenwich, CN: JAI. Podsakoff, P.M., & Mackenzie, S.B. (1996). An examination of the psychometric properties and nomological validity of some revised a nd reduced substitutes for leadership scales. Journal of Applied Psychology, 79, 702 713.

PAGE 253

253 Podsakoff, P.M., MacKenzie, S.B., Lee, J.Y., & Podsakoff, N.P. (2003). Common method biases in behavioral research: A critical review of the literature and recommend ed remedies. Journal of Applied Psychology, 88, 879 903. Podsakoff, P.M., MacKenzie, S.B., Podsakoff, N.P., Lee, J. (2003). The mismeasure of man(agement) and its implications for leadership research. L eadership Quarterly, 14, 615 656. Podsakoff, P.M., M acKenzie, S.B., Paine, J.B & Bachrach, D.G. (2000). Organizational citizenship behaviors: A critical review of the theoretical and empirical literature and suggestions for future research. Journal of Management, 26, 513 563. Probst, T. M. (2005). Economic stressors. In J. Barling,, E.K. Kelloway, and M.R. Frone (Eds.), Handbook of work stress (pp. 267 297). Thousand Oaks, CA; Sage. Randall, M.L., Cropanzano, R., Bormann, C.A., & Birjulin, A. (1999). Organizational politics and organizational support as p redictors of work attitudes, job performance, and organizational citizenship behaviors. Journal of Organizational Behavior, 20, 159 174. Rizzo, J.R., House, R.J., & Lirtzman, S.I., (1970). Role conflict and ambiguity in complex organizations. Administrat ive Science Quarterly, 15, 150 163. Robinson, S.L. & Bennett, R.J. (1995). A typology of deviant workplace behaviors: A multidimensional scaling study . Academy o f Management Journal, 38, 555 572. Rotter, J.B. (1966). Generalized expectancies for internal versus external control of reinforcement. Psychological Monograph General and Applied, 80, (Whole of No. 609). Rotundo, M. & Sackett, P.R. (2002). The relative importance of task, citizenship, and counterproductive performance to global ratings of job performance: A policy capturing approach. Journal of Applied Psychology, 87, 66 80. Rozin, P. & Royzman, E.B. (2001). Negativity bias, negativity dominance, and c ontagion. Personality and Social Psychology Review, 5, 296 320. Sandman, B.A. (1992). The measurement of job stress: development of the Job Stress Index. In C.J. Cranny, P.C. Smith, and E.F. Stone (Eds.), Job satisfaction: How people feel about their job s and how it affects their performance (pp. 241 254). New York, NY: Lexington Books. Sarason, I.G. & Johnson, J.H. (1979). Life stress, organizational stress, and job satisfaction. Psychological Reports, 44, 75 79. Schaubroeck, J., Cotton, J.L., and Je nnings, K.R. (1989). Antecedents and consequences of role stress: A covariance structure analysis. Journal of Organizational Behavior, 10, 35 58.

PAGE 254

254 Schaubroeck, J., Ganster, D.C., Fox, M.L. (1992). Dispositional affect and work related stress. Journal of A pplied Psychology, 77, 322 335. Scheck, C.L., Kinicki, A.J., & Davy, J.A. (1995). A longitudinal study of a multivariate model of the stress process using structural equations modeling. Human Relations, 48, 1481 1510. Scheck, C.L., Kinicki, A.J., & Davy , J.A. (1997). Testing the mediating processes between work stressors and subjective well being. Journal of Vocational Behavior, 50, 96 123. Schwab, D.P. (1980). Construct validity in organizational behavior. In B.M Staw and L.L. Cummings (Eds.), Research in Organizational Behavior , 2 , 3 43. Scott, W.E., Jr. (1966). Activation theory and task design. Organizational Behavior and Human Performance, 1, 3 30. Shalley, C.E., Zhou, J., & Oldham, G.R. (2004). The effects of personal and contextual characteris tics on creativity: Where should we go from here? Journal of Management, 30, 933 958. Society for Human Resource Management (2001). Layoffs and job security survey. Alexandria, VA; Author. Siegall, M., & McDonald, T. (1995). Focus of attention and empl oyee reactions to job change. Journal of Applied Social Psychology, 25, 1121 1141. Sonnentag, S. & Zijlstra, F.R.H. (2006). Job characteristics and off job activities as predictors as need of recovery, well being, and fatigue. Journal of Applied Psycholog y, 91, 330 350. Sonnentag, S. & Frese, M. (2003). Stress in Organizations. In W.C. Borman, D.R. Ilgen, and R.J. Klimoski (Eds.), Handbook of psychology (Vol. 12, pp. 453 491). Hoboken, NJ; Wiley. Spector, P.E. (1992). Summated rating scales: An introd uction. Newbury Park, CA: Sage. Spector, P.E. (1994). Using self report questionnaires in OB research: A comment on the use of a controversial method. Journal of Organizational Behavior, 15, 385 392. Spector, P.E., Dwyer, D.J., & Jex, S.M. (1988). Relat ion of job stressor to affective, health, and performance outcome: A comparison of multiple data sources. Journal of Applied Psychology, 73, 11 19. Spector, P.E. & Jex, S.M. (1998). Development of four self report measures of job stressors and strain: In terpersonal conflict at work scale, organizational constraints scale, quantitative workload inventory, and physical symptoms inventory. Journal of Occupational Health Psychology, 3, 356 367.

PAGE 255

255 Spector, P.E., Zapf, D., Chen, P.Y., & Frese, M. (2000). Why ne gative affectivity should be Journal of Organizational Behavior, 21, 79 95. Steiger, J.H. & Lind, J.C. (1980, May). Statistically based tests for the number of common factor s. Paper presented at the annual meeting of the Psychometric Society, Iowa City, IA. Storms, P.L. & Spector, P.E. (1987). Relationships of organizational frustration with reported behavioral reactions: the moderating effect of locus of control. Journal of Occupational Psychology, 60, 227 234. Sullivan, S.E. & Bhagat, R.S. (1992). Organizational stress, job satisfaction, and job performance: Where do we go from here? Journal of Management, 18, 353 374. Sverke, M., Hellgren, J., & Naswall, K. (2002). No security: A meta analysis and review of job insecurity and its consequences. Journal of Occupational Health Psychology, 7, 242 264. The Dream in Danger. (1994, November 29 th ). The Wall Street Journal, p. B1. 90). New York, NY; Research and Forcasts, Inc. Thompson, H.B., & Werner, J.M. (1997). The impact of role conflict/facilitation on core and discretionary behaviors: Testing a mediated model. Journal of Management, 23, 583 602. Torangeau, R., Rips, L.J., & Rasinski, K. (2000). The psychology of survey response. Cambridge, England: The Cambridge University Press. Tucker, L.R. & Lewis, C. (1973). A reliability coefficient for maximum likelihood factor analysis. Psychometrika, 38, 1 10. Turner, R.J. & Whe aton, B. (1995). Checklist measurement of stressful life events. In S. Cohen, R.C. Kessler, and L.U. Gordon (Eds.), Measuring stress: A guide for health and social scientists (pp. 29 58). Oxford, England; Oxford University Press. Van Dyne, L. & LePine, J .A. (1998). Helping and voice extra role behaviors: Evidence of construct and predictive validity. Academy of Management Journal, 41, 108 119. Venkatraman, N. & Grant, J.H. (1986). Construct measurement in organizational strategy research: A critique and proposal. Academy of Management Journal, 11, 71 87. Watson, D., Clark, L.A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54, 1063 107 0. Wernerfelt, B. (1984). A resource based view of the firm. Strategic Management Journal, 5, 171 180.

PAGE 256

256 Westman, M. & Eden, D. (1996). The inverted U relationship between stress and performance: A field study. Work & Stress, 10, 165 173. Williams, L.J . & Anderson, S.E. (1991). Job satisfaction and organizational commitment as predictors of organizational citizenship and in role behaviors. Journal of Management, 17, 601 617. Witt, L.A. (1998). Enhancing organizational goal congruence: A solution to or ganizational politics. Journal of Applied Psychology, 83, 666 674. Wright, T.A. & Bonett, D.G. (1997). The contribution of burnout to work performance. Journal of Organizational Behavior, 18, 491 499. Xie, J.L. & Johns, G. (1995). Job scope and stress: C an job scope be too high? Academy of Management Journal, 38, 1288 1309. Yerkes, R.M., & Dodson, J.D. (1908). The relation of strength of stimulus rapidity of habit formation. Journal of Comparative Neurology and Psychology, 18, 459 482. Zapata Phelan, C . P. , Hurst, C. & Colquitt, J.A. (Under review). Challenge stressors, hindrance stressors, and core self evaluations: Main and interactive effects on discretionary performance behaviors. Personnel Psychology. Zohar, D. (1999). When things go wrong: The ef fects of daily work hassles on effort, exertion and negative mood. Journal of Occupational and Organizational Psychology, 72, 265 283.

PAGE 257

257 BIOGRAPHICAL SKETCH Nathan Philip Podsakoff was born on August 30th, 1978 in Bloomington, Indiana. An only child, Natha n was conceived during the great blizzard of the previous winter. When he asked hi most of his youth in Bloomington, he spent approximately 7 months i n Fountain e ble a u, Fr ance when he was 13 years old. Following this time abroad, he enrolled in Bloomington High School South (Bloomington, IN) a nd graduated in 1997. He earned his B.S. in Psychology from Princeton University (Princeton, NJ) in 2001. Nathan participated in foot ball at Princeton earning honors as a Second Team All Ivy league d efensive end his senior year. O n completion of his Ph.D. program, Nathan will begin an assistant professorship at the University of Arizona in Tucson . This position will move him significant ly closer to over 40 of his relatives who live in s outhern California.