Citation
A Formal Model of Leadership in Team Goal Pursuit

Material Information

Title:
A Formal Model of Leadership in Team Goal Pursuit Team Design, Team Composition, and Dynamic Leader Regulatory Processes
Creator:
Zhou, Le
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (13 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Business Administration
Management
Committee Chair:
WANG,MO
Committee Co-Chair:
KAMMEYER-MUELLER,JOHN DANIEL
Committee Members:
BONO,JOYCE
LEITE,WALTER LANA
Graduation Date:
8/9/2014

Subjects

Subjects / Keywords:
Comparators ( jstor )
Computational modeling ( jstor )
Conscientiousness ( jstor )
Leadership ( jstor )
Learning ( jstor )
Perceptual learning ( jstor )
Problem solving ( jstor )
Sensory perception ( jstor )
Statistical bias ( jstor )
Time management ( jstor )
Management -- Dissertations, Academic -- UF
leadership -- team
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Business Administration thesis, Ph.D.

Notes

Abstract:
In this dissertation, I examine the role of leadership in the team goal pursuit process. Team is a basic organizational unit used to achieve collective goals. Theory and research have identified leadership as one of the most influential forces in setting directions for teams and overseeing teams while they strive to achieve the goals. Most of the existing team leadership studies have examined static team leader attributes and behaviors as predictors of team performance. This research is limited in terms of (a) understanding the dynamics in team leadership, (b) explaining why some team leadership functions arise as reactions to the team performance and the task environment, and (c) taking the effects of team composition and team design factors into account. Therefore, I propose and test an integrated formal model on team leadership, drawing on self-regulation theory and functional leadership theory. In Study 1, a formal computational model was specified. Qualitative model evaluation suggested that computational modeling results were consistent with previous empirical findings. In Study 2, an experimental lab study was conducted to test hypotheses derived from the computational model. Supporting the theoretical model, Study 2 showed that over time leader spent more time regulating a subordinate when his/her relative discrepancy was larger comparing to other team members and this relationship was stronger when team task structure was functional rather than divisional or disjunctive. Study 2 also showed that differences across team positions in task difficulties were positively related to the differentiation in leader time allocation across subordinates. In addition, differences among subordinates in learning ability were positively related to the differentiation in leader time allocation across subordinates, and this positive relationship was stronger when team task structure was functional rather than divisional or disjunctive. Study 3 further tested the computational model with data from a field sample. Study 3 found that team task interdependence strengthened the positive effects of differences across team positions in knowledge characteristics and differences across team members in learning ability and conscientiousness on average and differentiated individual-focused leadership behaviors. Theoretical and practical implications from these three studies are discussed. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2014.
Local:
Adviser: WANG,MO.
Local:
Co-adviser: KAMMEYER-MUELLER,JOHN DANIEL.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2015-08-31
Statement of Responsibility:
by Le Zhou.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Embargo Date:
8/31/2015
Resource Identifier:
968131352 ( OCLC )
Classification:
LD1780 2014 ( lcc )

Downloads

This item has the following downloads:


Full Text



PAGE 1

A FORMAL MODEL OF LEADERSHIP IN TEAM GOAL PURSUIT: TEAM DESIGN, TEAM COMPOSITION, AND DYNAMIC LEADER REGULATORY PROCESSES By LE ZHOU A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2014

PAGE 2

© 2014 Le Zhou

PAGE 3

To my parents and my committee

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank Mo Wang, Joyce B o no, John Kammeyer Mueller, Walter Leite , Jeff Vancouver, and Leslie DeChurch for their invaluable feedback on this research. This dissertation was awarded the 2013 Meredith P. Crawford Fellowship by the Human Resources Research Organization (HumRRO) and th e 2014 Mary L. Tenopyr graduate student scholarship by the Society for Industrial Organizational Psychology (SIOP).

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ............... 4 LIST OF TABLES ................................ ................................ ................................ ........................... 8 LIS T OF FIGURES ................................ ................................ ................................ ......................... 9 ABSTRACT ................................ ................................ ................................ ................................ ... 11 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .................. 13 2 LITE RATURE REVIEW ................................ ................................ ................................ ....... 20 Team Leadership ................................ ................................ ................................ .................... 20 Functional Leadership Theory ................................ ................................ ......................... 21 Transformational Leadership Theory ................................ ................................ .............. 23 Leader Member Exchange Theory ................................ ................................ .................. 26 Leadership for Self managing Teams ................................ ................................ ............. 29 Summary ................................ ................................ ................................ .......................... 30 Team Composition ................................ ................................ ................................ .................. 31 Team Composition Constructs ................................ ................................ ........................ 32 Team Composition and Team Performance ................................ ................................ .... 33 Team Design ................................ ................................ ................................ ........................... 34 Team Task Characteristics ................................ ................................ .............................. 34 Team Structure ................................ ................................ ................................ ................ 36 Task Environment Uncertainty ................................ ................................ ........................ 38 Joint Effects of Team Le adership, Team Composition, and Team Design: Previous Findings ................................ ................................ ................................ ............................... 40 Team Composition and Team Leadership ................................ ................................ ....... 40 Team Design and Team Lea dership ................................ ................................ ................ 41 Joint Effects of Team Composition, Team Design, and Team Leadership ..................... 43 3 A FORMAL MODEL OF TEAM LEADERSHIP ................................ ................................ . 48 Team Development System ................................ ................................ ................................ .... 51 Team Member regulated Learning System ................................ ................................ ..... 51 Team Leader regulated Learning System ................................ ................................ ....... 52 Team Performance System ................................ ................................ ................................ ..... 54 Team Member regulated Performance System ................................ ............................... 54 Team Leader regulated Performance System ................................ ................................ . 58 The Development of Dyadic Relationship ................................ ................................ ............. 60 Ass umptions of the Model ................................ ................................ ................................ ...... 61

PAGE 6

6 4 STUDY 1: COMPUTATIONAL MODEL SPECIFICATION AND EVALUATION ......... 75 Computational Model Specification ................................ ................................ ....................... 75 Qualitative Model Evaluation ................................ ................................ ................................ . 76 Effect of Disturbance Magnitude ................................ ................................ .................... 76 E ffect of Disturbance Duration ................................ ................................ ....................... 77 Effect of Leader Reactivity ................................ ................................ ............................. 78 Interaction between Leader Reactivity and Team Task Structure ................................ ... 79 Interaction between Team Composition in Learning Ability and Team Task Structure ................................ ................................ ................................ ....................... 81 Sensitivity Analyses ................................ ................................ ................................ ................ 83 Discussion ................................ ................................ ................................ ............................... 83 5 STUDY 2: HYPOTHESES TESTING USING A LAB EXPERIMENT .............................. 99 Method ................................ ................................ ................................ ................................ .. 103 Participants ................................ ................................ ................................ .................... 103 Task ................................ ................................ ................................ ............................... 103 Procedure ................................ ................................ ................................ ....................... 105 Manipulations ................................ ................................ ................................ ................ 106 Analytic Strategy ................................ ................................ ................................ ........... 110 Results ................................ ................................ ................................ ................................ ... 111 Relationships between Discrepancy and Leader Time Allocation ................................ 111 Effects of Manipulated Factors on Leader Time Allocation ................................ ......... 113 Supplementary Analyses ................................ ................................ ............................... 116 Quantitative Model Evaluation ................................ ................................ ...................... 120 Discussion ................................ ................................ ................................ ............................. 123 Theoretical Implications ................................ ................................ ................................ 123 Practical Implications ................................ ................................ ................................ .... 125 Limitations and Future Research Directions ................................ ................................ . 127 6 STUDY 3: HYPOTHESES TESTING USING A FIELD SAMPLE ................................ ... 145 Method ................................ ................................ ................................ ................................ .. 147 Sam ple and Procedure ................................ ................................ ................................ ... 147 Measures ................................ ................................ ................................ ........................ 149 Confirmatory Factor Analyses ................................ ................................ ...................... 152 Analytic Strategy ................................ ................................ ................................ ........... 153 Results ................................ ................................ ................................ ................................ ... 154 Discussion ................................ ................................ ................................ ............................. 158 Theoretical Impli cations ................................ ................................ ................................ 159 Practical Implications ................................ ................................ ................................ .... 162 Limitations and Future Research Directions ................................ ................................ . 162 7 GENERAL DISCUSSION ................................ ................................ ................................ ... 177 Summary of Key Findings ................................ ................................ ................................ .... 177

PAGE 7

7 Theoretical Implications ................................ ................................ ................................ ....... 179 Practical Implications ................................ ................................ ................................ ........... 182 Limitations and Future Research Directions ................................ ................................ ........ 183 APPENDIX A COD ES FOR SPECIFYING THE COMPUTATIONAL MODEL IN SOFTWARE VENSIM PROFESSIONAL ................................ ................................ ................................ . 186 B TASK AND MANIPULATIONS USED IN STUDY 2 ................................ ....................... 199 Simu lation Task Interface ................................ ................................ ................................ ..... 199 Descriptions Used in Team Task Structure Manipulation ................................ .................... 199 Functional Condition ................................ ................................ ................................ ..... 199 Divisional Condition ................................ ................................ ................................ ..... 200 Disjunctive Condition ................................ ................................ ................................ .... 200 Descriptions Used in Differentiation in Learning Ability Manipulation .............................. 200 Same Learning Ability Condition ................................ ................................ ................. 200 Different Learning Ability Condition ................................ ................................ ............ 201 Descriptions Used for Each Team Performance Episode ................................ ..................... 202 Same Task Difficulty, Small Disturbance Magnitude ................................ ................... 202 Same Task Difficulty, Large Disturbance Magnitude ................................ ................... 202 Different Task Difficulty, Small Disturbance Magnitude ................................ ............. 203 Different Task Difficulty, Large Disturbance Magnitude ................................ ............. 203 C MEASURES USED IN STUDY 3 ................................ ................................ ........................ 204 Time 1 Subordinate Survey Q uestions ................................ ................................ ................. 204 Demographic Information ................................ ................................ ............................. 204 Skill Variety, Job Complexity, and Problem Solving ................................ ................... 204 Learning Ability ................................ ................................ ................................ ............ 204 Conscientiousness ................................ ................................ ................................ .......... 205 Time 1 Team Leader Survey Questions ................................ ................................ ............... 205 Individual focused Leadership Behaviors ................................ ................................ ..... 205 Time 1 Network Survey Questions ................................ ................................ ....................... 206 Time 2 Team Manager Survey Questions ................................ ................................ ............ 206 Team Performance ................................ ................................ ................................ ......... 206 LIST OF REFERENCES ................................ ................................ ................................ ............. 207 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ....... 220

PAGE 8

8 LIST OF TABLES Table page 2 1 A summary of previous team leadership theories and research. ................................ ........ 45 4 1 SD of time spent on individual subordinate (per team performance episode) by conditions. ................................ ................................ ................................ .......................... 87 4 2 Results from sensitive analyses. ................................ ................................ ......................... 89 5 1 External negative disturbances. ................................ ................................ ........................ 130 5 2 Relationship between discrepancy and time spent on the focal subordinate by conditions (including team focused actions). ................................ ................................ .. 131 5 3 Relationship between discrepancy and time spent on the focal subordinate by conditions (excluding team focused actions). ................................ ................................ .. 134 5 4 Leader time allocation by conditions (including team focused actions). ......................... 137 5 5 Leader time allocation by conditions (exclu ding team focused actions). ........................ 139 5 6 Estimates of dynamic model parameters and model data fit aggregated across individuals. ................................ ................................ ................................ ....................... 141 6 1 Means, standard deviations, and bivariate correlations among study variables. .............. 165 6 2 Hierarchical multiple regression results. ................................ ................................ .......... 167

PAGE 9

9 LIST OF F IGURES Figure page 3 1 The general structure of negative feedback loop. ................................ ............................... 62 3 2 A formal model of team leaders hip. ................................ ................................ ................... 63 3 3 Team development system. ................................ ................................ ................................ 64 3 4 Team member regulated performance system. ................................ ................................ .. 67 3 5 Team leader regulated performance system. ................................ ................................ ...... 71 3 6 Transition between team performance cycles. ................................ ................................ ... 74 4 1 Task contingent team leadership behavior and team task state over time by disturbance magnitude conditions. ................................ ................................ ..................... 90 4 2 Task contingent team leadership behavior and team ta sk state over time by disturbance duration conditions. ................................ ................................ ........................ 91 4 3 T ask contingent team leadership behavior, team expectancy, and team task state over time by leader reactivity conditions. ................................ ................................ .................. 92 4 4 Task contingent team leadership behavior and team task state over time by leader reactivity and team task structure conditions. A) team task structure = functional, B) team task structu re = divisional, C) team task structure = disjunctive. ............................. 93 4 5 SD of individual expectancy and team task state over time by team composition in learning ability and team task structure cond itions. A) team task structure = functional, B) team task structure = divisional, C) team task structure = disjunctive. ...... 98 5 1 Predicted (baseline), predicted (calibrated), and observed ch oices over time. ................. 144 6 1 Task interdependence moderates the relationship between differentiation in skill variety and average information giving. ................................ ................................ .......... 171 6 2 Task interdependence moderates the relationship between differentiation in skill variety and average autonomy giving. ................................ ................................ ............. 171 6 3 Task interdependence moderat es the relationship between differentiation in skill variety and average socioemotional support. ................................ ................................ ... 172 6 4 Task interdependence moderates the relationship between differentiation in conscien tiousness and average autonomy giving. ................................ ........................... 172 6 5 Task interdependence moderates the relationship between differentiation in skill variety and differentiated information giving. ................................ ................................ . 173

PAGE 10

10 6 6 Task interdependence moderates the relationship between differentiation in problem solving and differentiated information giving. ................................ ................................ 173 6 7 Task interdependence moderates the relationship between workflow network centralization and differentiated information giving. ................................ ...................... 174 6 8 Task interdependence moderates the relat ionship between differentiation in learning ability and differentiated information giving. ................................ ................................ .. 174 6 9 Task interdependence moderates the relationship between differentiation in skill variety and team performance. ................................ ................................ ......................... 175 6 10 Task interdependence moderates the relationship between differentiation in task complexity and team performance. ................................ ................................ .................. 175 6 11 Task interdependence moderates the relationship between workflow network centralization and team performance. ................................ ................................ .............. 176

PAGE 11

11 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy A FORMAL MODEL OF LEADERSHIP IN TEAM GOAL PURSUIT: TEAM DESIGN, TEAM COMPOSITION, AND DYNAMIC LEADER REGULATORY PROCESSES By Le Zhou August 201 4 Chair: Mo Wang Major: Business Administration In this dissertation, I examine the role of leadership in the team goal pursuit process. Team is a basic organizational unit used to achieve collective goals. Theory and research have identified leadershi p as one of the most influential forces in setting directions for teams and overseeing teams while they strive to achieve the goals. Most of the existing team leadership studies have examined static team leader attributes and behaviors as predictors of tea m performance . This research is limited in terms of (a) understanding the dynamics in team leadership, (b) explaining why some team leadership functions arise as reactions to the team performance and the task environment, and (c) tak ing the effects of team composition and team design factors into account. Therefore, I propose and test an integrated formal model on team leadership, drawing on self regulation theory and functional leadership theory. In S tudy 1, a formal computational model was specified. Qual itative model evaluation suggested that computational modeling results were consistent with previous empirical findings. In Study 2, an experimental lab study was conducted to test hypotheses derived from the computational model. S upporting the theoretical model, Study 2 showed that over time leader spent more time regulating a su bordinate when his/her relative discrepancy was larger comparing to other team

PAGE 12

12 members and this relationship was stronger when team task structure was functional rather than divisi onal or disjunctive. Study 2 also showed that differences across team positions in task difficulties were positively related to the differentiation in leader time allocation across subordinates . In addition, differences among subordinates in learning abili ty were positively related to the differentiation in leader time allocation across subordinates, and th is positive relationship was stronger when team task structure was functional rather than divisional or disjunctive. Study 3 further tested the computati onal model with data from a field sample . Study 3 found that team task interdependence strengthened the positive effects of differences across team position s in knowledge characteristics and differences across team member s in learning ability and conscient iousness on average and differentiated individual focused leadership behaviors. Theoretical and practical implications from these three studies are discussed.

PAGE 13

13 CHAPTER 1 INTRODUCTION For a long time in human history, organizations have used teams as basic units to organize manpower and achieve collective goals (e.g., to gather food, raise off spring, and expand territory; Kozlowski & Ilgen, 2006). However, management research and practice in the first half of the 20 th century has largely assumed teams as s imple compositions of individuals, which provides limited understanding on the nature of team based activities (Kozlowski & Klein, 2000). With an increasing number of organizations redesigning work around team based structures in the past three decades (Ma rtin & Bal, 2006), newly developed theory and research on team based phenomena are proliferating (Kozlowski & Bell, 2003). Organizational researchers define team interact (face to face or, increasingly, vir tually) (c) possess one or more common goals; (d) are brought together to perform organizationally relevant tasks; (e) exhibit interdependencies with respect to workflow, goals, and outcomes; (f) have different roles and responsibilities; and (g) are toget her embedded in an encompassing organizational system, with boundaries and linkages this definition, one purpose of organizing work around team based structure i s to better utilize resources possessed by the team and its members to achieve a common goal. This process in which individuals work together to allocate resources and use resources to strive toward a common goal can be defined as team goal pursuit (DeShon , Kozlowski, Schmidt, Milner, & Wiechmann, 2004). Among all the factors that might influence team goal pursuit, scholars generally agree that team leadership plays an important role (Day, Gronn, & Salas, 2006; Morgeson, DeRue, & Karam, 2010; Zaccaro, Rittm an, & Marks, 2001; Wageman, Gardner, & Mortensen, 2012).

PAGE 14

14 Ample evidence has already demonstrated that leadership is indeed one of the most influential forces that drive team effectiveness. For example, studies found that leadership style of team supervisor team (e.g., Colbert, Kristof Brown, Bradley, & Barrick, 2008; Kearney & Gebert, 2009), collective efficacy (e.g., Wu, Tsui, & Kinicki, 2010), and interpersonal interactions (e.g., Zhang & Peterson, 2011). Studies have also shown that even self managing teams are subject to the influence from external leaders, who can significantly increase team effectiveness by empowering team members (e.g., Chen, Kirkman, Kanfer, Allen, & R osen, 2007), facilitating team learning (e.g., Wageman, 2001), and managing boundaries of the teams (e.g., Druskat & Wheeler, 2003). Despite the large volume of literature on team and leadership, most of the existing team leadership studies paid little att ention to disentangling the complicated dynamic and structural issues in team leadership phenomena (Day et al., 2006; Klein, Ziegert, Knight, & Xiao, 2006; Zaccaro et al., 2001). First, in terms of the dynamics in team leadership process , existing team lea dership theories are not clear about the mechanisms that drive team leadership behaviors over time . This is partly because research has not yet systematically scrutinized the assumption that team leadership is merely antecedents or inputs into the team per formance system. Recent empirical and theoretical research has begun to question this assumption (Day et al., 2006). In particular, s tudies from different theoretical perspectives have consistently suggested that some team leadership behaviors may be a rea ction to changes in the team task environment. For demonstrated that intervention from leaders to the team was driven by the unexpected events the team encounters when pe rforming team tasks

PAGE 15

15 transformational leadership suggested that the amount of transformational leadership behaviors of team leaders varied according to the team context (e.g., virtual vs. face to face interactions ). Kle in et al. (2006) , a qualitative study , also suggested that empowering team leadership varied over the course a task From a theoretical standpoint, narrowly viewing team leadership as s table inputs to team process contradicts the functional view of team leadership, which argues that part of team leadership functions are reacting to discrepancy in current and desired team task state s (DeChurch & Marks, 2006; McGrath, 1962; Morgeson et al. , 2010). As Vancouver, Tamanini, and Yoder (2010) pointed out, interactions between the leader and the subordinate happen on a two can impact the subordinate. As suc h, it is premature to assume that team leadership can only serve as inputs to team performance system. More research is needed to clarify how and why some leadership functions are triggered by team itself and its task environment. With team leadership vie wed as a static factor, there is also limited research on the influence of team leadership on team goal pursuit over multiple team performance episodes (Day et al., 2006; Morgeson et al., 2010). Team design research has recognized that some factors in the task environment change systematically (e.g., task and knowledge characteristics; Morgeson & Humphrey, 2006), whereas some other factors change randomly (e.g., disturbances from the environment; Morgeson, 2005; Morgeson & DeRue, 2006). However, h ow these t ime variant factors influence team goal pursuit and team leadership behaviors has not been fully understood. As such, it is necessary for team leadership research to start incorporating other theoretical perspectives that explicate dynamics in goal directe d activities, such as self regulation theory (Carver & Scheier, 1998; Vancouver, 2008).

PAGE 16

16 I n terms of the structural issues, existing team leadership theories seldom consider how characteristics of the team (e.g., team member learning abilities , team task st ructure ) influence team leadership. As noted by several researchers (Hackman & Wageman, 2005; Harrison & Humphrey, 2010; Hollenbeck, DeRue, & Guzzo, 2004; Wageman, 1999), team leadership research should incorporate our knowledge about team composition (i.e ., characteristics of the individuals occupying the roles within the team) and team design (i.e., characteristics of the roles within the team and how they are coupled together) . One of the reasons is that team design and team composition might manifest as boundary conditions of the relationship between team leadership and team performance (e.g., Chen et al., 2007). Incorporating team design and team composition is also justified by recent development in the contextual view on organizations (Johns, 2006). S pecifically, an emerging stream of research suggests that when dimensions of the organizational context are studied as a configuration or a set (e.g., leadership, team composition, and team design), they are more interpretable and meaningful, and have stro nger effects on behaviors of lower than independent discrete dimensions (Bowen & Ostroff, 2004; Delery & Doty, 1996; Johns, 2006; Rousseau & Fried, 2001). In team leadership research, this contextual p erspective has seldom been adopted (see Hollenbeck et al., 2002 for an exception). In addition, from the levels of analysis perspective (Kozlowski & Klein, 2000), taking team design and team composition into consideration is necessary for developing a tea m leadership theory that is well aligned with the phenomenon in question. In particular, what makes leading individuals in a team different from leading multiple independent persons is that the former is essentially a team level phenomenon (i.e., the distr ibution of leadership across multiple people). In this situation, the influence of leadership on one individual can affect the influence of

PAGE 17

17 leadership on another individual (an analog example is that when playing chess, moving one piece is going to change cognitive and behavioral processes happen in a team context. Thus, pure reductionism approach to team leadership (i.e. studying team leadership as a simple additive of leading multiple in dependent persons) may not capture the complex interaction between the context and the individual processes. As such, it is necessary to consider how the individual focused leadership leadership theory requires an explicit treatment of team design and team composition features. Given the above research gaps , the aim of the current research is to develop and test a formal model on team leadership that explains how team leadership behavi ors change over time and how team design and team composition factors influence team leadership as well as the overall team performance system. The proposed model draws on and integrates the functional theory of leadership (McGrath, 1962; Morgeson et al., 2010; Zaccaro et al., 2001) and the self regulation theory of goal pursuit (Carver & Scheier, 1998; Lord & Levy, 1994; Steel & König, 2006; Vancouver, 2000; 2008). In Study 1, I specify a formal computational model for team leadership. Qualitative model ev aluation was conducted through reviewing key arguments and hypotheses proposed in previous studies, and comparing previous empirical findings with the computational simulation results. In Study 2, a lab experiment was conducted to test hypotheses derived f rom the model, which concerns the joint effects of team member differences in learning ability, team position differences in task difficult y , team task structure, and external disturbance on team leader time allocation within and between team performance episodes . In Study 3, field data were collected to test the hypotheses about the joint effects of knowledge characteristics of

PAGE 18

18 the team positions, team member differences in learning ability and conscientiousness, and team task structure on average and dif ferentiat ed individual focused leadership behaviors. The current research makes several theoretical and empirical contributions to team leadership research and related fields. First, this research integrates functional leadership theory and self regulatio n theory to propose a theoretical model on the dynamics in team leadership. Integrating these two theories allows conceptualizing team leadership as part of the team regulation system. This integrated framework clarifies the theoretical mechanisms through which team leadership and team performance states influence each other over time, which contributes to team leadership research. This framework also contributes to leadership theory in general by providing an explanation for how functional leadership behav as well as changes in the task environment. In addition, by proposing and testing the integrated model, this research demonstrates that the structure and function of organizational units an d their leaders can be studied from a system based perspective and explained by a coherent set of mechanisms (Katz & Kahn, 1978). Second, this research integrates team design and team composition literatures with team leadership literature, and examine s te am design ( e.g. , team structure ) and team composition factors (e.g., team member composition in learning ability) as structural input of dynamic team performance system. This helps explain how the team design can impact behaviors of team members and leader s, and team performance over time. Third, this research proposes a computational model that formally delineates the mechanisms that drive dynamics in team leadership and the la rger team regulation system. Taking the computational modeling approach allows t his research to provide more precise description of the dynamic processes in the phenomenon . Therefore, this research also contributes to the emerging stream of research that takes the computational modeling approach

PAGE 19

19 to examine goal directed dynamic behavi ors (e.g., Vancouver, Tamanini et al., 2010; Vancouver, Weinhardt , & Schmidt , 2010). Finally, the computational model developed by this research can be used to guide team member selection, training, team leader development, and team design practices.

PAGE 20

20 CHAP TER 2 LITERATURE REVIEW In this chapter, I first review four theoretical perspectives that have been applied by previous research to understand team leadership, including functional leadership theory (McGrath, 1962), transformational leadership theory (Bas s, 1985; Podsakoff, MacKenzie, Moorman, & Fetter, 1990), leader member exchange theory (LMX; Graen & Scandura, 1987), and theory of leadership for self managing teams (Manz & Sims, 1987). The purpose of the review is to identify and summarize (a) key team leadership constructs, (b) theoretical mechanisms linking team leadership to team performance, and (c) empirical findings regarding the mechanisms articulated in each theoretical perspective (a summary is provided in Table 2 1). Next, I review team composi tion and team design literature, in order to pinpoint key team composition and team design variables that may influence team goal pursuit. Finally, I review studies that have examined the joint effects of team leadership, team composition, and team design on team performance. Constructs, mechanisms, and empirical findings reviewed in this section together serve as the foundation for the later develop ment and evaluat ion of the formal computational model. Team Leadership The current body of team leadership research is largely built on four leadership theories. First, based on the idea of functional leadership theory (McGrath, 1962), some research examines how leadership develops, facilitates, and ensures team functioning that leads to team goal accomplishmen t. Second, built on transformational leadership theory (Bass, 1985; Podsakoff et al., 1990), some research examines how behaviors of transformational leaders influence team performance. Third, drawing on LMX theory (Graen & Scandura, 1987), researchers hav e examined whether and how team performance is affected by average and

PAGE 21

21 self managing teams, some studies have examined how managers can facilitate self managi ng teams to manage their team functioning and achieve team performance. In the following sections, I review each theoretical perspective in terms of its key leadership variables, theoretical mechanisms proposed for team leadership team performance relation ship, and related empirical findings (see a summary in Table 2 1). Functional Leadership Theory h, 1962, p. 5). Based on this idea, leadership can be conceptualized as a set of functions that enable the subsystems (e.g., individual followers, roles, sub teams) in a larger system to achieve a shared goal (Fleishman, Mumford, Zaccaro, Levin, Korotkin, & Hein, 1991). The extension of the functional leadership idea to the team level is built on two theoretical foundations. First, research conducted by Kozlowski and colleagues (Bell & Kozlowski, 2002; Kozlowski, Gully, Salas, & Cannon Bowers, 1996) suggest s that in order for teams to function effectively, leadership should satisfy developmental and performance needs of the team. Early in team building, team leadership should serve the functions of developing required competencies of team members. Later, tea m leadership should serve the functions of monitoring and facilitating team performance. Second, Marks, Mathieu, and Zaccaro (2001) proposed a theoretical framework delineating specific team functions involved in transition and action phases of each team p erformance cycle. Taken works suggest that team leadership can be conceptualized as leadership functions that satisfy needs for team functioning in different team development and performance sta ges.

PAGE 22

22 Drawing on previous theorizing and research (Bell & Kozlowski, 2002; DeChurch & Marks, 2006; Hackman & Wageman, 2005; Kozlowski et al., 1996; Marks et al., 2001; Zaccaro et al., 2001), Morgeson et al. (2010) summarized a list of team leadership funct ions. Specifically, d uring the transition phase of team performance cycle, the core function of team leadership is to plan for team actions that can lead to goal achievement. Specific team leadership functions in this stage include composing teams, definin g mission, setting goals and expectations, structuring and planning, developing competencies of the team, facilitating sense making, and providing feedback. During the action phase of team performance cycle, the core function of team leadership is to monit or and facilitate team activities that strive toward goal accomplishment. Specific team leadership functions include monitoring team progress in reaching its goal, managing team boundaries, challenging teams, performing team tasks for the team, solving pro blems, providing resources, encouraging self management, and supporting developing social climate. Several mechanisms linking team leadership functions to team performance have been proposed. First, team leadership functions can influence team performance through influencing cognitive team states and processes, including team member knowledge and skills, shared mental models, team metacognition, and collective information processing (Hackman & Wageman, 2005; Kozlowski et al., 1996; Randall, Resick & DeChur ch, 2011; Zaccaro et al., 2001). Second, team leadership functions can influence motivational states of the teams, which in turn influence team performance. Such motivational states include team performance strategy, collective efficacy, team cohesion, and the amount of effort spent on the team task by team members (Hackman & Wageman, 2005; Kozlowski et al., 1996; Zaccaro et al., 2001). Third, team leadership functions can influence team performance through influencing team affective states,

PAGE 23

23 such as conflic t and emotion control norms and shared affective experience (Kozlowski et al., 1996; Zaccaro et al., 2001). Fourth, in terms of collective behaviors, team leadership function can influence team performance through influencing intrateam and interteam coordi nation processes, and information sharing among team members (DeChurch & Marks, 2006; Randall et al., 2011; Zaccaro et al., 2001). Empirical test of the functional team leadership theory found that sense giving of team leaders can increase team effectivene ss by increasing accuracy and similarity of mental models among team members (Randall et al., 2011). In addition, it was found that team leader strategizing and coordinating behaviors can increase team performance through improving interteam coordination (DeChurch & Marks, 2006). These results suggest that team leadership function indeed can influence team performance through influencing cognitive and behavioral team states and processes. However, so far there is no empirical study conducted aiming toward offering evidence regarding the motivational and affective mechanisms underlying the team leadership functions team performance relationship. Transformational Leadership Theory Transformational leadership theory has been applied to the team setting to un derstand leadership theory argues that leaders can be described by the extent to which they use the transformational leadership style, which includes four dimensio ns. Idealized influence , or sometimes referred to as charisma, is defined as serving as moral and ethical role models for followers. Inspirational motivation is defined as articulating and clarifying vision to followers. Intellectual stimulation is defined as challenging the status quo and inspiring followers to be creative in problem solving. Individual consideration is defined as providing instrumental and

PAGE 24

24 extended the conceptualization of transformational leadership to include two additional dimensions. Supporting followers toward goal accomplishment describes leadership support provided while followers work toward the goal, such as actively engaging in task problem solving and acting as role models. Encouraging cooperation describes team leader behaviors that improve collaborative teamwork among members. Motivational, affect, and behavioral mechanisms underlying the effects of transformational leadership on team pe rformance have been theorized. Specifically, transformational leadership can influence team performance through influencing several team motivational states. First, transformational leadership can increase follower perceived person organization value congr uence (Hoffman, Bynum, Piccolo, & Sutton, 2011), similarity in team goal importance perceived by team members (Colbert et al., 2008), and identification with the team (Kearney & Gebert, 2009; Wang & Howell, 2012; Wu et al., 2010). Because of such strong an d shared focus on team interests instead of personal interests induced by transformational leadership, team members are more likely to align their individual goals with team goals and put more effort into striving for team goal (Eisenbeiss, Knippenberg, & Boerner, 2008). Second, the leader and trust in the leader (Schaubroeck, Lam, & Peng, 2011; Wang & Howell, 2012). Therefore, followers of transformational leader shared beliefs about the capabilities of the team to achieve team tasks (team efficacy, team potency, or tea m empowerment), which in turn can increase team performance (Bass, Avolio, Jung, & Berson, 2003; Cole, Bedeian, & Bruch, 2011; Schaubroeck, Lam, & Cha, 2007; Schaubroeck et al., 2011; Wu et al., 2010). Fourth, some researchers suggest that

PAGE 25

25 transformational leadership can build team cultures that favor learning and adaptation, which leads to higher team performance (Nemanich & Vera, 2009). It should be noted that Wu et al. (2010) argue that variation in transformational leadership can be detrimental to team performance given that it can increase variation in self efficacy among team members, which is negatively related to team performance. In addition, transformational leadership can influence team performance through affective team states. Transformational l based trust in the leader, psychological safety, and trust in the team (Gundersen, Hellesoy, & Raeder, 2012; Schaubroeck et al., 2011). These positive affective states can encourage team members to put more effor t into team tasks and help each other while working toward the team goal, and thus result in higher team performance. Transformational leadership can also help teams buffer stress resulted from unexpected events, which may lead to higher team performance ( Kearney & Gebert, 2009). Finally, transformational leadership can improve team performance by improving information sharing and elaboration of task related information in peer interactions, which can in turn improve team performance (Kearney & Gebert, 2009 ; Zhang & Peterson, 2011). A number of quantitative studies have directly tested the mediating role of team motivational states between transformational leadership and team performance. Specifically, empirical studies show that the indirect effect of tran sformational leadership on team performance is mediated by average follower perceived follower organization value congruence (Hoffman et al., 2011), variation in follower perceived goal importance (Colbert et al., 2008), team identification (e.g., Kearney & Gebert, 2009), team empowerment (e.g., Bass et al., 2003), and cognition based trust in the leader (Schaubroeck et al., 2011). There is also evidence showing that the indirect effect of transformational leadership is mediated by affective and

PAGE 26

26 behavioral team functions (e.g., Gundersen et al., 2012; Kearney & Gebert, 2009; Schaubroeck et al., 2011; Zhang & Peterson, 2011). However, so far there is no study that has examined how team cognitive states and processes mediate the relationship between transforma tional leadership and team performance. Leader Member Exchange Theory Team leadership research based on LMX theory (Graen & Scandura, 1987) has mainly examined four team leadership constructs: group average LMX, median LMX, within group LMX differentiation , and leader leader exchange relationship (LLX) . Average and median LMX describe the central tendency of the quality of the exchange relationships between group leader and the subordinates. Average LMX indicates the average level of LMX of multiple leader subordinate dyads of a group. Median LMX indicates the level of LMX that separates the higher half of a group from the lower half of the group in terms of the ranking of LMX . Median LMX is used when the distribution of LMX within a group is highly skewed ( Le Blanc & Gonzalez Roma, 2012; Liden , Erdogan, Wayne, & Sparrowe , 2006). LMX differentiation (or variation of LMX) indicates the amount of separation of individual subordinates in terms of their LMXs with a common leader (Henderson, Liden, Glibkowski, & C haudhry, 2009; Liden et al., 2006). LLX captures the quality of the exchange relationship between team leaders and their own supervisors, which indicates the amount and quality of resources exchanged between team leaders and higher level managers (Zhou, Wa ng, Chen, & Shi, 2012). Boies and Howell (2006) argue that average LMX is positively related to team performance through influencing team motivational states and intrateam communication process. owell (2006) argue that average LMX can increase team potency through verbal persuasion and enacted mastery. When team members enjoy high quality relationship with the leader on average, they receive more verbal

PAGE 27

27 encouragement, information, and emotional su pport, which can persuade the team as a whole to hold higher efficacy beliefs. Team members may also receive more developmental opportunities that are similar to the team tasks. Higher team efficacy developed from verbal persuasion and enacted mastery can in turn increase team performance. Boies and Howell (2006) also argue that average LMX can improve communication among team members, which can reduce team conflict and increase team performance. Empirical research showed that average LMX was positively rel ated to team performance indirectly through team potency (Boies & Howell, 2006). Different reasoning and predictions exist regarding whether LMX differentiation is beneficial or detrimental to team performance. One point of view argues that LMX differentia tion can be seen as a type of division of labor which resulted from differentiation of roles within a team (Liden et al., 2006). Given that individuals are assigned to roles that require their special expertise, division of labor within a unit can increase unit performance through the strategic use of team resources (Naidoo, Scherbaum, Goldstein, & Graen, 2011; Stogdill, 1959). In addition, team members are likely to work harder when they perceive high quality LMX as a type of incentive and privilege (Le Bl anc & Gonzalez Roma, 2012). Therefore, LMX differentiation may increase group performance. However, as Liden et al. (2006) pointed out, LMX differentiation also has the potential to decrease group performance due to the hierarchy (i.e., status differentiat ion) created. LMX may be viewed as disparity in personal values and rewards among group members, which has been shown to reduce collective performance (Bloom, 1999). Empirical findings regarding the main effect of differentiation of LMX on team performance have remained inconclusive so far (Boies & Howell, 2006; Le Blanc & González Romá, 2012; Liden et al., 2006). Studies found that median LMX, task interdependence, and team developmental stage could be moderators on the LMX differentiation team performance

PAGE 28

28 relationship, such that LMX differentiation was positively related to team performance when median LMX was low, when task interdependence was high, or when team was more mature (Le Blanc & González Romá; 2012; Liden et al., 2006; Naidoo et al., 2011). Te am leaders are both supervisors of lower level subordinates and subordinates of higher level managers. Due to this unique structural position of team leaders, their relationships with their own managers (i.e., LLX) can influence the team and its members. T angirala, Green, and Ramanujam (2007) propose that LLX moderate s the strength of the positive effect s of LMX (Cronpanzano & Mitchell, 2005) , these authors argue that when team leaders have a high quality exchange relationship with their own supervisors, they receive more resources from the higher level managers. The resources bestowed by higher level managers allow team leaders to treat those subordinate s with whom they have a high quality rel ationship (i.e., higher LMX) better. Empirical findings showed that LLX could strengthen the positive relationship s attitude s toward the organization and the customers (Tangirala et al., 2007). Venkataramani, Green, and Schleicher (2010) argue that LLX is positively related to LMX. Based on social network research, these authors argue that supervisors with higher LLX can be viewed as enjoying higher status in the organization, which motivates subordinates to develop h igher quality LMX with them. Supporting this argument, empirical findings showed that perceived leader status mediated the positive effect of LLX on LMX, which was in turn sfaction and turnover intention . Zhou et al. (2012 ) further integrated and extended previous research on LLX. These authors propose three mechanisms through which LLX influence the team and individual subordinates. First, based on social learning perspective of leadership (Bass, 1990; Mayer,

PAGE 29

29 Kuenzi, Green baum, Bardes, & Salvador, 2009), supervisors with a good relationship with their own supervisors can learn from their role models and develop good relationships with their subordinates. Therefore, LLX can positively influence the average LMX team leaders h ave with subordinates in the team . Second , LLX could allow the team leaders to empower the team as a whole. Team empowerment could in turn motivate individual subordinates and increase their job per formance and job satisfaction. Third , according to social exchange theory (Cronpanzano & Mitchell, 2005), team leaders with higher LLX receive more resources from the higher level managers, which facilitates the vational states, job performance, and job attitude. Empirical findings supporte d all three mechanisms proposed (Zhou et al., 2012). Leadership for Self managing Teams Self managing teams are collectives of individual workers that can self regulate their activities in interdependent tasks (Goodman, Devadas, & Griffith Hughson, 1988). Manz and Sims (1984; 1987) argue that external leaders can still actively use leadership behaviors to influence the effectiveness of self managing teams. These leadership beha viors include setting team performance goals, communicating expectations, monitoring team performance, facilitating problem solving, and preparing teams to cope with and make sense of disruptive events (Manz & Sims, 1987; Morgeson, 2005; Morgeson & DeRue, 2006; Tesluk & Mathieu, 1999). Based on has extended leadership for self managing teams to include empowering leadership behaviors , ich leaders enhance autonomy, control, self management, and leadership behaviors include encouraging subordinates to set their own goals and involving subordinate s in decision making (Kirkman & Rosen, 1999; Manz & Sims, 1987).

PAGE 30

30 Several mechanisms that underlie the positive effect of leadership for self managing teams on team performance have been discussed in the literature. First, empowering behaviors of the team leaders can influence team motivational states, including collective efficacy (Carmeli, Schaubroeck, & Tishler, 2011; Srivastava, Bartol, & Locke, 2006) and perceived motivational characteristics of team tasks (e.g., decision making autonomy, meaningfulnes s, and impact; Chen et al., 2007; Kirkman & Rosen, 1999). Second, external leaders can increase team performance through improving team problem management actions (Tesluk & Mathieu, 1999). Third, empowering leadership encourages team members to share decis ion latitude, knowledge, and information, and collaborate with each other to accomplish tasks (Carmeli et al., 2011; Srivastava et al., 2006). However, there has been no empirical test of whether leadership for self managing teams can influence team perfor mance through affective team states. Summary Based on the above review, the functional approach to leadership seems to be the most suitable for understanding team leadership, for at least three reasons. First, leadership functions identified by functional approach include the specific behaviors identified by the other three theories. As Hackman and Wageman (2005) pointed out, taking the functional approach ensures a complete but parsimonious conceptual analysis of team leadership. This is because functiona l approach avoids describing all the specific behaviors or styles that team leaders may exhibit in all sorts of specific situations. Instead, this approach identifies which team leadership functions are critical for the team based on team functions that ma nifest in the same structure (i.e., team functions in transition and action phases of team performance cycles) across different task situations. Second, since the functional approach to team leadership is rooted in team functioning in team performance cyc les, functional leadership theory can be easily extended to account for the

PAGE 31

3 1 dynamics in team leadership. As I will delineate in detail in the formal model, as team functioning changes over time and from task to task, the specific amount of resources, time, and effort that leaders spend on performing different team leadership functions may vary accordingly. This association between dynamics in team leadership functioning and team ensure that the team is not lagging behind its ideal team goal pursuit state. Third, functional team leadership theory is suitable for serving as the foundation for incorporating team composition and team design factors into an integr ated team leadership model. Functional team leadership theory argues that an important part of team leadership by effectively performing team tasks collectively (Hackman & Wageman, 2005; Kozlowski et al., 1996). As such, it is conceivable that how much team leaders engage in each specific team as well as the task design features (e.g., task difficulty, task interdependence). Based on the above three considerations, I use functional team leadership theory as the basis for developing the integrated formal model. Team Composition As I explained earlier, examination of te am leadership should incorporate the knowledge from team composition and team design research. Therefore, in this section and the next section, I review literature in these two areas in order to identify key team composition and team design variables that influence team performance. I also review existing research on the theoretical mechanisms underlying the influence of team composition and design.

PAGE 32

32 Team Composition Constructs Team composition refers to how team is composed in terms of certain individual c haracteristic (LePine , Hollenbeck, Ilgen, & Hedlund , 1997). Both competence and motivation related individual characteristics have been examined in team composition research. In terms of team composition in competenc e , specific characteristics that have be en examined include general cognitive ability (e.g., Barrick , Stewart, Neubert, & Mount , 1998), emotional intelligence (e.g., Jordan & Troth, 2004), and goal orientation (e.g., LePine, 2005). With respect to team composition in motivational characteristics , specific characteristics that have been examined include demographics (Fisher, Bell, Dierdorff, & Belohlav, 2012), Big Five personality factors (e.g., Barrick et al., 1998; Barry & Stewart, 1997; LePine, 2003), core self evaluation (CSE; e.g., Zhang & Pe terson, 2011), psychological collectivism (Dierdorff, Bell, & Belohlav, 2011), and Four different configuration models are often used to conceptualize team composition: Team level characteristic as minimum, maximum, average (mean or median), and variation of team member standing on a particular characteristic (Barrick et al., 1998). (1972) typology of team task structure, LePine et al. (1997) and Barrick et al. (1 998) propose that team composition in competenc e should be configured according to the specific team task structure. When team task structure is additive (or divisional ; i.e. , individuals complete their work independently and group output is the simple s ummation ; Hollenbeck & Spitzmuller, 2012) , average team member ability is the appropriate team composition measure. When team task structure is conjunctive (or functional ; i.e., individuals complete their work interdependently and group completes its work only when every individual completes his/her work ) , the lowest ability level among team members is the appropriate team composition measure. When team task structure is disjunctive (i.e., individuals complete their

PAGE 33

33 work independen tly and group completes its work as long as one individual completes his/her work ) , the highest ability level is the appropriate team composition measure. In addition, when heterogeneity is of theoretical interest, variation is the appropriate measure of t eam composition (Barrick et al., 1998). Team Composition and Team Performance Empirical studies found that team composition can influence team performance, but the strength of the relationship varies depending on the specific individual characteristic exam ined, the operationalization of team composition, and the study setting (Bell, 2007). In a meta analysis mental ability, and average emotional intelligence were sign ificantly related to team agreeableness and average conscientiousness, openness to experience, teamwork preference, and collectivism were strongly related to team performanc e. In addition to demonstrate the main effects of team composition factors on team performance, a few studies have examined the mechanisms linking team composition to team performance. Barry and Stewart (1997) examined how team composition in personality (extraversion and conscientiousness) affected team performance through influencing team processes (task focus, group cohesion, open communication, and leadership emergence) in small groups. They found that the proportion of extraverted members had a curvil inear relationship with team performance that was mediated by task focus. However, team composition in conscientiousness was not significantly related to team processes and performance. Fisher et al. (2012) examined how team compositions in personality (a verage agreeableness) and demographic characteristics (heterogeneity in race and gender) influenced team performance through influencing team mental models and coordination process. Fisher et

PAGE 34

34 al. argue that team that is characterized by a high level of agr eeableness has higher quality interpersonal interactions because team member in general are more cooperative and trust each other more. Therefore, high average agreeableness fosters similar team mental models among team members. The authors also argue that surface level demographic similarity influences the social integration process within the team, which may impede team members from developing similar team mental models. E mpirical findings from Fisher et al. suggested that team average cooperation and tru st (facets of agreeableness) and racial heterogeneity were related to similarity performance. Team Design In addition to team leadership and team composition, team design features also have significant impact on team performance. T eam design creation of organizational support for the team and (Morgeson & Humphrey, 2008, p. 46). Three aspects of team design have been found to influence team performance: team task characteristics, team structure, and task environment uncertainty. Team Task Characteris tics As summarized by Morgeson and Humphrey (2006), task characteristics include two aspects, motivational and knowledge related characteristics. Motivational task characteristics include autonomy in work scheduling, decision making, and work methods, tas k variety, task significance, task identity, and feedback from job. Knowledge related task characteristics include job complexity, information processing, problem solving, skill variety, and specialization. In addition, Humphrey, Morgeson, and Mannor (2009 ) proposed that in the team setting, different

PAGE 35

35 positions that are in charge of special types of tasks can also be evaluated by its strategic importance. A position that is strategically important is one that (a) encounters more problems to be solved, (b) h andles more work, and (c) is more central in the workflow in the team. Individual level task characteristics can be aggregated to the team level, using average, minimum, maximum, and variation, to capture characteristics of the tasks within the same team ( Morgeson & Humphrey, 2008). Several mechanisms linking team level task characteristics to team performance have been examined. Some researchers argue that team tasks that are characterized by high autonomy and meaningfulness can increase team empowerment, which in turn leads to higher team performance (Kirkman & Rosen, 1999; Mathieu, Gilson, & Ruddy, 2006). In addition, Humphrey et al. (2009) argue that team performance is a function of the match between team member ability, incentives given to the team mem bers, and the strategic importance of the positions. Humphrey et al. found that when team members who occupied strategically important positions were more competent (having higher experience and job related skills) and were given more financial incentives, team performance was better. Although results from a meta analysis study showed that team level task characteristics were significantly related to team performance (Stewart, 2006), there is very limited discussion on the theoretical mechanisms that under lie team level task characteristics and team performance relationship. Some researchers suggest that self regulation theory may be a promising theoretical perspective to help understand the underlying mechanisms of task characteristics effect (Holman, Cleg g, & Waterson, 2002; Morgeson & Humphrey, 2008; Parker & Ohly, 2008). For example, Parker and Ohly (2008) suggest that task characteristics can influence both goal choice and goal striving processes over time, which in turn influence team performance on di fferent tasks.

PAGE 36

36 However, so far there is a lack of detailed theoretical account on how such influence from task characteristics on team goal pursuit may unfold over time. Team Structure Team structure refers to how tasks, responsibilities, and authority ar e distributed among different positions within the team (Hollenbeck et al., 2004; Stewart & Barrick, 2000). According to Hollenbeck and Spitzmuller (2012), team structure includes multiple dimensions. Team task structure (or team task allocation structure) is the extent to which team task is organized by functional specialization. When team task is structured functionally (i.e., high departmentalization, specialization of subunits, or high horizontal interdependence), different positions are responsible for different functions for team performance and, thus, team cannot accomplish its task unless all functions accomplish their responsibilities. Team decision making structure (also called centralization dimension of team structure or vertical interdependence) describes how decision latitude or authority is distributed within the team. When decision latitude is highly centralized, only one or few people in the team have the power to make decisions. In contrast, teams with less centralized decision latitude may allow all members to participate in decision making. Team reward structure (also called outcome interdependence) task performance. In teams with high outcome inter dependence, team based goals are used to align team member actions. The reward systems that teams use can be cooperative, competitive, or a mixture of both. Finally, team communication structure (also called spatial interdependence) is the extent to which team design facilitates team members to exchange information. Team structure has been shown to influence team performance. Stewart and Barrick (2000) found that team task structure ha d a curvilinear relationship with team performance, which was mediated by intrateam processes, including communication and conflict. The strength

PAGE 37

37 of this relationship differed according to the type of task performed by the team (conceptual vs. behavioral task). For teams that mostly engaged in conceptual task, the level of inte rdependence within the team was related to team performance in a U shaped way whereas for teams that mostly engaged in behavioral tasks, the level of interdependence was related to team performance in a reversed U shaped way. Team decision making structure can also influence team performance and its effect depends on the type of team task as well (Stewart & Barrick, 2000). For teams that mostly engage in conceptual tasks, centralization of team decision making is negatively related to team performance, wher eas for teams that mostly engage in behavioral tasks, centralization of team decision making is positively related to team performance. In addition, the effect of current team decision making structure on team performance can be influence by previous decis ion making structure used. For example, Hollenbeck, Ellis, Humphrey, Garza, and Ilgen (2011) found that when team decision making structure changed from decentralized to centralized structure as compared to the reverse direction, teams experienced less eff iciency and had worse performance on non routine tasks. Team reward structure can influence team performance as well. Wageman (1995) found that teams performed best when the team reward structure was either cooperative or competitive, the effect of which suggested that teams with hybrid reward system had worst member interaction and member satisfaction, which resulted in lowest team performance. In addition, research suggests that t he effect of current team reward structure on team process and outcomes can be dependent on previous team reward structure. For example, Johnson, Hollenbeck, Humphrey, Ilgen, Jundt, & Meyer (2006) examined how teams adapted to changes in team reward struct ure. They found that

PAGE 38

38 comparing team s whose reward structure changed from competitive to cooperative to those whose reward structure changed from cooperative to competitive, the former teams had lower information sharing after the change , which led to lower collective decision making accuracy and less time spent in reaching decision. Task Environment Uncertainty Although task environment external to the team performance system can be described along various dimensions, task environment uncertainty (or stabi lity, dynamism, predictability; Burns & Stalker, 1961; Katz & Kahn, 1978) is the one dimension that has been consistently shown to significantly influence team processes and performance. In addition, task environment uncertainty may fluctuate over time, th us is important to consider when examining the dynamic team goal pursuit process. Task environment uncertainty is influenced by competition in the environment, changing task demands, organizational strategy, and technology development (Burns & Stalker, 196 1). Literature on top management teams has well documented that environment uncertainty can change the strength and shape of the effect of top management 2006; Cannell a, Park, & Lee, 2008; Haleblian & Finkelstein, 1993). For example, Cannella et al. (2008) found that environmental uncertainty could strengthen the positive effect of functional diversity of top management team members on firm performance. It was also foun d that as environmental uncertainty became higher, the positive linear relationship between top management team international experience, education heterogeneity, and tenure diversity on firm performance became reverse U shaped (Carpenter & Fredrickson, 20 01). Uncertainty also exists in the work environment of lower level work teams, which arises from unexpected disruptive events and changing task demands for the team (Klein et al., 2006; Kozlowski et al., 1996; Marks et al., 2001; Morgeson, 2005; Morgeson & DeRue, 2006). Team

PAGE 39

39 research has started to examine how uncertainty in team task environment influences team performance. For example, Wall, Cordery, and Clegg (2002) argue that the effect of empowerment on unit performance is only significant when uncert ainty is high, because the impact of empowerment is mostly realized through learning and proactive orientations among employees. In an empirical study, Cordery, Morrison, Wright, and Wall (2010) examined whether unpredictable problems would interact with t ask autonomy to influence team performance. The authors argue that task uncertainty raises non routine issues that require higher decision latitude to allow teams to choose non routine method of solutions. Otherwise, team performance will be decreased by t ask uncertainty. Their study found that there was a negative effect of task uncertainty on task performance before autonomy was given to the team. Nevertheless, after an intervention was implemented to provide the team with more autonomy, the negative effe ct of task uncertainty was alleviated. In addition, Gibson (1999) found that only when team task environment is predictable, team efficacy was positively related to team performance. She expected this effect basing on the rationale that when task was highl y uncertain, the strategy chosen based on team efficacy could not be efficiently carried out. Hollenbeck et al. (2002) proposed and tested a model on how structural fit affected team performance. Hollenbeck et al. argue that there are two types of structu ral fit in terms of aligning team environment, team task structure, and team member composition in ability and personality. External fit is the fit between environment and team task structure. When team task environment is unstable, additive (i.e., divisio nal) task structure is more suitable (i.e., a better fit) for team performance than conjunctive (i.e., functional) structure. Internal fit is the fit between team task structure and team member composition in ability and personality. When team task structu re is additive (vs. conjunctive), high cognitive ability is required for all team members to perform

PAGE 40

40 well, because individuals need to be responsible for all aspects of the task. When team task structure is conjunctive, high agreeableness is positively rel ated to team performance, because coordination and cooperation are of higher quality for a group of agreeable individuals. The authors further argue that when there is low external fit (i.e., misalignment between task environment and team task structure), neuroticism is negatively related to individual performance because neurotic people are more easily affected by unstable environment. The study found that teams with additive structure performed better under unpredictable environment whereas teams with con junctive structure performed better under predictable environment. Joint Effects of Team Leadership, Team Composition, and Team Design: Previous Findings In this section, I review previous studies that have attempted to examine the joint effects of team leadership, team composition, and team design on team goal pursuit. The review in this section provides empirical foundations for later development and evaluation of the formal model, particularly regarding the joint effects of the three sources of influen ce on team performance. Team Composition and Team Leadership Randall et al. (2011) examined and found that team composition in cognitive ability and psychological collectivism, and sense giving provided by external leader had independent effects on team pe rformance. However, Zhang and Peterson (2011) found that transformational leadership interacted with team average and variation in CSE to influence team performance. Specifically, Zhang and Peterson found that there was interaction effect between transform ational leadership and team composition in CSE on team advice network density, which in turn influenced team performance. Zhang and Peterson argue that team composition is s all have similarly high CSEs (i.e., low variation and high average in CSE), they are more receptive

PAGE 41

41 affective responses to leadership behavior are similar. In contra st, groups with high heterogeneity advisory ties by having more conflict and less psychological safety. One possible reason for the conflicting findings between Randall et al. (2011) and Zhang and Peterson (2011) is that in Randall et al. (2011), only sense giving behavior of team leaders was examined, which might not have captured those team leadership behaviors that could magnify or complement the effects of team compos ition in cognitive ability and psychological collectivism on team performance (e.g., study, transformational leadership, a construct with broader conceptual space th an sense giving, was examined. Thus, it might be easier to detect the interaction between team leadership and team composition on performance with this broader operationalization of team leadership. Team Design and Team Leadership Team task characteristi cs and team leadership. As Edmonson (2003) pointed out, studying the role of leadership over multiple team tasks in different situations originates from situational or contingent perspective on leadership (Hersey & Blanchard, 1977; Vroom & Jago, 2007). The core idea of situational leadership is that to lead effectively leadership behaviors should vary across different situations according to the demands of task characteristics and adership behaviors should change according to the changing task characteristics and team member composition in knowledge and skills (Edmondson, 2003; Klein et al., 2006). Few empirical studies have examined the role of team leadership over multiple team pe rformance cycles. Klein et al. (2006) qualitative ly examined how leaders of surgical teams allocate the active leadership responsibility to different team members over time. Surgical teams are one type of action teams, which perform

PAGE 42

42 unpredictable and highl y consequential tasks interdependently. In addition, the composition of surgical teams constantly changes. These characteristics make surgical teams particularly suitable for researchers to explore the dynamic processes involved in team goal pursuit. Klein et al. suggest s is beneficial for team performance as well as developing junior team leaders, particularly when the team task demands and team composition are both changing constantly. Morgeson and colleagues (Morgeson, 2005; Morgeson & DeRue, 2006) qualitatively and effectiveness were influenced by characteristics of the unexpected events encountered by the team (event novelty, disruption, criticality, urgency, and duration). Based on functional leadership theory (McGrath, 1962), they proposed and found that team leadership has a larger impact on team performance when the events were novel or disruptiv e because teams had less experience with these events. Further, team leadership played an important role in helping reduce the discrepancy between team competency and team task requirement. In addition, these authors found that the amount of team leader in tervention varied from event to event, depending on the characteristics of the disruptive events. Team task structure and team leadership. Liden et al. (2006) argue that when there is high task interdependence in the team (i.e., teams with conjunctive task structure), team performance is influenced by the level of coordination and strategic distribution of resources enacted by the leader. Therefore, LMX differentiation, a way through which team leaders can allocate resources strategically, can increase team performance. Supporting their argument, these authors found that LMX differentiation was positively related to team performance for teams with high task interdependence, but there was not a significant relationship between LMX

PAGE 43

43 differentiation and team per formance for those teams with low task interdependence. In addition, Chen et al. (2007) found that the effect of empowering leadership behavior on team motivational state and team performance was stronger when task interdependence was higher. These authors argue that one explanation for the moderating effect of interdependence is that teams with high motivational states. Thus, when task interdependence is high, individ actions are more likely to be coupled together to work toward the team goal. Task environment uncertainty and team leadership. Research on top management team consistently found that the relationship between behaviors of the leaders hip team and firm effectiveness was contingent on the stability of task environment (Agle et al., 2006; Cannella et al., 2008; Haleblian & Finkelstein, 1993). For example, for organizations in industries with high unpredictability, performance of top manag ement team was better when CEOs demonstrated directive leadership rather than empowering leadership behaviors, whereas the opposite was found for organizations in industries with a relatively stable environment (Hmieleski & Ensley, 2007). At the team level , Carmeli et al. (2011) found that the positive relationship between empowering leadership and team efficacy was stronger when team members perceived that the environmental uncertainty was high. Joint Effects of Team Composition, Team Design, and Team Lea dership Based on the path goal theory of leadership (House, 1971), the substitutes for leadership theory (Howell, Bowen, Dorfman, Kerr, & Podsakoff, 199 0 ), and the situational leadership theory (Hersey & Blanchard, 1977), Yun, Faraj, and Sims (2005) argue that leadership should have a weaker effect on team performance when subordinates have sufficient information and experience to accomplish the task given a task difficulty level. Therefore, team composition in experience and team task difficulty should in teract with team leadership to impact team

PAGE 44

44 performance. When teams are composed by highly experienced members and when the tasks allow longer time to accomplish, empowering (vs. directive) leadership should result in higher team performance. In contrast, w hen team members are inexperienced and there is not sufficient time allowed, directive (vs. empowering) leadership is more effective because it saves arguments, Yun et al. found that when team experience was low and task difficulty was high, directive leadership led to higher team effectiveness than empowering leadership; when team experience was high and task difficulty was low, empowering leadership was more effect ive in influencing team performance than directive leadership. It is important to note that despite the empirical evidence reviewed here, there is no integrated theoretical framework that delineates the mechanisms through which team composition, team desig n, and team leadership simultaneously and interactively influence team goal pursuit. Therefore, in the next section, I propose a more integrated formal model on how team leadership, team composition, and team design shape the dynamic team goal pursuit proc ess.

PAGE 45

45 Table 2 1. A summary of previous team leadership theories and research. Theory Leadership variable Mechanism for team leadership team performance relationship Cognitive Motivational Affective Behavioral Functional leadership (McGrath, 1962) Tra nsition phase: compose team, define mission, establish expectations and goals, structure and plan, train and develop team, sensemaking, provide feedback Action phase: monitor team, manage team boundaries, challenge team, perform team task, solve problems, provide resources, encourage team self management, support social climate Team member knowledge and skills Performance strategy Conflict control Intrateam coordination Shared mental models (Randall et al., 2011) Collective efficacy Emotion control norm s Interteam coordination (DeChurch & Marks, 2006) Collective information processing Task cohesion Emotional contagion/shared affect Information sharing team metacognition Effort spent on task

PAGE 46

46 Table 2 1. Continued. Theory Leadership variable Mechanism for team leadership team performance relationship Cognitive Motivational Affective Behavioral Transformational leadership (Bass, 1985; Podsakoff et al., 1990) Average and variation of idealized influence, inspirational motivation, i ntellectual stimulation, individualized consideration, supporting followers toward goal accomplishment, encouraging cooperation Mean perceived follower organization value congruence (Hoffman et al., 2011) Affect based trust in leader (Schaubroeck et al., 2011) Information sharing (Zhang & Peterson, 2011) Goal importance variation (Colbert et al., 2008) Trust in team (Gundersen et al., 2012) Elaboration of task related information (Kearney & Gebert, 2009) Team identification (Kearney & Gebert, 2009; Wang & Howell, 2012; Wu et al., 2010) Stress Knowledge sharing (Srivastava et al., 2006) Team empowerment/ potency/efficacy (Bass et al., 2003; Cole et al., 2011; Schaubroeck et al., 2007; Schaubroeck, et al. 2011; Sivasubramaniam et al., 2002; Wu et al., 2010) Psychological safety (Schaubroeck et al., 2011) Team culture (Nemanich & Vera, 2009) Leader identification variation (Wu et al., 2010) Cognition based trust in leader (Schaubroeck et al., 2011) Self efficacy diverge nce (Wu et al., 2010)

PAGE 47

47 Table 2 1. Continued. Theory Leadership variable Mechanism for team leadership team performance relationship Cognitive Motivational Affective Behavioral Leader member exchange (Graen & Scandura, 1987) Mean LMX, median LMX, LMX differentiation Team potency (Boies & Howell, 2006) Communication Effort spent on task Perceived fairness Leadership of self managing teams (Manz & Sims, 1987) Empowering leadership, goal setting, communicating expectations, monitor t eam performance, facilitate problem solving, prepare team to cope with disruptive events, sensemaking Team potency/empowerment (Carmeli et al., 2011; Chen et al, 2007; Kirkman & Rosen, 2006) Problem management actions (Tesluk & Mathieu, 1999) Knowl edge sharing (Srivastava et al., 2006)

PAGE 48

48 CHAPTER 3 A FORMAL MODEL OF TEAM LEADERSHIP In this chapter, I propose a formal team leadership model drawing on two theories: functional leadership theory (McGrath, 1962) and self regulation theory (Carver & Sc heier, 1998; Vancouver, 2008). As I discussed earlier, functional leadership theory is the most suitable leadership perspective for developing an integrated dynamic model on team leadership. However, despite identifying key leadership functions in team dev elopment and team performance stages (Morgeson et al., 2010; Kozlowski et al., 1996; Zaccaro et al., 2001), functional team leadership theory does not offer explanations regarding how leadership influences goal pursuit over time, neither does functional te am leadership theory incorporate team composition and team design features into the team goal pursuit process. In addition, the current treatment of the reciprocal interplay between team performance state and team leadership (Morgeson, 2005; Morgeson & DeR ue, 2006) has not clarified the underlying mechanisms that link team leadership, team goal choice, and team goal striving. Therefore, to account for these dynamic processes involved, I apply self regulation theory to further our understanding of leadership in team goal pursuit. S elf regulation theory views individual performance system and leadership as subsystems in a larger performance system (i.e., the team). These subsystems dynamically regulate the critical states and actions of themselves and of each other (Vancouver, Tamanini et al., 2010). The collective goal choice and goal striving processes enacted by these subsystems follow the structure of negative feedback loops, in which effort to change the current state is determined by the discrepancy betwe en current and desired states. For example, when there is a discrepancy between expected and actual task states of a subordinate, both the leader and the subordinate can r action can

PAGE 49

49 be viewed as efforts spent on the task driven by perceived discrepancy, i.e., the self regulation viewed as to act on what the subordinate is not doing adequately (Morgeson et al., 2010), which can be conceptualized as other regulation or manager regulation process (Vancouver, Tamanini et al., 2010). The same logic as in this example can be used to conceptualize how leadership and individuals in the competencies and accomplishing a collective goal. As such, self regulation theory can provide a unified framework for conceptualiz ing regulatory actions enacted by leadership and a group of individuals (i.e., team members) simultaneously in team goal pursuit. In this res e arch I use the general negative feedback loop formally delineated by Vancouver ( 2008) as the building block to account for the dynamics among multiple subsystems. S pecifically, self regulation theory views processes underlying goal pursuits as a set of discrepancy reducing negative feedback loops hierarchically organized in the same system (Carver & Scheier, 1998; Powers, 1973). The general form of the negative feedb ack loop can be illustrated by the diagram shown in Figure 3 1. In the negative feedback loop, current state of a variable ( v ; e.g., the current team task state) is incorporated in the system by an input function. Assuming that the perception is formed dir ectly and accurately, perception of the current state ( p ) can be defined as identical to the current state of v . Goals are represented as desired states of a variable ( p* ; e.g., the desired team task state). The perception of the current state is then comp ared to the desired state by the comparator, which gives discrepancy ( d ) between the perceived state and the desired state. Comparator is generally asymmetric, which gives zero when desired state is less than perceived state. Discrepancy together with an a mplifying or attenuating factor, called gain ( k ), determines the output ( o ; e.g., putting effort into the team

PAGE 50

50 task). Output in turn changes the current state of the variable at a certain rate ( r ; e.g., how fast a team member makes progress on team task). The current state of a variable can also be influenced by external influences that are called disturbances ( D ). When change in the state of the variable is modeled in continuous time, the current state v at time t can be represented using an integral funct ion. In a negative feedback loop, input, comparator, and output functions together are referred to as one regulatory agent or regulatory (sub)system, which operates upon the external environment to obtain or maintain its desired state. Based on this gener al structure of self regulation theory (Vancouver, 2008) and the two major types of leadership functions identified by functional team leadership theory (Morgeson et al., 2010), I propose a formal model on team leadership. As illustrated in Figure 3 2, the team regulat ion system is composed by two sub systems: team development system and team performance system. Each of these two systems consist multiple lower level systems ( i.e., subsystem s ) and is subject to the joint influence of leader individual differe nces, team composition factors, and team design factors. The team development system includes team member regulated and leader competency level. The desired amount of competency is influenced by the variety of skills required for accomplishing the team task (i.e., a team design factor). How fast a team increases two characteristics that form team c omposition factors). The extent to which team leaders spen d effort and time in developmental leadership behaviors is influenced by leader reactivity and leader sensitivity to relative discrepancy among subordinates (i.e., leader individual difference facto rs). The team performance system includes team member regulated and team leader regulated performance systems, which together regulate task state of the team. Team

PAGE 51

51 des ign factors (team task structure, assigned team goal, deadline for team task, job complexity, and problem solving), disturbances from the environment, leader reactivity, and leader sensitivity to relative discrepancy. I detail the structure and process in each system of this formal model in the following sections. Team Development System Team development system is illustrated in Figure 3 3. Team development system includes two subsystems: team member regulated and team leader regulated learning systems. Bot leadership behaviors. Through these two subsystems, teams develop their compet encies to meet the level of competencies required by the task (Marks et al., 2001; Morgeson et al. 2010; Vancouver, Tamanini et al., 2010). The output from the team development system, i.e., team ams progress toward the team goals (Barrick et al., 1998; LePine et al., 1997). Team Member regulated Learning System As illustrated in Figure 3 3, team member j k ) learning system regulates his/her competency level over time ( member competency ijk t , with i indicates the i th team performance episode, ranging from 1 to I; j indicates the j th subordinate/position, ranging from 1 to J; k indicates the k th team/leader, ranging from 1 to K; t equals time since beginning of the first team per formance episode ) . Specifically, the current level of individual competency is perceived member compet ency bias jk ). Perception of current level of individual competency is compared to the desired level of

PAGE 52

52 competency ( desired member competency ij k ) by the competency comparator (Vancouver, Tamanini et al., 2010). Desired competency is a function of the skill variety required of the task to which the individual is assigned (Morgeson & Humphrey, 2006). Skill variety is defined as the extent to which the task requires different sets of skills and knowledge to perform (Morgeson & Humphrey, 2006). When task require s higher skill variety, desired competency for accomplishing a specific task should be higher. Discrepancy between current and desired individual engages in learnin g behavior ( member learning behavior ijk t ). Another factor that may influence the extent to which individuals engage in learning behavior is conscientiousness of the individual. Some team members may be more motivated to learn due to their dispositional tra its, as compared to others (Colquitt, LePine, & Noe, 2000). Team member learning behavior in turn Team Leader regulated Learning System the desired competency level. The discrepancy influences the amount of developmental leadership behaviors that a leader performs t o influence the team members (Vancouver, Tamanini et al., 2010). The extent to which a leader engages in training and developing a particular subordinate leader reactivity k ). Dispo sitional differences of a leader describe the consistent ways in which the leader mobilizes

PAGE 53

53 the subordinates toward goal accomplishment. Research on leader traits, behavioral styles, and leader subordinate relationship suggest that leaders may tend to enga ge in leadership behaviors to different levels based on personal characteristics (e.g., Anand, Vidyarthi, Liden, & Rousseau, 2010; Graen & Uhl Bien, 1995). Importance of the position to which the subordinate is assigned ( position importance ij k ) might also influence leadership behaviors in the developing system. When a position is more critical to team effectiveness than others, team leader is more likely to provide critical resource to the subordinate on the position to ensure that the subordinate is qualif ied for the job (Humphrey et al., 2009). In addition, whether leaders engage in behaviors to regulate a subordinate at a particular time point is influenced by which target the leader chooses to regulate. Which regulation target is chosen is influenced by well ( the specific function s are described w hen introducing team leader regulated performance system ) . competency level. However, the rate of change in individual competency may differ from person Learning ability from self regulated learning and provided by external sources, which can be operationalized as ability is an individual difference factor that is influenced by gen eral cognitive ability (i.e., g ; Vancouver, Tamanini et al., 2010). In skill acquisition processes, individuals differ in their capacity to encode information which results in different levels of learning within a given time period (Kanfer & Ackerman, 1989 ). Team composition literature also suggests that general cognitive ability of a team is related to team performance because teams differ in their ability to

PAGE 54

54 learn and absorb new task related information (LePine, 2003; LePine, Colquitt, & Erez, 2000). Ther efore, learning ability can be a key team composition factor affecting the team development ( competency disturbance ijk t ) . For example, when organizations switch the equipment or technologies, domain knowledge about the specific prior equipment or technologies are no longer relevant, which can result in reduced team member competencies. Team Performance System Through team performance system, teams strive toward their goals. Team performance system includes two subsystems: team member and team leader regulated performance systems (illustrated in Figures 3 4 and 3 5), which maintains and regulates individual and team task states. Team leader regulated performance system controls task contingent team leadership behaviors, sets new task goals for the team, and decides whether team as a whole can move onto the next performance cycle (Morgeson, 2005; Morgeson & DeRue, 2006; Morgeson et al., 2010). Inputs to the team performa nce system include team composition in conscientiousness, job complexity, problem solving, assigned team goal, deadline s , team task structure, disturbances from the environment, leader reactivity, and leader sensitivity to relative discrepancies among subo rdinates. Team Member regulated Performance System As illustrated in Figure 3 4, team member regulated performance system includes five lower lever subsystems: individual performance, individual expectancy, individual task choice, team performance, and te am expectancy subsystems. These five lower level subsystems together regulate individual and team task states. Individual performance system. Individuals perceive their own task states ( individual task state ijk t ) through the individual task state input fun ction (Vancouver, Weinhardt et al., 2010).

PAGE 55

55 Perceived individual task state is compared with desired task state (i.e., individual goal ij k ). Discrepancy between perceived and desired task states decides individual task state output. Individual task state output is the input to individual expectancy and task choice systems, which decide s the perceived utility of the focal task. Outputs from the task choice agent ( member task choice output ijk t ) decide how much effort the individual puts into working on the task. In addition to individual themselves, team leaders can also actively influence individual task state through task contingent leadership behaviors. How fast efforts from team leaders and individual themselves transform into progress on the task is a function of task difficulty and individual competency. Individual competency decides how fast the individual can work on the task. Task difficulty, which is influenced by job complexity and problem solving required by the task (Morgeson & Humphrey, 20 06), decides how many efforts are needed to make the same amount of progress in a given amount of time. Individual task state is also subjective to external disturbances ( task disturbance ijk t ). Negative external disturbances can set back the progress made toward the goal. Individual expectancy system . Vancouver, Weinhardt et al. (2010) proposed a formal model that provides a dynamic account of expectancy and valence in individual goal pursuit over time. The current model follows the structure they proposed . Specifically, expectancy input is a function of how fast an individual perceives him/herself can progress on a task ( i.e., member perception of competency ijk t ) and the discrepancy between current and desired states of the task. The ratio between discrepa ncy and competency decides the expected time needed to complete the task. Expectancy comparator compares the expected time needed and the perceived time left ( accepted deadline ij k time ). The bigger the difference between expected time needed and perceive d time left, the larger individual expectancy ( member individual expectancy output ijk ).

PAGE 56

56 Considering that lack of expectancy can result in task withdrawal (Carver & Scheier, 1998), the value of expectancy is always equal to or larger than zero. Individual task choice system. Given that team members might have responsibilities that are not directly related to team goal while performing on team tasks, task choice system controls how likely individuals spend efforts on task rather than off task activities. Bas ed on Vancouver, Weinhardt et al. (2010), task choice agent serves the function of comparing utilities of the focal task and off task activities. In the current model, utility of off task activities is considered as a summarized value of all other alternat other than working on the team task. Utility of the focal task is derived from multiple perceptions, including individual and team expectancy outputs, discrepancy of the current and desired task states, and task valence to the individual. Expectancy theory argues that expectancy and valence can both increase efforts spent on task performance (Vroom, 1964). Translated into utility measure, expectancy and valence can increase the perceived utility of working on the task as compared to other choices of action. Discrepancy in task state can increase the utility of performing on the task as well. There is larger utility from reducing a larger discrepancy (Vancouver, Weinhardt et al., 2010). Task valence, defined as the value of the outcome from the characteristics. The present study focuses on examin ing conscientiousness. Individuals who are more conscientious ar e likely to see a larger value in accomplishing the goal given their dispositional tendency for accomplishment striving (Barrick & Mount, 1991). Team performance system. task state, the form of which d epends on team task structure. There are three typical team task structures that represent three different levels of task interdependence, as identified by Steiner

PAGE 57

57 (1972). When team task is additive , all positions in the team are responsi ble for team task is conjunctive , each position in the team is responsible for a single unique function. Team task is accomplished only when every function accomplishes its duties. When teak team is disjunctive, each position is responsible for the same set of functions but team task state is only determined by the task state of the best performer. In other words, the whole team accomplishes th e team task when one person accomplishes his/her task. For example, consider a consulting team that is composed of four subordinates. When team task structure is additive, every subordinate is responsible for independently solving one fourth of the total amount of problems from the client. The team accomplishes the team goal when all four members accomplish their individual goals. When team task structure is conjunctive, every subordinate is responsible for providing a unique type of service (e.g., legal, accounting, finance, and operation) to the client. The team accomplishes the team goal when all four team members accomplish the specific types of service for the client. When team task structure is disjunctive, every subordinate is responsible for all the problems from the client. The team accomplishes the team goal when any one of the four team members finishes solving all the problems for the client. In the formal model proposed, team task structure is specified as different composition models that aggre gate individual states (e.g., competency, expectancy, and task state) to the team level. Team expectancy system. Based on the formal model of dynamic individual expectancy in Vancouver, Weinhardt et al. (2010) , I propose a dynamic conceptualization of tea m competency ( member perception of team competency ijk t ) and the discrepancy between current

PAGE 58

58 task state and desired task state ( i.e., team goal ij k ). Th e ratio between team task state discrepancy and perceived team competency decides the expected time needed to complete the team task. Team expectancy comparator compares the expected time needed and the perceived time left ( accepted deadline ij k time ). The larger the difference between expected time needed team expectancy output ijk t actions. Team Leader regulated Performance System In addition to team member regulated performance system, team leader regulated performance system also actively regulates individual and team task states. Three specific functions are carried out by team leader regulated performance system: directing leaders to engage in task contingent leadership behaviors, setting team goals, and deciding when the team can move onto the next task cycle. As illustrated in Figure 3 ate to the desired task state (i.e., leader assigned goal ij k ). If there is a discrepancy, team leaders can engage in task contingent leadership behaviors (Hackman & Walton, 1986; Morgeson et al., 2010; Mumford, Campion, & Morgeson, 2006). The extent to whi ch leaders engage in task contingent leadership behaviors is also influenced by leader reactivity. Team leadership literature suggests that team leaders who prefer hands on coaching instead of empowering and enabling are more likely to intervene sooner and engage in a broader range of task activities themselves (Hackman & Wageman, 2005; Wageman, 2001). In addition, importance of the positions in the team is also likely to influence the priority of the tasks that the leaders take over. Research on action tea ms suggest that when team leaders perceive team members occupying strategically important positions are relatively less experienced, they are more likely to take over those

PAGE 59

59 assigned responsibilities first (Klein et al., 2006; Yun et al., 200 5 ). In addition , similar to individual team members, leaders are more motivated to engage in task contingent leadership behaviors when leaders perceive higher team expectancy. At a particular time point, leader can only engage in regulatory actions toward one target. For a given leader, when relative discrepancy among the subordinates is below his/her threshold (defined as leader sensitivity to relative discrepancy k ), leader is likely to regulate the team as a whole. However, when one of the task state is higher than the threshold, team leader is likely to switch the regulatory focus to the individual subordinate. In team member regulated performance system, individual j set goal (i.e., individual goal ij k ) is influenced by le ader assigned goal. A team leader can facilitate team members to internalize leader assigned goal by defining team goal clearly to everyone in the group. Transformational leadership theory argues that by defining the collective goal, leaders can help subor dinates understand the specific level of performance required of them (Bass , 1985 ). Leader team performance episode. When the team finished the last task before the deadline (i.e., spen t less time than expected) or did not meet the deadline (i.e., failure to accomplish the goal on time), it is likely for team leaders to set a more difficult or easier goal for the team as compared to the previously assigned team goal. Team leaders also c ontrol when team transits from one team task to the next one. Team leaders compare their perceived team task state on a given task with its desired state. If the discrepancy is larger than zero, team leaders are likely to command the teams to stay on the c urrent task. When team leaders perceive that there is no discrepancy between the desired and current team task states, team leaders can direct the team into working on the next task. F or team

PAGE 60

60 utput can direct task number system to change the current task number from i to i +1. In addition, when time reaches the deadline and team task is still not completed, team task number may also increase from i to i +1. In this case, task i is considered as i ncomplete (i.e., goal pursuit fails). The transition process is illustrated in Figure 3 6. The Development of Dyadic Relationship I also incorporate leader member dyadic relationship into the formal model (illustrated in Figure 3 1) based on the social exc hange theory (Blau, 1964; Foa & Foa, 1980). Social exchange theory argues that interpersonal relationship develops from a series of exchange s between two attitudes, a nd interactions with each other (Cropanzano & Michell, 2005). In other words, there is a reciprocal influence between relationship developed and the way exchanges are handled between two persons. Applying this perspective to the team leadership setting, te am leader subordinate relationship and their interactions can be viewed as constantly reinforcing each other. The current model specifies LMX as a product developed from leader subordinate similarity, developmental and performance related regulat ory behaviors directed at the interactions between leaders and subordinates. Specifically, the baseline of LMX is influenced by leader subordinate similarity. Prev ious research suggested that leader subordinate similarity in deep level characteristics can influence leader treatment of subordinates and LMX (Bauer & Green, 1996; Sparrowe & Liden, 1997; Zhang, Wang, & Shi, 2012). The amount of regulatory behaviors lead ers spent on a particular subordinate can also influence the quality of dyadic

PAGE 61

61 quality (Bauer & Green, 1996). According to social exchange theory, the relationship formed between supervisors and subordinates can in turn influence how likely supervisors interact with subordinates. Assumptions of the Model Th is formal model has several theoretical assumptions that together define the boundaries of the model and justify which constructs are included or excluded from the model. First, it assumes that the teams have already been formed and the team membership is relatively stable over the multiple team performance cycles examined. Each individual team member is allocated t o a fixed role. Therefore, composing team and allocating team members to different roles (two functions of team leadership identified by Morgeson et al., 2010) and team member exit (i.e., turnover) are not specified in the model. Second, this model assumes that the team interacts with other teams in a stable pattern. Therefore, the part of team leadership that manages the team boundary related functions is not specified in the model. Events in the external environment that are not routine to the team are sp ecified as disturbances to individual team members and the team as a whole. Third, in terms of interactions among team members, the model assumes that the team members interact with each other in a relatively stable manner. Therefore, within team interacti ons and team leadership functions that affect social interactions (e.g., developing team climate) are not specified in the model.

PAGE 62

62 Figure 3 1 . The general structure of negative feedback loop . Note. This figure is adapted f rom Vancouver, Weinhardt et al. ( 2010). Variables in t he illustrated model are formally specified by the following equations. p = v If ( p * p ) > 0, then d = ( p * p ); else, d = 0 o = kd

PAGE 63

63 Figure 3 2 . A formal model of team leadership.

PAGE 64

64 Figure 3 3. Team development system. Note . Variables in this figure are specified by the following equations. i indicates the i th team performance episode , ranging from 1 to I . j indicates the j th subordinate/position , ranging from 1 to J . k indicates the k th team/leader , ranging from 1 to K . t = time since beginning of the 1 st team performance episode.

PAGE 65

65 i f , then ; else, 0 if , then ; else, 0 if , then ; else, if , then ; else, if (| < , then ; else, if [ ], then ; else if [ ], then ; else

PAGE 66

66 k1 = 0 k2 = 0

PAGE 67

67 Figure 3 4. Team member regulated performance system. Note. Variables in this figure are specified by the following equations. i indicates the i th team performance episode, ranging from 1 to I. j indicates the j th subordinate/position, ranging from 1 to J. k indicates the k th team/leader, ranging from 1 to K. t = time since beginning of the 1 st team performance episode. +

PAGE 68

68 if , then ; else, 0 if , then ; else, 0 if , then ; else, 0 + if , then ; else, 0

PAGE 69

69 if team task structure , then ; else, if team task structure , then ; else, if team task structure , then ; else, if team task structure = , then ; else, if , then ; else, 0 if , then ; else, 0

PAGE 70

70 team task structure

PAGE 71

71 Figure 3 5. Team leader regulated performance system. Note. Variables in this figure are specified by the following equations. i indicates the i th team performance episode, ranging from 1 to I. j indicates the j th subordi nate/position, ranging from 1 to J. k indicates the k th team/leader, ranging from 1 to K. t = time since beginning of the 1 st team performance episode. +

PAGE 72

72 if , then ; else, 0 if , then ; else, if , then ; else, + k3 + if , then ; else, 0 if , then 0; else, 1 if team task structure , then ; else, if team task structure , then ; else,

PAGE 73

73 team goal = 1 k3 = 0

PAGE 74

74 Figure 3 6. Transition between team performance cycles. Note. Variables in this figure are specified by the following equations. i indicates the i th team performance episode, ranging from 1 to I. j indicates the j th subordinate/position, ranging from 1 to J. k indicates the k th team/leader, ranging from 1 to K. t = time since beginning of the 1 st team performance episode. if , then ; else, 0 if , then ; else, 0 total num ber of task = 1

PAGE 75

75 CHAPTER 4 STUDY 1: COMPUTATIONAL MODEL SPECIFICATION AND EVALUATION The purpose of Study 1 is to specify and qualitatively e v aluate the computational model. Computational models are the ories formally described by computational language (Tab er & Timpone, 1996). Therefore, Study 1 programmed the model by using equations to specify mathematical relationships between the constructs in the model. Model data fit of the c omputational model can be assessed qualitatively and quantitatively. Qualitati ve model evaluation compares the computational modeling results with inferences drawn from empirical data (see Vancouver, Ta manini et al., 2010 for an example). Outcome validity of the model is dictions) from the co mputational modeling results are consistent with the empirical findings (Taber & Timpone, 1996). This is the focus of Study 1 . In Study 2, longitudinal data were collected from participants in a lab environment to provide an opportunity to quantitative ly e valuate the model. Computational Model Specification A computational model was specified according to the description of the model in Chapter 3. The model (a) includes formal representations of key motivational constructs, (b) describes the individual goal pursuit process over multiple tasks, (c) specifies team leadership behaviors in the team goal pursuit process over time, and (d) specifies the relationships between team leadership behaviors, team design factors, team composition, and team performance ove r time. Mathematical e quations used to define all the constructs in the model are provided under the figures listed at the end of Chapter 3. All codes used to program the model using the software Vensim Professional are reported in Appendix A.

PAGE 76

76 Qualitative Model Evaluation To qualitatively evaluate the model proposed, I examined whether the model behaved consistently with key empirical findings reported in the literature. I manipulated the values of the input factors (independent variables) to the model and observed the values of the endogenous variables (dependent variables) generated through mechanisms specified in the model . For the input factors, I manipulated three levels of team task structure (functional, divisional, and disjunctive), two levels of te am member composition in learning ability (same and different learning ability) with the average learning ability being constant, two levels of leader reactivity (low and high leader reactivity), two levels of environmental disturbance magnitude (small and large disturbance), and two levels of disturbance duration (short and long duration). Each run of the simulation lasted for 30 time units. Following suggestions from Ilgen and Hulin (2000) , all input factors were manipulated simultaneously so that both ma in effects and joint effects of the input factors on the endogenous variables could be analyzed. Together, there were 48 combinations of input factors. Effect of Disturbance Magnitude Morgeson and DeRue (2006) examined whether the amount of intervention fr om team leader is influenced by the characteristics of the unexpected events encountered by the team. Their study asked subordinates to recall past events and rated criticality ( ), urgency (e.g , and duration ( ) of the unexpected events that the team faced . They also asked the team leaders to assess how di the time they spent intervening the team s . Their study found that all three characteristics were related to disruptiveness of the events rated by the

PAGE 77

77 leaders and that urgency of unexpected events was positively related to the proportion of team leader s spent on intervention. Their findings suggested that the magnitude of external negative disturbance cou ld be positively related to team leader ship regulatory behaviors. I n the computational modeling study, I manipulated the magnitude of negative external disturbances input to the model and observed the outputs of team leadership behaviors. Two levels of disturbance magnitude were man ipulated: sm all and large. Small disturbance was specified as deducting 5% from the task already accomplished by the team. Large disturbance was specified as deducting 10% from the task already accomplished by the team. Computational modeling results for the effects o f disturbance magnitude on task contingent team leadership behavior and team task state over time are illustrated in Figure 4 1. Computational modeling results suggested that there was more task contingent team leadership behavior when disturbance magnitud e was large (vs. small). Therefore, computational modeling results were consistent with In addition, computational modeling results suggested that the discrepancy between team goal and team task state increased more a fter the disturbance when disturbance magnitude was large (vs. small). As shown in Figure 4 1, computational modeling results also suggested that the difference between small and large disturbance conditions in task contingent team leadership behavior was larger when disturbance happened closer to deadline. In other words, time left to accomplish the task moderated the positive relationship between disturbance magnitude and task contingent team leadership behavior, such that the positive relationship was st ronger when disturbance took place closer to the deadline. Effect of Disturbance Duration Morgeson and DeRue (2006) found that duration of negative disruptive events was positively related to the disruptiveness of unexpected events fac ed by the teams . Thei r qualitative

PAGE 78

78 study also suggested that team leaders might need to make more efforts to prepare the teams to cope with longer unexpected events. In the computational modeling study, I manipulated two levels (short vs. long) of the duration of negative exte rnal disturbances and observed the outputs of team leadership behaviors. Short disturbance happened during time units 11 to 13 and during time units 21 to 23. Long disturbance happened during time units 11 to 15 and time units 21 to 25. Computational model ing results for the effects of disturbance duration on task contingent team leadership behavior and team task state over time are illustrated in Figure 4 2. Computational modeling results suggested that there was more task contingent team leadership behavi or when the disturbance duration was long (vs. short). Therefore, computational modeling on this aspect . In addition, computational modeling results suggested that the discrepancy between t eam goal and team task state increased more after the disturbance when disturbance duration was long (vs. short). As shown in Figure 4 2, computational modeling results also suggested that the difference between short and longer disturbance conditions in t ask contingent team leadership behavior was larger when disturbance happened closer to the deadline. In other words, time left to accomplish the task moderated the positive relationship between disturbance duration and task contingent team leadership behav ior, such that the positive relationship was stronger when disturbance took place closer to the deadline. Effect of Leader Reactivity Previous studies (e.g., Cole et al., 2011; Kark , Sharma, & Chen, 2003; Shin & Zhou, 2007; Wu et al., 2010) found that disp ositional characteristics of team leaders (e.g., transformational leadership style, empowering leadership style) was positively related to team efficacy, which in turn was positively related to team performance. In the computational model, I manipulated th e input factor that represents individual differences of team leaders in reacting to

PAGE 79

79 the discrepancy and observed its effects on team leadership behaviors, team expectancy, and team task state. Two levels (low vs. high) of leader reactivity were specified , such that leaders with higher reactivity would engage in more regulatory actions given the same amount of discrepancies between team goals task states. Computational modeling results for the effects of leader reactivity on task contingent team leadership behavior, team expectancy, and team task state over time are illustrated in Figure 4 3. Computational modeling results suggested that there was more task contingent team leadership behavior, higher team expectancy, and higher te am performance over time when leader reactivity was high (vs. low). Therefore, computational modeling results were consistent with empirical findings regarding the positive associations among leadership style, team efficacy, and team performance. Computat ional modeling results also suggested that the relationships among leader behavior, team expectancy, and team performance were consistently positive over time. However, as shown in Figure 4 3, the strength of the positive relationship between leadership be havior and team performance was stronger at the beginning of the team performance episode than near the deadline. This is because the difference in team leadership behavior between high and low leader reactivity conditions was larger at the beginning of th e team performance episode, whereas the difference in team task state between high and low leader reactivity conditions was relatively consistent over time. Interaction between Leader Reactivity and Team Task Structure Chen et al. (2007) examined whether the effect of team leadership style on team performance differed in teams with high vs. low task interdependence. Their study measured empowering leadership style and team performance of high interdependent freight teams and low interdependent receiving te ams in a home improvement company. Their study found that

PAGE 80

80 empowering leadership style was positively related to team performance only in teams with high task interdependence. These results suggested that team task structure could be a boundary condition on the strength of the relationship between individual differences of team leaders and team performance. In the computational modeling study, I manipulated leader reactivity and team task structure, and observed their joint effects on team leadership behavio rs and team task state. Two levels of leader reactivity were specified: low and high. Three conditions of team task structure were specified: disjunctive, divisional, and functional, which ranged from low to high task interdependence. Computational modeli ng results for the effect of leader reactivity on team leadership behavior and team task state in different conditions of team task structure (i.e., functional, divisional, and disjunctive) are illustrated in Figure 4 4. Results suggested that the positive relationships among leader reactivity, task contingent team leadership behavior, and team performance were stronger when team task structure was functional and divisional (vs. disjunctive). This was partly because the difference in team leadership behavio r between higher and lower leader reactivity conditions was larger when team task structure was functional and divisional (vs. disjunctive). Therefore, computational modeling results were consistent with empirical findings on the moderating effect of team task structure on the positive relationship between leadership styles and team performance (e.g., Chen et al., 2007). In addition, computational modeling results suggested that the effect of leader reactivity on team leadership behavior could weaken over time, especially when team task structure was functional (i.e., high task interdependence). As shown in Figure 4 4, the difference between low and high leader reactivity conditions in team leadership behavior was larger at the beginning of team performance episode than closer to the deadline. However, the change of the difference was

PAGE 81

81 not linear, such that the decreasing of the difference was slower when there were negative external disturbances. Interaction between Team Composition in Learning Ability and Team Task Structure Wu et al. (2010) examined whether differentiated individual focused transfo rmational leadership ( individualized consideration and intellectual stimulation ) was related to divergence among tea m members in self efficacy and group effecti veness (indicated by group performance and group viability) . These authors collected data from full time employees in work groups in different industries, and treated task interdependence as a covariate in their analyses . Wu et al. found that differentiate d individual focused leadership was positively related to differentiation in self efficacy which was negatively related to group effectiveness. In addition, Liden et al. (20 06) found that when task interdependence was high, LMX differentiation was positive ly related to team performance . F indings in Liden et al. suggested that the effect of differentiated leadership on team performance could differ according to the team task structure. S imulating the formal computational model, I examined whether team compo sition in learning ability could be the input factor that led to differentiated leadership behaviors (indicated as SD of time spent on regulating individual subordinate), differentiation in individual expectancy within the team (indicated as SD of individu al expectancy within the team ), and team task state over time. In addition, I examined whether the relationships among these variables differed across team task structure conditions. Two conditions of team composition in learning ability were manipulated: same and different learning abilities among subordinates. Three conditions of team task structure were specified: disjunctive, divisional, and functional, which ranged from low to high task interdependence. Computational modeling results for the effect of team composition in learning ability ( same vs. different ) on differentiated leadership behaviors are reported in Table 4 1. Results

PAGE 82

82 different (vs. same) , there was differentiation in leadership be havior s among team members, which was consistent in all three conditions of team task structure (i.e., functional, divisional, and disjunctive). Computational modeling results for the effect of team composition in learning ability ( different vs. same) on d ifferentiation in individual expectancy and team task state over time are reported in Figure 4 5. The r ies were different (vs. same) , there was differentiation in individual expectancy among team members, which was consistent in all three conditions of team task structure. In addition, when team task structure was functional differentiation in learning abili ty and team performance. However, the negative relationship was weaker when team task structure was divisional and there was a positive relationship between in learning ability and team performance when team task structure was disjunctive (i.e., low task interdependence). Taken together, these results suggested that when team task structure was functional (i.e., in learning ability could lead to higher differentiation i n team leadership behaviors among subordinates as well as individual expectancy, and lower team performance. However, when team task structure was disjunctive higher differentiation in team leadership behaviors among subordinates as well as individual expectancy, and higher team performance. These results provided an alternative explanation for the positive relationship between differentiated individual focused leadership and differentiation in self efficacy and the negative relationship between differentiated individual focused leadership and team effectiveness found in Wu et al. ( 2010). It is possible that the observed

PAGE 83

83 correlations among differentiated leaders hip, differentiation in individual expectancy, and team performance were all outcomes resulting from differentiation among team members in learning ability, i.e., a structural input factor to the team performance system. Moreover, the computational modelin g results suggested the strength and direction of the observed correlations could change depending on the team task structure conditions. Sensitivity Analyses In the sensitivity analyses, I examined the above computational modeling results under different conditional values of individual differences and bias variables, including individual competency bias, initial competency, individual task state bias, utilities of off task activities, esults suggested that patterns of the simulation results reported above were robust within the range of values tested. I list all the parameters examined and the specific values assessed in Table 4 2. Discussion The present study specified a computational model that formally represents the team leadership model proposed in Chapter 3. Qualitative model evaluation results suggested that the model behaved consistently with relevant empirical findings reported in the literature. In addition, the computational modeling results extended existing verbal theories. The present study had several important theoretical implications. First, the computational modeling results suggested that the specified dynamic model could account for important team leadership phenomen a documented in previous studies. In particular, s imulated data generated from the present computational model were consistent with empirical findings about various aspects of team leadership. For example , the computational modeling results were consistent was positively related to the amount of team leader intervention. When the external negative

PAGE 84

84 disturbance is large or lasts long er , team leaders engage in more te am leadership behaviors to help the team counter the negative influence from the disturbance s so that the team can progress toward the goal. Both the findings supported the functional view on t eam leadership (Kozlowski et al., 1996; McGrath, 1962), which describes team leadership as a regulatory system that supports and complements The computational modeling results also accounted for the positive relationship among studies (e.g., S hin & Zhou, 2007). In addition, the computational modeling results explicated that the positive effect of team l (e.g., leader reactivity) on team performance was realized through team leadership behaviors. Furthermore, the computational modeling results accounted for the moderating effect of team task structure on team leadershi p team performance relationship (e.g., Chen et al., 2007). The simulation of the computational model also provided causal explanations for the positive correlations among differentiated individual focuse d leadership, differentiation in individual expectanc y, and team performance found by Wu et al. (2010). Second, the present study helps integrate existing team leadership models and explicitly addresses dynamic issues in team leade rship phenomenon. As reviewed above , a number of team leadership theories hav e been proposed to understand the role of leadership in team goal pursuit pro cess. Different theories focus on different aspects of team leadership. For example, exp ectancy, which in turn influences team performance (e.g., Chen et al., 2007). Some other theories focus on the composition of individual focused leadership in the team context, such as

PAGE 85

85 the differentiat ed individual focused leadership and its relationship w ith differentiation in individual expectancy and team performance (e.g., Wu et al., 2010). Although similar constructs were involved in these theories (e.g., team expectancy and individual expectancy), it was not clear how these theories could be integrate d to achieve a more generalizable and coherent theoretical framework. In addition, mechanisms of the dynamics of leadership and team goal pursuit are missing from existing team leadership theories. The current dynamic model addresses these limitations. The computational modeling results showed that this integrated model was capable of deriving hypotheses that were consistent with predictions fr o m existing theories. Moreover, the current model was described by formal computational language, which helps furth er integrate other theoretical mechanisms (e.g., inter member interaction processes) into the model. It should be noted that the current model achieves the goal of integrating existing theories without sacrificing conceptual parsimony (Kuhn, 1962; Myung, 2000; 2003; Taber & Timpone, 1996). The present model was developed based on the core assumptions of self regulation theory and functional leadership theory. In this way, all hypotheses derived from the model are coherent and consistent. The simplicity of the basic structure of the model also allows the model to incorporate more components (e.g., peer team members serving as one an agent s ) without introducing additional assumptions. Third, the simulation of the computational model suggest ed several directions for future research. Specifically, the simulation showed that under different team task structures, the relationships among team member composition in learning ability, differentiat ed individual focused leadership, differentiation in individual expectancy, and team performance could be different. Given that previous study ( Wu et al. , 2010) did not examine team member composition

PAGE 86

86 in learning ability and team task structure, future studies could test this relationship empirically. In add ition, the present computational model treated bias parameters as constants in the analyses. It is conceivable that certain biases can change over time. For example, leaders might form more accurate perceptions of the subordinates as the time working toget her increases. Future computational or empirical studies could further examine whether these bias parameters changed Furthermore, computational modeling results suggested that some effects found in previous cross sectional research could vary over time. For example, the simulation results revealed that the effects of disturbance magnitude and disturbance on team leadership behaviors were stronger when the disturbance happened closer to the deadline. Future empirical research should observe team leadership beh aviors over the course of a team performance episode to examine this relationship.

PAGE 87

87 Table 4 1. SD of time spent on individual subordinate (per team performance episode ) by conditions. Team Task Structure Team Composition in Learning Ability Leader Reactivity Disturbance Magnitude Disturbance Duration SD of Time Spent Functional Same Low Small Short .0000 Functional Same Low Small Long .0000 Functional Same Low Large Short .0000 Fun ctional Same Low Large Long .0000 Functional Same High Small Short .0000 Functional Same High Small Long .0000 Functional Same High Large Short .0000 Functional Same High Large Long .0000 Functional Different Low Small Short .0577 Functional Differen t Low Small Long .0577 Functional Different Low Large Short .0577 Functional Different Low Large Long .0577 Functional Different High Small Short .1067 Functional Different High Small Long .1036 Functional Different High Large Short .0877 Functional Different High Large Long .0995 Divisional Same Low Small Short .0000 Divisional Same Low Small Long .0000 Divisional Same Low Large Short .0000 Divisional Same Low Large Long .0000 Divisional Same High Small Short .0000 Divisional Same High S mall Long .0000 Divisional Same High Large Short .0000 Divisional Same High Large Long .0000 Divisional Different Low Small Short .2833 Divisional Different Low Small Long .2560 Divisional Different Low Large Short .2449 Divisional Different Low Larg e Long .2299 Divisional Different High Small Short .1067 Divisional Different High Small Long .1067 Divisional Different High Large Short .0877 Divisional Different High Large Long .0903

PAGE 88

88 Table 4 1. Continued. Team Task Structure Team Composition in Learning Ability Leader Reactivity Disturbance Magnitude Disturbance Duration SD of Time Spent Disjunctive Same Low Small Short .0000 Disjunctive Same Low Small Long .0000 Disjunctive Same Low Large Short .0000 Disjunctive Same Low Large Long .0000 D isjunctive Same High Small Short .0000 Disjunctive Same High Small Long .0000 Disjunctive Same High Large Short .0000 Disjunctive Same High Large Long .0000 Disjunctive Different Low Small Short .3448 Disjunctive Different Low Small Long .3448 Disjun ctive Different Low Large Short .3448 Disjunctive Different Low Large Long .3448 Disjunctive Different High Small Short .4167 Disjunctive Different High Small Long .4167 Disjunctive Different High Large Short .4167 Disjunctive Different High Large Lon g .4167

PAGE 89

89 Table 4 2. Results from sensitive analyses. Parameter Default value Values tested Effect Individual competency bias 0 .01, .01 Negatively related to individual learning Positively related to developmental leadership behavior Initial competenc y .00005 .00001, .00009 Negatively related to individual learning and development leadership behavior Individual task state bias 0 .01, .01 Positively related to individual expectancy Positively related to task contingent leadership behavior Utilities o f off task activities 0 .10, .20 Negatively related to individual task state Positively related to task contingent leadership behavior Individual goal bias 0 .01, .01 Positively related to individual task state Negatively related to task contingent leade rship behavior Leader bias about individual competency 0 .01, .01 Negatively related to task contingent leadership behavior

PAGE 90

90 Figure 4 1. Task contingent team leadership behavior and team task state over time by disturbance magnitude conditions.

PAGE 91

91 Fi gure 4 2. Task contingent team leadership behavior and team task state over time by disturbance duration conditions.

PAGE 92

92 Figure 4 3. Task contingent team leadership behavior, team expectancy, and team task state over time by leader reactivity conditions.

PAGE 93

93 A Figure 4 4. Task contingent team leadership behavior and team task state over time by leader reactivity and team task structure conditions. A) t eam task structure = functional . B) t eam task structure = divisional . C) team task structure = d isjunctive .

PAGE 94

94 B Figure 4 4. Continued.

PAGE 95

95 C Figure 4 4. Continued.

PAGE 96

96 A Figure 4 5. SD of individual expectancy and team task state over time by team composition in learning ability and team task structure conditions. A) t eam task structure = functional . B) t eam task structure = divisional . C) team task structure = disjunctive .

PAGE 97

97 B Figure 4 5. Continued.

PAGE 98

98 C Figure 4 5. Continued .

PAGE 99

99 CHAPTER 5 STUDY 2: HYPOTHESES TESTING USING A LAB EXPERIMENT A number of hypotheses could be derived from the current computation al model. One important issue that the model could speak to is how leaders allocate time to regulatory actions toward different subordinates as well as the team as a whole over the course of a team performance episode . In other words, what influences over time. This is an important research question for both theoretical and practical reasons. Theoretically, m oving from traditional leadership research to the team context, it is critical for team leadership theories to articulate how leaders treat multiple individuals in the same unit (Kozlowski & Klein, 2000). One fundamental question is to whom the leader dev otes time to different regulatory targets . Understanding the mechanisms that regulation target can help advance our understand ing about the regulatory process in the team context and the multilevel nature of leadership phenomenon. From a practic e standpoint, understanding how leaders choose regulatory targets over time can help improve decision making strategies and skills, particularly in dynamic task environment . Different theoretical perspectives have been taken to und erstand who is regulated by team leaders . Research taking the classic leadership approaches assumes that leaders treat the team as a whole, regardless of the within team differences in position or person characteristics (e.g., Schaubroeck et al., 2007 ). Th e observed within team/within leader differences are treated as random variances in the conceptual and empirical analyses. Some other researchers acknowledge that leaders might treat different subordinates differently and argue that differentiation among s ubordinates in individual focused leadership can be detrimental to team performance ( e.g., Wu et al., 2010). However, e xisting research has not explicated what brought about the differences in leaders treatment of subordinates. Specifically, it is not cle ar that (a) at a given time point, what

PAGE 100

100 factors ; and (b) over the course of a team performance episode, which factors lead to differentiation in time of the leader allocated to different subordinates. Drawing on f unctional leadership theory and self regulation theory, the current computational model provides predictions to both questions. First, t h e current model specifies regulation target is decided by the relat ive discrepancy among subordinates, with the specific form of the relationship contingent on team task structure (mathematical equation used to define regulation target ijkt is listed under Figure 3 3) . The model specifies this relationship based on the sel f regulation theory (Schmidt & DeShon, 2007; Vancouver, Weinhart et al., 2010) . Specifically, self regulation theory argues that when individuals work on multiple tasks simultaneously, the choice of regulation target is decided by comparing the weighted di screpancies of the alternative tasks . The weighting factor is a function of the importance of the task. at a given time point should be decided by comparing weigh ted discrepancies among multiple subordinates. The weighting factor is a function of team task structure . When team task is functional, subordinates with the largest discrepancy receives the largest weight. When team task is divisional, the weights are equ al across the subordinates. When team task is disjunctive, subordinates with the smallest discrepancy receives the largest weight. Therefore, with all else being equal, team task structure moderates the relationship between relative discrepancy and the lik elihood of being regulated by the leader. Hypothesis 1: W ith all else being equal among the subordinates, the positive relationship between relative discrepancy of a subordinate comparing to other team members and the time

PAGE 101

101 spent by the leader on regulatin g the subordinate is stronger when task structure is functional (vs. divisional and disjunctive). T he current model specifies that the differentiation in task difficulty across positions and the differentiati on among team members in learning ability are p ositively related to the differentiation in leader time allocation across subordinates. Specifically, w hen team members mainly driven by external disturban ces and team task structure. When team positions differ in task difficulty, positions that are responsible for more difficult task s are more likely to have larger discrepancy relative to less difficult positions and thus receive more time from the leaders. Similarly, when team members differ in learning ability, members with relatively lower learning ability are more likely to have larger discrepancy than members with higher learning ability and thus receive more time from the leaders to help address such d iscrepancy . Given that differentiation in task difficulties and learning abilities are task level input factors, they can systematically influence the differentiation in leader time allocation over the entire team performance episode. Hypothesis 2: When t ask difficulty is different (vs. same) among the positions, (a) more time of the leader is allocated to more difficult position , and (b) the differentiation in leader time allocation is larger. Hypothesis 3: When learning ability is different (vs. same) am ong the subordinates, (a) more time of the leader is allocated to the subordinate with lower learning ability , and (b) the differentiation in leader time allocation is larger. The current model also specifies that the magnitude of external disturbance can change the strength of the relationship s between the differentiation in task difficulty and member

PAGE 102

102 learning ability , and the differentiation in leader time allocation. When disturbance magnitude is large (vs. small), the impact of task difficulty and learn relatively smaller as compared to external unexpected events . Larger external disturbance could diminish the differences among subordinates in task states to a greater extent. W hen disturbance magnitude is large, it is more difficult fo r team leaders to differentiate subordinates in di fferent positions or with different characteristics because changes in task states are more heavily triggered by external forces . Therefore, given the same amount of differentiation in task difficulty or member learning ability, it is less likely for team leaders to differentiate among subordinates when disturbance magnitude is large (vs. small). Hypothesis 4: When disturbance magnitude is small (vs. large), (a) the positive relation ship between task difficulty of a position and time spent on this position by the leader is stronger , and (b) the positive relationship between differentiation in task difficulty and differentiation in leader time allocation is stronger. Hypothesis 5: When disturbance magnitude is small (vs. large), (a) the negative relationship between learning ability of a subordinate and time spent on this subordinate by the leader is stronger , and (b) the positive relationship between differentiation in learning ability and differentiation in leader time allocation is stronger. The current model specifies that team task structure would moderate the relationship between differentiation in task and person characteristics and differentiation in leader time allocation. When team task structure is functional (vs. divisional and disjunctive), regulating the subordinate with more difficult task contributes more significantly to increasing the overall performance of the team. Accordingly, there should be a stronger relationship b etween differentiation in task difficulty and differentiation in leader time allocation when team task

PAGE 103

103 structure is functional (vs. divisional and disjunctive). As for team member differentiation in learning ability, regulating the subordinate with lowest learning ability contributes more significantly to increasing the overall performance of the team when team task structure is functional (vs. divisional and disjunctive). Accordingly, there should be a stronger relationship between differentiation in learn ing ability and differentiation in leader time allocation when team task structure is functional (vs. divisional and disjunctive). Hypothesis 6: When task structure is functional (vs. divisional and disjunctive), (a) the positive relationship between task difficulty of a position and time spent on this position by the leader is stronger , and (b) the positive relationship between differentiation in task difficulty and differentiation in leader time allocation is stronger. Hypothesis 7: When task structure i s functional (vs. divisional and disjunctive), (a) the negative relationship between learning ability of a subordinate and time spent on this subordinate by the leader is stronger , and (b) the positive relationship between differentiation in learning abili ty and differentiation in leader time allocation is stronger. Method Participants A total of 86 students in a southeast university voluntarily participated in the study. Participants were either rewarded twelve dollars in cash or extra course credit. The average age of the participants was 22.16 years old ( SD = 4.02). 61% ( n = 52) of the participants were female. 55% ( n = 45) of the participants were Caucasian. Task A simulation platform was developed specifically for this study. In this simulation envir onment , the participants played the role of a head chef (i.e., team leader) who managed a team composed by four simulated subordinates working in the kitchen (a sous chef, a pastry

PAGE 104

104 chef, a preparation chef, and a wine steward). At the beginning of each tea m performance episode , participants would be informed of the specific needs of the guests attending a particular food event (e.g., catering a family event, hosting a food tasting festival , etc. objective was to make sure the team as a w hole delivered the dishes ordered by the guests. In order to do so, the participants needed to regulate the competency level and task state of the time in the simulation environment ) on vario us actions built into the task interface. Participants were allowed to allocate time to two general types of actions: team focused and individual focused regulating actions. When the same amount of time was allocated to a team focused action and an individ ual focused action, the team focused action was less effective on an individual subordinate than the individual focused action. Team focused actions had four choices: team focused giving instruction, demonstration, sense making, and reviewing past events. Individual focused actions had four choices as well: individual focused giving instruction, demonstration, sense making, and reviewing past events. In the current study, these four choices did not differ from each other in terms of the effects on subordina tes. In the current study, participants were required to allocate a total of 600 minutes into 120 five minute actions, i.e., to make 120 choices, during each team performance episode . Finishing making the choices for each team performance episode took about 10 15 minutes actual time. Throughout the team performance episode , participants were presented with information competency level (i.e., percentage of overall task state (i.e., percentage of team task accomplished). This information was refreshed each time the participants allocated time into an action. Participants were also informed that

PAGE 105

105 random e vents could disrupt the progress of the individual subordinates and/or the team, which resulted in individual or team task state being set back. In addition, throughout the task, view subordinat characteristics (e.g., learning ability, position responsibilities , etc. ). Record most recent five choices was displayed on the interface throughout the team performance episode . All interactions between the participants and the interface (e.g., which subordinate was being regulated at a given time point) were recorded throughout the study . The simulation started with zero percent of individual and team task accomplished. Throughout the team performance episode , particip ants were reminded of the objective, i.e., to increase the team task state to deliver 100% of the dishes ordered by the customers. Participants could proceed to the next team performance episode when the team accomplished the objective. Negative disturbanc e appeared periodically. S creenshot of the study interface is included in Appendix B. Procedure The present study examined the main effects and joint effects of four independent variables: team task structure, differentiation in learning ability across tea m members , differentiation in task difficulty across positions, and magnitude of external disturbance. A between participant (team task structure: functional, divisional, disjunctive; differentiation in learning ability: same vs. different ) and within part icipant ( differentiation in task difficulty: same vs. different ; disturbance magnitude: small vs. large) mixed design was used in this study. Each study session included up to six participants, with about two participants in a typical study session. Partic ipants were randomly assigned into one of the six between participant conditions. Upon arriving at the lab, participants were introduced to the study and assigned to individual computers which had the study platform set up. Partic ipants first completed a s urvey

PAGE 106

106 about demographic information and prior knowledge about the kitchen management scenario used in the simulation task. After completing the survey, participants watched a seven minute video which included the tutorial of the task (introducing objective of the task and operations of the task interface) and manipulations of team task structure and team composition in learning ability. Following the tutorial, participants practice d over the course of one team performance episode . Participants then performe d over four team performance episode s. In each team performance episode , participants first read about a specific task assigned to the kitchen (manipulation of differentiation in task difficulty embedded). Participants then made choices about time allocati on over the course of a team performance episode . Manipulations of within participant factors were presented at a random order for each participant. At the end of the study, participants were debriefed and compensated. Manipulations Team task structure man ipulation. Team task structure was manipulated through both descriptions of the basic information of the team and algorithms programmed into the platform. Descriptions of the teams were embedded in the tutorial. The full text of descriptions used in the th ree conditions is presented in Appendix B. In the functional condition, team task structure was performer will Correspondingly, i n the simulation, team task state was programmed as a function of the lowest individual subordi random factor . In the divisional structure condition, collective performance of the kitchen is determined by the average (emphasized) performance of Correspondingly, i n the simulation, team task state was

PAGE 107

107 factor . In the disjunctive perf ormance of the kitchen is determined by the best (emphasized) performer among the four Correspondingly, i n the simulation, team task state was programmed as a funct ion of the highest individual task state and a random factor. Differentiation in learning ability manipulation. Differentiation in learning ability across team members was manipulated through both descriptions of the basic information of the team members and algorithms programmed into the platform. Descriptions of the individual team members were embedded in the tutorial. The full text of descriptions used in the two conditions (same vs. different ) is presented in Appendix B. In the same learning ability condition, team members all had the same level of education. Correspondingly, i n the simulation, the competency level of all four subordinates increased at the same speed. In the different learning ability condition, one team member had a middle school dip loma, two had high school diploma s , and one had an associate degree. Correspondingly, i n the simulation, the competency level of the four subordinates increased at different speeds, with the subordinate with middle school diploma learned the slowest and th e subordinate with associate degree learned the fastest. D ifferentiation in task difficulty manipulation. Differentiation in task difficulties across positions was manipulated through both descriptions of the task presented at the beginning of each team p erformance episode and algorithms programmed into the platform. The full text of descriptions used in each team performance episode is presented in Appendix B. In the same task difficulty Correspondingly, i n the

PAGE 108

108 programmed to be the same. In the different task difficulty condition, one of the fo Correspondingly, i n the simulation, the task difficulty level of one position was programmed as higher than the other three positions, which r esulted in slower task progress at the same competency level. Disturbance magnitude manipulation. During each team performance episode , state. In the small magni tude condition, each negative disturbance set back the task state at the given time point by 20%. In the large magnitude condition, each negative disturbance set back the task state at the given time point by 40%. Pilot testing suggested that these magnitu de levels were noticeable by the participants. Timing, target, and magnitude of the disturbance are presented in Table 5 1. Measures Following previous research on goal related regulatory actions over time (e.g., Schmidt & DeShon, 2007), repeated measurem ents over the course of a team performance episode were divided into five segments. Each segment included data from 24 choices made by the participants. Following previous research on choices between competing targets (Schmidt & DeShon, 2007), one subordin ate was chosen as the referent subordinate in the analyses. T he subordinate with the highest task difficulty level in the diff erent task difficulty condition was assigned as the referent/focal subordinate. Relative discrepancy. As all participants had the same goal (i.e., to deliver 100% of the dishe s ordered by the guests), relative discrepancy was calculated as the difference between a

PAGE 109

109 focal subordinate and average task states of the other three subordinates ( i.e., [non referent subordinate 1 task state + non referent subordinate 2 task state + non referent subordinate 3 task state ]/3 referent subordinate task state). Higher score indicated larger relative discrepancy of the focal subordinate as compared to the other three subordinates. Con sidering the limitations of difference score approach (Edwards, 1994), I also conducted analyses using the analyses results are also reported . Leader time allocation. At the within episode level, time allocated to a subordinate was coded as the proportion of time that a focal subordinate received in each segment (time allocated to the focal subordinate divided by the total time available in each segment). At the team performance episode level, leader time allocation was coded as the proportion of time a focal subordinate received in each team performance episode (time allocated to the focal subordinate divided by the total time available in each team performa nce episode ) and the SD of the time received by the four subordinates. Considering that certain amount of time could be allocated to the team focused actions, I also coded leader time allocation by excluding time allocated to the team focused actions. Anal yses results from using this alternative coding scheme are included as well . Covariates. Gender, age, and previous knowledge about kitchen management were included as covariates when te sting H ypothesis 1. Previous knowledge about kitchen management w as measured by able do you think you are about the following jobs: executive chef, sous chef, pastry chef, preparation chef, and wine point scale ranging from 1 (not at all

PAGE 110

110 knowledgeable) to 5 (extremely knowledgeable). The average score from these five items was used to indicate previous knowledge about kitchen management. Analytic Strategy Data collected in the present study had a nested structure. Specifically, obser vations of time allocation in each segment were nested within team performance episodes . Team performance episode level data were nested within individuals whom were assigned to different between person conditions. Considering this structure of the present data, Hypothesis 1 was tested using multilevel modeling with multiple group comparison. Within each condition , t he multilevel model was specified as: Level 1: Y ij 0j 1 j ( T ij ) 2 j (X ij ) + r ij Level 2: 0j 00 01 (W1 j ) 0 2 (W2 j 0 3 (W3 j ) + u 0 j 1 j 1 0 + u 1j 2 j 2 0 + u 2j where: variance of r ij 2 = Level 1 residual variance variance of u 0j 00 = random intercept residual variance variance of u 1j = 11 = random slope of T ij residual vari ance variance of u 2j 22 = random slope of X ij residual variance time spent on the focal subordinate relative discrepancy , gender , representing age , and W3 presenting prior knowledge. Hypotheses 2 7 were tested with repeated measure multivariate analysis of variance ( M ANOVA) with time allocation at the team performance episode level as the outcome variables. The general linear model was specified as:

PAGE 111

111 Y ij = + A + B + C + D + E + AB + AC + AD + AE + BC + BD + BE + CD + CE + DE + ABC + ABD + ABE + ACD + ACE + ADE + BCD + BCE + BDE + CDE + ABCD + ABCE + ABDE + ACDE + BCDE + ABCDE + ij where: = overall mean across all conditions A, B, C, D, E = fac tor main effects AB, AC, AD, AE, BC, BD, BE, CD, CE, DE = two way interaction effects ABC, ABD, ABE, ACD, ACE, ADE, BCD, BCE, BDE, CDE = three way interaction effects ABCD, ABCE, ABDE, ACDE, BCDE = four way interaction effects ABCDE = five way interaction effects ij = error time allocation A team task structure B differentiation in learning ability , differentiation in task difficulty representing disturbance magnitude , E presenting time allocation measurement . Results Relationships between Discrepancy and Leader Time Allocation Hypothesis 1 stated that with all else being equal among the subordinates, there was a e discrepancy to others and time of the leader allocated to this focal subordinate, and this positive relationship was stronger when task structure was functional (vs. divisional and disjunctive). To test Hypothesis 1, I conducted a multiple group multilev el modeling using a subset of the data that included equivalent task model, I specified the within episode level fixed effect of discrepancy (either as differe nce score or as difference score components) on time allocation to the focal subordinate. Gender, age, and

PAGE 112

112 previous knowledge about kitchen management were included as episode level control variables. Time (coded as measurement wave ranging from 1 to 5) wa s included as a within episode level control variable. In order to facilitate interpreting the results, episode level control variables were grand mean centered and within episode level predictors were group mean centered. Multilevel modeling results for each condition are reported in Table 5 2. Multiple group analyses results suggested that the effect of relative discrepancy (i.e., difference score) on time allocated to the focal subordinate differed significantly across team task structure conditions ( 2 s[2] = 13.88 and 8.24, p s < .05, for small and large disturbance conditions , respectively). There was a stronger positive relationship between relative discrepancy and time allocated to the focal subordinate when team task structure was functional ( s = 1.15 and 1.40, p s < .01, for small and large disturbance conditions , respectively) than when team task structure was divisional ( s = .36 and .45, p s > .05, for small and large disturbance conditions , respectively) and disjunctive ( = 1.24, p < .01 and = .50, p > .05, for small and large disturbance conditions , respectively). Multiple (i.e., difference score component) on time allocated to the focal subordinate d iffered significantly across team task structure conditions ( 2 s[8] = 22.33 and 26.48, p s < .01, for small and large disturbance conditions , respectively). There was a stronger negative relationship between focal time allocated to the focal subordinate when team task structure was functi onal ( s = 3.69 and 4.42, p s < .01, for small and large disturbance conditions , respectively) than when team task structure was divisional ( = 1.03, p > .05 and = 1.98, p < .05, for small and large disturbance conditions , respectively) and disjuncti ve ( = 3.66, p < .01 and = 1.43, p > .05, for small and large disturbance conditions , respectively). Therefore, Hypothesis 1 was supported.

PAGE 113

113 Effects of Manipulated Factors on Leader Time Allocation Means and standard deviations of the time allocated t o the focal subordinate in each team performance episode and the differentiation in leader time allocation across subordinates in each team performance episode are reported in Table 5 4. A 3 (team task st ructure: functional, divisional, and dis junctive) by 2 (differentiation in learning ability: same and different) by 2 (differentiat ion in task difficulty: same and different) by 2 ( disturbance magnitude: small and large) mixed M ANOVA with repeated measures on the last two factors was conducted on the time a llocation measures (time allocated to the focal subordinate and differentiation in leader time allocation across subordinates) . Multivariate tests r evealed significant main effect of differentiation in task difficulties ( = .41, F [1, 80] = 116.24, p < .01). It also revealed significan t two way interactions between differentiation in learning ability and disturbance magnitude = .91, F [1, 80] = 7.68, p < .01), and between differentiation in task = .87, F [1, 80] = 12.50, p < .01). The mixed MANOVA also revealed significant three way interaction among differentiation in learning = .95, F [1, 80] = 4.28, p < .05). Th ese results warranted further univariate analyses. Further univariate analyses results suggested that the main effect of task difficulty differentiation on time allocation to the focal subordinate was significant ( F [1, 80] = 93.36, MSE = .01, p < .01, 2 = .54). More time was allocated to the focal subordinate (i.e., the one with highest task difficulty) when task difficulty were different across positions ( Mean = .26, SE = .01) than when task difficulty was the same for all positions ( Mean = .18, SE = .01) . Therefore, Hypothesis 2a was supported. The main effect of learning ability differentiation on time allocation to the focal subordinate was significant ( F [1, 80] = 5.44, MSE = .02, p < .05, 2 = .06). More time was allocated to the focal subordinate (i.e., the one with lowest learning ability) when

PAGE 114

114 learning ability were different across team members ( Mean = .23, SE = .01) than when learning ability was the same for all team members ( Mean = .2 0, SE = .01). Therefore, Hypothesis 3a was supported. The interaction effect of task difficulty differentiation and disturbance magnitude on time allocat ed to the focal subordinate was significant ( F [1, 80] = 21.36, MSE = .01, p < .01, 2 = .21). However, the differenc es in time allocated to the focal subordinate between task difficulty differentiation conditions ( different vs. same) was larger when disturbance magnitude was large ( mean difference = .12, SE = .01) than when disturbance magnitude was small ( mean difference = .04, SE = .01). Therefore, Hypothesis 4a was not supported. The interaction effect of team composition in learning ability and disturbance magnitude on time allocation to the focal subordinate was significant ( F [1, 80] = 7.40, MSE = .01, p < .01, 2 = .09). The difference in time allocated to the focal subordinate between differentiation in learning ability conditions ( different vs. same) was larger when disturbance magnitude was small ( mean difference = .06, SE = .01) than when disturban ce magnitude was large ( mean difference = .003, SE = .01). Therefore, Hypothesis 5a was supported. The interaction effects between task difficulty differentiation and task structure , and between learning ability differentiation and task structure on time a llocation to the focal subordinate were not significant. Therefore, Hypotheses 6a and 7a were not supported. Results suggested that the three way interaction among task difficulty differentiation, learning ability differentiation , and disturbance magnitude was significant ( F [1, 80] = 7.70, MSE = .01, p < .01, 2 different and disturbance magnitude was large, the difference in time allocated to the focal subordinate between task difficulty differentiation co nditions ( different vs. same) was largest ( mean difference = .14, SE = .02).

PAGE 115

115 Results also suggested that the other three way interactions and the four way interaction were not significant. R esults suggested that the main effect of differentiation in task difficulty on differentiation in leader time allocation across subordinate was significant ( F [1, 80] = 102.19, MSE = .002, p < .01, 2 = .56). There was larger differentiation in leader time allocation when task difficulty were different across positions ( Mean = .09, SE = .004) than when task difficulty was the same for all positions ( Mean = .04, SE = .002). Therefore, Hypothesis 2b wa s supported. The main effect of learning ability differentiation on differentiation in leader time allocation was also significant ( F [1, 80] = 21.89, MSE = .002, p < .01, 2 = .22). There was more differentiation in leader time allocation when learning abi lity were different across subordinates ( Mean = .08, SE = .004) than when subordinates had the same learning ability ( Mean = .06, SE = .004). Therefore, Hypothesis 3b was supported. The interaction effect of task difficulty differentiation and disturbance magnitude on differentiation in leader time allocation was not significant. Therefore, Hypothesis 4b was not supported. The interaction effect between learning ability differentiation and disturbance magnitude on differentiation in leader time allocation was significant ( F [1, 80] = 4.64, MSE = .001, p < .05, 2 = .06). The difference in differentiation in leader time allocation between learning ability differentiation conditions ( different vs. same) was larger when disturbance magnitude was small ( mean dif ference = .03, SE = .01) than when disturbance magnitude was large ( mean difference = .02, SE = .01). Therefore, Hypothesis 5b was supported. The interaction between task difficulty differentiation and task structure on differentiation in leader time allo cation was not significant. Therefore, Hypothesis 6b was not supported. The interaction effect between learning ability differentiation and task structure on differentiation in

PAGE 116

116 leader time allocation was marginally significant ( F [2, 80] = 2.58, MSE = .002, p = .08, 2 = .06). The difference in differentiation in leader time allocation between learning ability differentiation conditions ( different vs. same) was larger when task structure was functional ( mean difference = .04, SE = .01) than when task structu re was divisional ( mean difference = .01, SE = .01) and disjunctive ( mean difference = .02, SE = .01). Therefore, Hypothesis 7b was supported. Results also suggested that the three way interactions and the four way interaction were not significant. Supple mentary Analyses Alternative coding scheme of leader time allocation. I repeated the above analyses by using an alternative coding scheme of leader time allocation, i.e., excluding time allocated to team focused actions. Multiple group multilevel modeling results suggested that for small disturbance, relative discrepancy was positively related to time allocated to the focal subordinate when team task structure was functional ( = 1.84, p < .01) and disjunctive ( = 2.42, p < .01) but not when team task str ucture was divisional ( = .76, p > .05). For large disturbance, relative discrepancy was positively related to time allocated to the focal subordinate when team task structure was functional ( = 1.58, p < .01), but not when team task structure was divisi onal ( = .65, p > .01) and disjunctive ( = .79, p > .05). The effects differed significantly across team task structure conditions when disturbance magnitude was large ( 2 [2] = 6.71, p < .05). This finding supported Hypothesis 1. Multilevel modeling res ults for the relationship between relative discrepancy and time allocated to the focal subordinate for each condition are reported in Table 5 3. A mixed MANOVA was conducted on time allocation measures using the alternative coding scheme. M ultivariate tes ts revealed significant main effect of differentiation in task = .41, F [1, 80] = 113.28 , p < .01). It also revealed significant two way

PAGE 117

117 = . 94 , F [1, 80] = 4.97 , p < .05 ), and between differentiation in task difficulties and d isturbance = . 92 , F [1, 80] = 6.71 , p < .05 ). The mixed MANOVA also revealed significant three way interaction among differentiation in learning ability, differentiation in task = .95, F [ 1, 80] = 4. 17 , p < .05). These results warranted further univariate analyses. R esults suggested that the main effect of task difficulty differentiation on time allocated to the focal subordinate was significant ( F [1, 79] = 94.99, MSE = .01, p < .01, 2 = . 55). More time was allocated to the focal subordinate (i.e., the one with highest task difficulty) when task difficulty were different across positions ( Mean = .40, SE = .01) than when task difficulty was the same for all positions ( Mean = .27, SE = .01). The main effect of learning ability differentiation on time allocated to the focal subordinate was not significant. The interaction effect of task difficulty differentiation and disturbance magnitude on time allocation to the focal subordinate was signific ant ( F [1, 79] = 13.53, MSE = .02, p < .01, 2 = .15). The difference in time allocated to the focal subordinate between task difficulty differentiation conditions ( different vs. same) was larger when disturbance magnitude was large ( mean difference = .18, SE = .02) than when disturbance magnitude was small ( mean difference = .08, SE = .02). The interaction effect between learning ability differentiation and disturbance magnitude on time allocated to the focal subordinate was significant ( F [1, 79] = 5.96, MS E = .03, p < .05, 2 = .07). The difference in time allocated to the focal subordinate between learning ability differentiation conditions ( different vs. same) was larger when disturbance magnitude was small ( mean difference = .06, SE = .02) than when disturbance magnitude was large ( mean difference = .03, SE = .02). The interaction effect between task difficulty differentiation and team task structure on time allocated to the focal subordinate was not significant. The interaction effect between

PAGE 118

118 learning ability differentia tion and team task structure on time allocated to the focal subordinate was significant ( F [2, 79] = 5.04, MSE = .01, p < .01, 2 = .11). The difference in time allocated to the focal subordinate between learning ability differentiation conditions ( different vs. same) was larger when team task structure was functional ( mean difference = .07, SE = .02) than when team task structure was divisional ( mean difference = .02, SE = .02) and disjunctive ( mean difference = .01, SE = .02). Therefore, Hypotheses 2a, 5a, and 7a were supported, but not Hypotheses 3a, 4a, and 6a. Results suggested that the three way interaction among learning a bility differentiation , disturbance magnitude, and team task structure was significant ( F [2, 79] = 3.25, MSE = .03, p < .05, 2 = .08). When team task structure was functional and disturbance magnitude was small, the difference in time allocated to the focal subordinate between team learning ability differentiation conditions ( different vs. same) was largest ( mean difference = . 17, SE = .03). Results suggested that the three way interaction among task difficulty differentiation, learning ability differentiation , and disturbance magnitude was significant ( F [1, 79] = 7.65, MSE = .02, p < .01, 2 abilities were different and disturbance magnitude was large, the difference in time allocated to the focal subordinate between task difficulty differentiation conditions ( different vs. same) was largest ( mean difference = .20, SE = .02). Results also sugg ested that the other three way interactions and the four way interaction were not significant. R esults suggested that the m ain effect of differentiation in task difficulty on differentiation in time allocation across subordinate was significant ( F [1, 79] = 88.95, MSE = .01, p < .01, 2 = .53). There was larger differentiation in time allocation when task difficulty were different across positions ( Mean = .15, SE = .008) than when task difficulty was the same

PAGE 119

119 for all positions ( Mean = .07, SE = .004). The m ain effect of learning ability differentiation on differentiation in time allocation was marginally significant ( F [1, 79] = 3.23, MSE = .01, p = .08, 2 = .04). There was larger differentiation in time allocation when learning ability were different across subordinates ( Mean = .12, SE = .01) than when learning ability was the same for all subordinates ( Mean = .10, SE = .01). The interaction effects between task difficulty differentiation and disturbance magnitude, between learning ability differentiation an d disturbance magnitude, and between task difficulty differentiation and team task structure on differentiation in time allocation were not significant. The interaction effect between learning ability differentiation and team task structure on differentiat ion in time allocation was significant ( F [2, 79] = 5.34, MSE = .02, p < .01, 2 = .12). The difference in differentiation in time allocation between learning ability differentiation conditions ( different vs. same) was larger when team task structure was fu nctional ( mean difference = .07, SE = .01) than when team task structure was divisional ( mean difference = .01, SE = .01) and disjunctive ( mean difference = .00, SE = .01). Therefore, Hypotheses 2b, 3b, and 7b were supported, but not Hypotheses 4b, 5b, an d 6b. Results also suggested that the three way interactions and the four way interaction were not significant. Interaction between time and relative discrepancy. I tested whether the relationship between relative discrepancy and time allocated to the fo cal subordinate changed over time. The interaction term between time and relative discrepancy was included in the multilevel model. Multiple group analyses results suggested that the moderation effect of time on the relationship between relative discrepanc y and time allocated to the focal subordinate differed significantly across team task structure conditions when disturbance magnitude was large ( 2 [6] = 15.57, p < .05). Under large disturbance, the positive relationship between relative discrepancy and time

PAGE 120

120 allocated to the focal subordinate weakened over time when team task structure was disjunctive ( = .38, p < .01) but not when team task stru cture was divisional ( = .09, p > .05) and functional ( = .01, p > .05). Quantitative Model Evaluation I evaluated the model data fit of the theoretical model (i.e., the computational model specified in Study 1; M 0 ) and compared the theoretical model t o two alternative models: a more parsimonious model that eliminated team expectancy components from the theoretical model (M 1 ) and a less parsimonious model that added individual expectancy components to the theoretical model (M 2 ). These two alternative mo dels were chosen because previous research choices made during regulating multiple tasks (Vancouver , Weinhardt et al., 2010). Model data fit of leader time allocation in each team performance episode . Expected values of time spent on the focal subordinate and differentiation ( SD ) of time spent on subordinate s that were derived from the theoretical models are reported in Table 5 4 (including team focused actions) and Ta ble 5 5 (excluding team focused actions). Expected values were derived by using baseline values of the input factors, which were calibrated in a pilot study. Model data fit was evaluated by comparing estimates of the outcome variables from the participants As for the amount of time spent on focal subordinate, the expected values and the s of the estimated marginal means) overlapped in 8 out of the 24 condit amounts of time spent on the referent subordinate when task difficulty was the same across positions and disturbance magnitude was large. When task difficulty was the same acros s positions and disturbance magnitude was small or when task difficulty was different across

PAGE 121

121 spent on the referent subordinate. The expected values and the e data overlapped in 6 and 8 out of the 24 conditions for the more parsimonious and less parsimonious models , respectively. As for the SD of time spent on the subordinates, the expected values and the estimates from the partic overlapped in 6 out of the 24 conditions. The estimates from the SD of time spent on the subordinates when task difficulty was different across positions and disturbance magnitude was small. Whe n task difficulty was different across positions and disturbance magnitude was large or when task than the expected SD of time spent on the subordinates. The e xpected values and the estimates overlapped in 5 and 3 out of the 24 conditions for the more parsimonious and less parsimonious models , respectively. Model data fit of choices over time. I also examined how the theoretical mode l and alternative models fit choices made by participants over time. Optimization was conducted to obtain values of overall model data fit index and estimates of the parameters that would minimize the differences between the modeled and observed time serie s data (i.e., parameter calibration; Oliva, 2003). Considering that the choices over time were categorical variables at each time point, model data fit was indicated by th e percentage of data points with accurate prediction (Vancouver , Weinhardt et al. , 20 10). In the optimization procedure, input values of relative discrepancy were calibrated. All other parameters were set at the default values. In order to obtain estimates of the parameters in a realistic range, lower and upper bounds of the

PAGE 122

122 parameters were pre constrained to between 0.5 and 0.5 . Leader reactivity and leader sensitivity to relative discrepancy were constrained to between 0 and 1 . Optimization was conducted for individual data fit values were aggregated across participants. Optimization was conduc ted using software Vensim Professional. As shown in Table 5 6, model data fit of the theoretical model varied, ranging from 19.56% to 51.44% accurate prediction (average = 36.19 %). These results suggested that the theoretical model could explain the parti under certain combinations. Parameter estimates suggested that leader biases of individual participants and between team performance episodes (i.e., within participants). To illustrate the differences of the predictions before and after calibration, I plotted the expected choices with baseline values of the parameters (i.e., leader biases of individual task state, leader reactivity, and leader sensitivity to relative discrepancy), expected choices with calibrated values of the parameters, and observed choices over time for one participant ( Figure 5 1 ) . Compared to the expected choices with baseline values of the parameters, the observed choices included fewer team focused actions in the first half the team performance episode , which suggested that the participant was more sensitive to relative discrepancy than the baseline level. After calibrating the paramet ers in the model, the expected choices over time were more similar to the observed choices over time. In addition, estimation results suggested that the model data fit of the more parsimonious model (M 1 ; 33.25 % ) and the less parsimonious model (M 2 ; 35.38 % ) was also smaller than the theoretical model.

PAGE 123

123 Discussion The present study tested a series of hypotheses derived from the proposed theoretical model. Data collected from a lab experiment supported the hypothesis that relative discrepancy was positively related to leader time allocation, the relationship of which was stronger when team task structure was functional (vs. divisional and disjunctive). This study also found that when task difficulties was different (vs. same) across the positions, more time o f the leader was allocated to the more difficult position and differentiation in leader time allocation among subordinates was larger. When team members had different (vs. same) learning abilities, more time was allocated to the subordinate with lower lear ning ability and differentiation in time allocation among subordinates was larger. In addition, the negative relationship between learning ability of a position and time allocated to regulating by the leader is stronger when disturbance magnitude was small er. The positive relationship between differentiation in learning ability and differentiation in leader time allocation is stronger when disturbance magnitude was smaller or when team task structure was functional (vs. divisional and disjunctive). Theoret ical Implications These findings from the present study have several theoretical implications. First, the finding supported the idea that the amount of time allocated from team leaders to subordinates current task state and ideal task state (i.e., team goal), and the strength of the relationship was contingent on the team task structure. This finding provides support to the theoretical model that draws on the discrepancy reduction argument of self regu lation theory a nd the task interdependence idea of coupling lower level performance systems to the higher level. The process of leader dynamically choosing regulation target according to relative discrepancy and team task structure is the fundamental mecha nism through which other relationships were derived from model.

PAGE 124

124 In addition, the supplementary analyses found that time significantly changed the strength of the relationship between relative discrepancy and leader time allocation when team task structur e was disjunctive, such that the positive relationship was stronger at the beginning of the team performance episode than closer to the deadline. This finding is consistent with goal pursuit, which sugges ted that individuals tended to pursue the task with less discrepancy when time is closer to deadline (Schmidt & DeShon, 2007; Vancouver , Weinhardt , 2010). The current study further advances this argument by showing that the positive effect of discrepancy o n time allocation only weakens when the regulation targets (i.e., different subtasks, different subordinates) are not dependent on each other (i.e., low task interdependence or disjunctive task structure). Second, the study supported the hypothesis that the differentiation in task difficulty across positions in the team could lead to more time of the leader being allocated to the more difficult position and larger differentiation in leader time allocation across positions. This finding is consistent with previous research which suggested that position characteristics can influence resource allocation in the team (Humphrey et al., 2009). The current study extends the previous research by showing that the underlying mechanism for this relationship is that le adership serves as other regulatory agent to ensure the functioning of each position. With all else being equal, when a position is tasked with difficult responsibilities, there is likely to be a larger discrepancy and the ideal state. Therefore, it is more likely for resource s to flow from the leader to the difficult positions . Third, supporting the effect of team composition on team leadership, this study found that differentiation in ties was positively related to differentiation in leader time allocation. Team leaders also tended to allocate more time to the subordinate with

PAGE 125

125 lowest learning ability. This finding is consistent with previous research which suggested that team compositio n can influence resource allocation in the team (Henderson et al., 2009). The current study further showed that the relationship between team composition in learning ability and team leadership was influenced by magnitude of external negative disturbance a nd team task structure. This finding extended previous research on team leadership by demonstrating the joint effects of team composition and team design features. The current study suggested that differentiation in kely to be strong when team members are different in their abilities to acquire new knowledge and their jobs rely on each other. This finding also demonstrates the importance of considering team design factors when examining the effect of team composition on team processes. Furthermore, quantitative model evaluation results suggested that leader biases of had both within and between individual variations. These results showed that the computational model established in Study 1 could account for individual differences of team leaders, which helps integrate team leadership research on leader individual differences (e.g., Shin & Zhou, 2007) and research on le adership behavior as reactions to the task environment (e.g., Morgeson & DeRue, 2006). In addition, examining the within individual variation helps provide a dynamic team performance episode t o on task characteristics. Practical Implications leaders alloca ted more time to the subordinate in more difficult posi tions and the differentiation in leader time allocation across subordinates was larger. Although spending more time on

PAGE 126

126 regulating the subordinate in more difficult position helped the subordinate striv e toward the goal, the increased differentiation in leader time allocation might create unwanted effects (e.g., decreased perceived justice among other subordinates). Finding of this study suggested that if the organization wants to decrease the amount of differentiated leadership among subordinates, tasks assigned to positions in the same work groups or teams can be designed to be similar in terms of difficulty level. This finding also suggested that organizations could use team design to motivate employee s to take on more difficult tasks, assuming employees observe that those who work in more difficult positions receive more attention from the team leaders. This study also found that when team members differed in learning ability, team leaders allocated more time to the subordinates with lower learnin g ability and differentiation in time allocation across subordinates was larger, the effect of which was larger when task interdependence was high in the team. Although selection procedures could help reduce the differences among employees in job related knowledge, skills, and abilities, differences might still exist. Finding of the present study showed that one critical function of team leadership was to make sure that team as a whole, especially when team me mbers depend on each other closely, suggested that in order for organizations to reduce differentiation in subordinates, team members compo sed into the same group should be relatively similar in learning ability. This study also provides practical implications for training and developing team leaders. The study found that team leaders could be biased in their perceived task state of individ ual who to regulate .

PAGE 127

127 ng rewards and punishments. Therefore, these biases could lead to negative consequences for both individual subordinates and the organizations. It is possible that through training in simulation tasks like the platform used in iases could be identified and corrected. Limitations and Future Research Directions The current study provides an empirical test of the proposed dynamic team leadership model. Using an experimental design , the study helps clarify causal relationships betw een input factors to the team performance system and leader time allocation of team leaders over time. As a first step toward testing the dynamic model of team leadership, the study is limited in several ways. First, this study examined actual discrepancy between current and ideal task state rather than perceived discrepancy, although the theoretical model specified that perceived discrepancy mediated the effect of task state on leader regulatory actions. Quantitative model evaluation results suggested that some participants actual task state. However, previous research suggested that the effects of actual ta sk state and perceived task state on regulatory actions are similar (Schmidt & DeShon, 2007). The current model was also able to account for the biases through optimization procedure. Nevertheless, future research should directly examine the effect of perc regulatory actions. predictions of disturbances (Vancouver, 2008). In the lab study, the same disturbance schedule was used in order t o control the effects of scheduling when testing between episode differences. This protocol might have allowed some participants to expect when disturbances would happen and become more attuned to the changes of task states of individual subordinates, and take

PAGE 128

128 expected disturbance into consideration when making decisions . Future studies should examine this possibility by comparing conditions with and without a fixed schedule of disturbances. Third, the current study did not examine the effects of goal att ainment or failure on subsequent team leadership behaviors. In the current simulation scenarios, each team task was independent from each other, which did not allow the leader to adjust their behaviors based on the previous performance episodes. In organiz ational settings, it is conceivable that the success or failure of a prior task can influence how leaders prepare the team members for the next task, given similarity between the tasks. Future research should examine how goal attainment or failure influenc es team leadership behaviors across multiple team performance episodes. Furthermore, the current theoretical model demonstrated varying levels of fit to the data from participants. One future research direction is to modify the model and test the revised model empirically. Several directions for modifying the model were suggested by comparing the parameter to include is the individual difference of leaders to differe ntiate across subordinates. In the LMX literature, Henderson et al. (2009) suggested that differential treatment of subordinates might be influe . Therefore, future research could introduce a parameter to captur Another possible modification to the m odel is to introduce a mechanism adjustment of sensitivity to relative discrepancy over time . a suggested that differentiation in task difficulty had a strongly positive relationship with differentiation in leader time allocation when disturbance was large (vs. small), which was opposite to the expectation from the theoretical model . Optimization r esults suggested that team team performance episodes .

PAGE 129

129 Therefore, it is possible that team leaders were more sensitive to the relative discrepancy among subordinates when disturbance mag nitude was larger. Another direction for future research is to adjust the initial states of individual subordinates or the team as a whole. Vancouver , Weinhardt et al. (2010) suggested that when alternative choices differ in large enough expected values at the beginning of the performance episode, individuals might focus their regulatory efforts on one task. Future empirical research could test the effects of initial states on team leadership behavior. In addition, the simulation platform used in the presen t study did not allow the participants to set or adjust the team goals. Therefore, the present research only examined team leadership behavior within a single performance episode. Motivational research suggested individuals adjust their goal choices across multiple performance episodes (Locke & Latham, 1990). It is important for future research to examine how leaders set team goals across multiple performance episodes.

PAGE 130

130 Table 5 1. External negative disturbances. Disturbanc e Magnitud e b Time (minutes al located) a Target Low High 25 Subordinate A 20% 40% 50 Subordinate B 20% 40% 75 Subordinate C 20% 40% 100 Subordinate D 20% 40% 125 Subordinate A 20% 40% 150 Subordinate B 20% 40% 175 Subordinate C 20% 40% 200 Subordinate D 20% 40% 280 Subordinate A 20% 40% 320 Subordinate B 20% 40% 360 Subordinate C 20% 40% 400 Subordinate D 20% 40% 440 Subordinate A 20% 40% 480 Subordinate B 20% 40% 520 Subordinate C 20% 40% 560 Subordinate D 20% 40% Note. a The same schedule was used for all team performan ce episodes . b Disturbance was p rogrammed as the percentage of progress that was deducted from the current task state.

PAGE 131

131 Table 5 2. Relationship between discrepancy and time spent on the focal subordinate by conditions (including team focused actions). Esti mate Predictor Same task difficulty, small disturbance Same task difficulty, large disturbance Different task difficulty, small disturbance Different task difficulty, large disturbance Team Task Structure = Functional, Team Member Learning Ability = Same Difference score Gender .07 .07** .11** .04 Age .01 .02** .02* .00 Prior knowledge .05 .02 .05* .05 Wave .02 .03** .01 .02* Relative discrepancy 1.15** 1.40** .61** .62** Difference score component Gender .07 .07** .11** .04 Age .01 .02** .02* .00 Prior knowledge .05 .02 .05* .05 Wave .01 .01 .01 .01 Referent task state 3.69** 4.42** 1.57** 2.00** Non referent 1 task state 1.51** 1.44** .89* 1.12* Non referent 2 task state .44 1.29** .22 .28 Non referent 3 task state 1.76** 2.00** .87 .54 Team Task Structure = Functional, Team Member Learning Ability = Different Difference score Gender .03 .02 .02 .06 Age .01* .00 .00 .00 Prior knowledge .02 .01 .02 .02 Wa ve .01 .01 .02 .01 Relative discrepancy .33 .62** .32* .38* Difference score component Gender .03 .02 .02 .07 Age .01* .01 .00 .00 Prior knowledge .02 .01 .02 .03 Wave .02 .02 .03 .04 Referent task state .77 1.96** .54 1.18* Non referent 1 task state .07 .84* 1.41* .41 Non referent 2 task state .14 .86** 1.62* 2.16** Non referent 3 task state .90 .09 2.30** 1.11

PAGE 132

132 Table 5 2. Continued. Estimate Predictor Same task difficulty, small disturbance Same task difficulty, large disturbance Different task difficulty, small disturbance Different task difficulty, large disturbance Team Task Structure = Divisional, Team Member Learning Ability = Same Difference score Gender .03 .03 .04 .04 Age .0 1** .01** .01** .00 Prior knowledge .03 .00 .03 .02 Wave .01 .01 .02 .02 Relative discrepancy .36 .45 .53** .45** Difference score component Gender .03 .03 .04 .04 Age .01** .01** .01** .00 Prior knowledge .03 .01 .03 .02 Wave .01 .01 .02 .03 Referent task state 1.03 1.98* 1.68** 1.88** Non referent 1 task state .69 .54 .05 1.93* Non referent 2 task state 1.78* .12 2.41 .28 Non referent 3 task state .01 1.37 .65 .19 Team Task Structure = Di visional, Team Member Learning Ability = Different Difference score Gender .04 .10** .01 .10 Age .00 .00 .02 .04* Prior knowledge .02 .02 .01 .01 Wave .01 .01 .06** .01 Relative discrepancy 1.32** .48** .71** .20 Difference s core component Gender .04 .10** .03 .10 Age .01 .00 .02 .04* Prior knowledge .03 .03 .01 .01 Wave .02 .01 .02 .01 Referent task state 3.84** 1.74** 2.06** .44 Non referent 1 task state 2.04** 1.57** 1.94* .19 Non refere nt 2 task state .15 .20 1.09 .71* Non referent 3 task state 1.49 .20 1.22 .97

PAGE 133

133 Table 5 2. Continued. Estimate Predictor Same task difficulty, small disturbance Same task difficulty, large disturbance Different task difficulty, small disturbance Di fferent task difficulty, large disturbance Team Task Structure = Disjunctive, Team Member Learning Ability = Same Difference score Gender .03 .01 .04 .03 Age .01* .00 .01** .01 Prior knowledge .02 .01 .00 .02 Wave .03* .01 .0 4** .02* Relative discrepancy 1.24** .50 .75** .44** Difference score component Gender .03 .01 .03 .03 Age .01** .00 .01** .01 Prior knowledge .02 .01 .01 .02 Wave .04* .02 .02 .03 Referent task state 3.66** 1.43 3.38** 1.49** Non referent 1 task state 2.24* .72* .58 1.24** Non referent 2 task state .98 2.11** .38 .36 Non referent 3 task state .39 .01 2.20* .12 Team Task Structure = Disjunctive, Team Member Learning Ability = Different Difference score Gender .04 .03 .03 .12* Age .01* .01 .02** .01 Prior knowledge .02 .04 .05* .01 Wave .01 .01 .05 .01 Relative discrepancy .11 .34 .49* .18* Difference score component Gender .04 .03 .02 .12* Age .01* .01 .01* . 01 Prior knowledge .02 .04 .05* .01 Wave .01 .01 .05 .03 Referent task state 1.98* 1.25 2.23* .48* Non referent 1 task state 1.18 .18 .09 1.27* Non referent 2 task state .96 1.37* 2.80** 1.29 Non referent 3 task state .07 .04 1.39 1.82** Note. * p < .05, ** p < .01, two tail tests.

PAGE 134

134 Table 5 3. Relationship between discrepancy and time spent on the focal subordinate by conditions (excluding team focused actions). Estimate Predictor Same task difficulty, small disturbance Same task difficulty, large disturbance Different task difficulty, small disturbance Different task difficulty, large disturbance Team Task Structure = Functional, Team Member Learning Ability = Same Difference score Gender .02 .01 .02 .05 Age .00 .00 .00 .01 Prior knowledge .03 .01 .08 .06 Wave .02 .01 .03 .03** Relative discrepancy 1.84** 1.58** 1.02** .62** Difference score component Gender .02 .02 .02 .05 Age .00 .00 .00 .01 Prior knowledge .03 .01 .08 .06 Wave .00 .01 .04 .06** Referent task state 4.84** 2.91** 2.14** 1.86** Non referent 1 task state 1.01 .16 1.63* 1.45** Non referent 2 task state 2.83* .39 1.70 .15 Non referent 3 task state 1.11 2.52 1.15 .48 Team Task Struc ture = Functional, Team Member Learning Ability = Different Difference score Gender .02 .08 .06 .06 Age .01 .01 .03** .03** Prior knowledge .06* .00 .11* .02 Wave .01 .02 .02 .03* Relative discrepancy .77 ( p = .07) .98** .40* .62** Difference score component Gender .01 .07 .06 .07 Age .01 .01 .03** .03** Prior knowledge .07** .01 .11* .02 Wave .04 .02 .04 .09* Referent task state 2.56 3.08* 1.12 (p = .08) 1.98** Non referent 1 task state .57 1 .04 .52 .79 Non referent 2 task state .77 1.90 1.32 2.19** Non referent 3 task state .91 .11 .56 .52

PAGE 135

135 Table 5 3. Continued. Estimate Predictor Same task difficulty, small disturbance Same task difficulty, large disturbance Different task diffic ulty, small disturbance Different task difficulty, large disturbance Team Task Structure = Divisional, Team Member Learning Ability = Same Difference score Gender .02 .01 .01 .02 Age .00 .01** .03** .00 Prior knowledge .02 .00 .02 .08 Wave .00 .01 .05** .05* Relative discrepancy .76 .65 1.02** .90** Difference score component Gender .04 .03 .01 .02 Age .00 .01 .03** .00 Prior knowledge .01 .04 .02 .08 Wave .02 .00 .03 .10 Referent task state 2.82 * 2.85** 3.60** 2.50* Non referent 1 task state .34 .95 .12 1.45 Non referent 2 task state 3.10* .09 1.82 .43 Non referent 3 task state .03 1.80 1.58 .92 Team Task Structure = Divisional, Team Member Learning Ability = Different Diff erence score Gender .07 .00 .02 .31* Age .01 .02 .03 .01 Prior knowledge .02 .02 .00 .08 Wave .04* .01 .07* .05* Relative discrepancy 2.08** .70** .81** .26 Difference score component Gender .07 .00 .05 .31* Age .01 .01 .02 .01 Prior knowledge .03 .02 .01 .08 Wave .02 .01 .03 .09 Referent task state 6.14** 2.43** 2.56** 3.45 (p = .08) Non referent 1 task state 2.46* 1.50** 1.80* 1.27* Non referent 2 task state 1.79 .22 3.87** .18 Non referen t 3 task state 1.90 .55* 3.55** 1.19

PAGE 136

136 Table 5 3. Continued. Estimate Predictor Same task difficulty, small disturbance Same task difficulty, large disturbance Different task difficulty, small disturbance Different task difficulty, large disturbance Team Task Structure = Disjunctive, Team Member Learning Ability = Same Difference score Gender .12 .02 .08 .03 Age .00 .01** .00 .01* Prior knowledge .05 .03 .01 .06 Wave .04 .01 .07** .04* Relative discrepancy 2.42** .79 1.55** .23 Difference score component Gender .14 .05 .08 .04 Age .01 .01** .00 .01* Prior knowledge .04 .04 .01 .06 Wave .02 .03 .02 .02 Referent task state 6.51* 1.02 7.19** 1.56* Non referent 1 task state 1.98 1.45 1. 82 (p = .06) 1.87 Non referent 2 task state 2.66 2.23** .03 .39 Non referent 3 task state 2.02 .23 4.74* .31 Team Task Structure = Disjunctive, Team Member Learning Ability = Different Difference score Gender .01 .02 .00 .12 Age .01 .01 .02 .03** Prior knowledge .03 .01 .03 .03 Wave .03 .01 .14** .04** Relative discrepancy .49 .34 ( p = .06) 1.14** .05 Difference score component Gender .01 .02 .03 .12 Age .01 .01 .01 .03** Prior knowledge .03 . 02 .00 .03 Wave .02 .00 .15 .00 Referent task state 3.67* 1.08 (p = .09) 5.76** .40 Non referent 1 task state 1.33 .57 .35 1.65* Non referent 2 task state 1.98 1.46* 2.19* .32 Non referent 3 task state .18 .14 .28 1.33 Note. * p < .05, ** p < .01, two tail tests.

PAGE 137

137 Table 5 4. Leader time allocation by conditions (including team focused actions). % time spent on referent subordinate Team task structure Team member learning ability Position difficulty Disturbance magnitude n Obser ved Mean ( SD ) 95% CI of estimated marginal mean Expected value (M 0 ) Expected value (M 1 ) Expected value (M 2 ) Functional Same Same Small 14 .18 (.09) (.135, .223) .14 .17 .15 Functional Same Same Large .19 (.05) (.154, .230) .23 .35 .24 Functional Same D ifferent Small .24 (.08) (.184, .290) .13 .22 .12 Functional Same Different Large .28 (.09) (.217, .341) .24 .30 .28 Functional Different Same Small 15 .28 (.08) (.232, .318) .18 .23 .22 Functional Different Same Large .12 (.06) (.086, .16 0) .35 .57 .39 Functional Different Different Small .30 (.08) (.252, .354) .09 .10 .09 Functional Different Different Large .27 (.13) (.212, .331) .12 .14 .10 Divisional Same Same Small 15 .15 (.10) (.105, .191) .19 .33 .19 Divisional Same Same Large .18 (.10) (.148, .221) .34 .47 .42 Divisional Same Different Small .23 (.13) (.182, .285) .07 .10 .06 Divisional Same Different Large .26 (.10) (.197, .316) .13 .14 .13 Divisional Different Same Small 12 .21 (.08) (.158, .253) .18 .19 .19 Divisional Different Same Large .17 (.07) (.127, .209) .36 .51 .35 Divisional Different Different Small .21 (.14) (.155, .270) .13 .19 .17 Divisional Different Different Large .31 (.12) (.239, .373) .31 .51 .34 Disjunctive Sam e Same Small 16 .14 (.07) (.100, .182) .10 .13 .11 Disjunctive Same Same Large .14 (.07) (.109, .180) .02 .03 .04 Disjunctive Same Different Small .18 (.07) (.130, .230) .17 .29 .17 Disjunctive Same Different Large .26 (.12) (.204, .319) .37 .41 .37 Disjunctive Different Same Small 14 .24 (.07) (.192, .280) .17 .29 .17 Disjunctive Different Same Large .17 (.07) (.133, .210) .27 .41 .34 Disjunctive Different Different Small .24 (.08) (.186, .292) .05 .06 .03 Disjunctive Different Diffe rent Large .30 (.13) (.239, .362) .07 .10 .12

PAGE 138

138 Table 5 4. Continued. SD of time spent on subordinates Team task structure Team member learning ability Position difficulty Disturbance magnitude n Observed Mean ( SD ) 95% CI of estimated marginal mea n Expected value (M 0 ) Expected value (M 1 ) Expected value (M 2 ) Functional Same Same Small 14 .03 (.01) (.013, .042) .10 .11 .16 Functional Same Same Large .03 (.02) (.017, .044) .13 .22 .19 Functional Same Different Small .07 (.03) (.044, .099) .09 .15 .09 Functional Same Different Large .07 (.04) (.042, .098) .14 .18 .26 Functional Different Same Small 15 .08 (.03) (.069, .097) .11 .14 .15 Functional Different Same Large .05 (.02) (.038, .065) .21 .35 .24 Functional Different Different Small .14 (.05) (.111, .165) .08 .08 .16 Functional Different Different Large .09 (.05) (.059, .113) .06 .08 .14 Divisional Same Same Small 15 .04 (.03) (.021, .049) .11 .20 .13 Divisional Same Same Large .04 (.03) (.025, .051) .21 .29 .2 7 Divisional Same Different Small .10 (.06) (.071, .125) .06 .09 .06 Divisional Same Different Large .09 (.04) (.062, .116) .07 .07 .24 Divisional Different Same Small 12 .05 (.03) (.032, .063) .12 .11 .19 Divisional Different Same Large .04 (.02) (.026, .056) .17 .24 .24 Divisional Different Different Small .12 (.05) (.090, .150) .09 .11 .17 Divisional Different Different Large .10 (.06) (.069, .130) .14 .24 .21 Disjunctive Same Same Small 16 .03 (.02) (.016, .043) .08 .10 .11 Disjunctive Same Same Large .03 (.02) (.017, .043) .09 .17 .13 Disjunctive Same Different Small .06 (.04) (.039, .091) .10 .16 .11 Disjunctive Same Different Large .08 (.06) (.054, .107) .18 .19 .32 Disjunctive Different Same Small 1 4 .05 (.03) (.033, .062) .10 .16 .10 Disjunctive Different Same Large .05 (.03) (.035, .063) .13 .19 .28 Disjunctive Different Different Small .09 (.07) (.060, .115) .06 .11 .08 Disjunctive Different Different Large .10 (.06) (.072, .129) .11 .15 .17 Note. M 0 = model with team expectancy components and without individual expectancy components M 1 = model without team expectancy components and without individual expectancy components M 2 = model with team expectancy components and with individual expect ancy components

PAGE 139

13 9 Table 5 5. Leader time allocation by conditions (excluding team focused actions). % time spent on referent subordinate Team task structure Team member learning ability Position difficulty Disturbance magnitude n Observed Mean ( SD ) 95% CI of estimated marginal mean Expected value (M 0 ) Expected value (M 1 ) Expected value (M 2 ) Functional Same Same Small 14 .24 (.08) (.190, .281) .31 .25 .28 Functional Same Same Large .28 (.07) (.240, .314) .42 .41 .41 Functional Same Different Small . 35 (.12) (.254, .444) .29 .29 .27 Functional Same Different Large .38 (.12) (.301, .466) .43 .38 .42 Functional Different Same Small 15 .45 (.11) (.402, .490) .42 .34 .39 Functional Different Same Large .20 (.10) (.160, .232) .60 .60 .63 F unctional Different Different Small .48 (.20) (.392, .576) .20 .17 .16 Functional Different Different Large .40 (.15) (.316, .475) .25 .23 .19 Divisional Same Same Small 15 .25 (.10) (.202, .290) .40 .39 .38 Divisional Same Same Large .28 (.08) (.242, .314) .61 .55 .65 Divisional Same Different Small .44 (.24) (.351, .535) .18 .18 .16 Divisional Same Different Large .46 (.15) (.385, .544) .26 .21 .20 Divisional Different Same Small 12 .32 (.06) (.268, .372) .34 .27 .31 Divi sional Different Same Large .22 (.03) (.185, .269) .62 .60 .57 Divisional Different Different Small .32 (.22) (.242, .456) .30 .27 .32 Divisional Different Different Large .47 (.14) (.352, .539) .60 .60 .59 Disjunctive Same Same Small 16 . 28 (.07) (.236, .321) .27 .22 .20 Disjunctive Same Same Large .25 (.07) (.214, .284) .04 .05 .07 Disjunctive Same Different Small .35 (.16) (.257, .434) .36 .37 .35 Disjunctive Same Different Large .41 (.19) (.331, .485) .61 .49 .54 Disju nctive Different Same Small 14 .31 (.07) (.261, .352) .37 .37 .37 Disjunctive Different Same Large .22 (.05) (.178, .253) .53 .49 .54 Disjunctive Different Different Small .32 (.12) (.222, .412) .15 .11 .08 Disjunctive Different Different Large .40 ( .18) (.317, .482) .17 .16 .19

PAGE 140

140 Table 5 5. Continued. SD of time spent on subordinates Team task structure Team member learning ability Position difficulty Disturbance magnitude n Observed Mean ( SD ) 95% CI of estimated marginal mean Expected value (M 0 ) Expected value (M 1 ) Expected value (M 2 ) Functional Same Same Small 14 .05 (.06) (.023, .087) .22 .16 .29 Functional Same Same Large .05 (.04) (.024, .070) .26 .25 .32 Functional Same Different Small .11 (.06) (.059, .167) .20 .20 .22 Functional Same Different Large .10 (.07) (.056, .147) .28 .22 .39 Functional Different Same Small 15 .14 (.07) (.110, .172) .27 .19 .30 Functional Different Same Large .09 (.04) (.065, .109) .39 .37 .39 Functional Different Different Small .22 (.10) (.172, .276) .17 .13 .27 Functional Different Different Large .13 (.08) (.088, .176) .13 .13 .24 Divisional Same Same Small 15 .09 (.08) (.060, .123) .24 .24 .28 Divisional Same Same Large .06 (.04) (.039, .083) .42 .33 .41 Divisional Sam e Different Small .20 (.14) (.149, .253) .15 .16 .15 Divisional Same Different Large .17 (.08) (.121, .209) .14 .10 .36 Divisional Different Same Small 12 .08 (.04) (.042, .115) .24 .17 .31 Divisional Different Same Large .06 (.04) (.034, .086) .29 .28 .37 Divisional Different Different Small .22 (.12) (.131, .253) .22 .17 .32 Divisional Different Different Large .16 (.08) (.092, .196) .26 .28 .36 Disjunctive Same Same Small 16 .06 (.04) (.031, .092) .21 .15 .23 Disjunctive Same Same Large .06 (.05) (.041, .084) .22 .20 .23 Disjunctive Same Different Small .14 (.11) (.087, .188) .23 .22 .23 Disjunctive Same Different Large .13 (.11) (.087, .172) .30 .23 .45 Disjunctive Different Same Small 14 .07 (.05) (.034 , .099) .22 .22 .23 Disjunctive Different Same Large .06 (.04) (.041, .087) .26 .23 .45 Disjunctive Different Different Small .12 (.09) (.068, .176) .14 .16 .18 Disjunctive Different Different Large .14 (.09) (.095, .187) .29 .19 .27 Note. M 0 = mode l with team expectancy components and without individual expectancy components M 1 = model without team expectancy components and without individual expectancy components M 2 = model with team expectancy components and with individual expectancy components

PAGE 141

141 T able 5 6. Estimates of dynamic model parameters and model data fit aggregated across individuals. M 0 Team task structure Team member learning ability Position difficulty Disturbance magnitude Task state bias A Task state bias B Task state bias C Task state bias D Leader reactivity Relative discrepancy sensitivity Model data fit (% accurate prediction) Functional Same Same Same Different Different Small .03 .07 .04 .03 .99 .02 33.47 Functional Same Large .01 .03 .00 .00 .99 .04 42.15 Functiona l Same Small .06 .02 .01 .02 .98 .01 30.85 Functional Same Large .01 .14 .01 .04 .98 .07 36.64 Functional Different Same Same Different Different Small .08 .01 .00 .11 1.00 .03 29.34 Functional Different Large .02 .00 .01 .00 .97 .00 42.1 5 Functional Different Small .18 .12 .01 .09 1.00 .01 31.27 Functional Different Large .05 .04 .01 .02 .99 .02 36.78 Divisional Same Same Same Different Different Small .01 .06 .01 .04 .99 .11 49.09 Divisional Same Large .05 .01 .00 .01 1.00 .01 43.64 Divisional Same Small .16 .01 .01 .02 1.00 .04 51.44 Divisional Same Large .03 .08 .00 .01 1.00 .01 40.33 Divisional Different Same Same Different Different Small .10 .03 .00 .14 .99 .01 27.94 Divisional Different Large .0 3 .03 .01 .02 .97 .01 32.64 Divisional Different Small .10 .02 .01 .03 .98 .01 19.56 Divisional Different Large .01 .02 .00 .01 .99 .01 43.14 Disjunctive Same Same Same Different Different Small .00 .06 .06 .00 .92 .01 42.70 Disjunctive S ame Large .02 .07 .02 .02 .99 .01 36.78 Disjunctive Same Small .01 .05 .00 .03 .98 .01 34.99 Disjunctive Same Large .03 .05 .01 .01 .98 .02 34.57 Disjunctive Different Same Same Different Different Small .01 .01 .04 .02 .99 .00 34.44 Dis junctive Different Large .02 .03 .03 .00 .98 .00 32.65 Disjunctive Different Small .01 .06 .00 .00 .97 .00 34.02 Disjunctive Different Large .06 .02 .04 .01 .98 .03 28.10

PAGE 142

142 Table 5 6. Continued. M 1 Team task structure Team member learning ability Position difficulty Disturbance magnitude Task state bias A Task state bias B Task state bias C Task state bias D Leader reactivity Relative discrepancy sensitivity Model data fit (% accurate prediction) Functional Same Same Same Different Differ ent Small .00 .05 .02 .00 .99 .01 31.82 Functional Same Large .02 .03 .01 .00 .99 .01 38.02 Functional Same Small .03 .08 .00 .02 .90 .01 26.86 Functional Same Large .00 .06 .00 .01 1.00 .01 37.19 Functional Different Same Same Different Different Small .05 .04 .00 .02 1.00 .01 29.55 Functional Different Large .04 .04 .02 .01 .99 .01 40.50 Functional Different Small .09 .02 .01 .08 .85 .05 33.20 Functional Different Large .02 .04 .02 .02 .99 .03 36.09 Divisional Same Sa me Same Different Different Small .05 .03 .03 .02 1.00 .04 37.69 Divisional Same Large .03 .06 .02 .05 1.00 .06 34.54 Divisional Same Small .12 .03 .01 .01 .98 .04 47.52 Divisional Same Large .00 .02 .01 .03 .97 .01 39.67 Divisional Diff erent Same Same Different Different Small .03 .03 .00 .03 .99 .01 22.48 Divisional Different Large .04 .03 .00 .04 .98 .02 34.99 Divisional Different Small .06 .05 .00 .09 .90 .04 22.31 Divisional Different Large .01 .03 .01 .00 1.00 .02 37.52 Disjunctive Same Same Same Different Different Small .01 .02 .06 .02 1.00 .02 26.31 Disjunctive Same Large .00 .04 .02 .01 1.00 .02 23.16 Disjunctive Same Small .05 .02 .01 .00 .99 .03 48.76 Disjunctive Same Large .03 .03 .03 .00 1.00 .05 40.3 6 Disjunctive Different Same Same Different Different Small .02 .00 .02 .01 .97 .01 26.72 Disjunctive Different Large .00 .01 .04 .00 1.00 .02 30.58 Disjunctive Different Small .01 .03 .03 .00 .99 .02 22.73 Disjunctive Different Large .0 3 .00 .01 .00 .99 .01 29.42

PAGE 143

143 Table 5 6. Continued. M 2 Team task structure Team member learning ability Position difficulty Disturbance magnitude Task state bias A Task state bias B Task state bias C Task state bias D Leader reactivity Relativ e discrepancy sensitivity Model data fit (% accurate prediction) Functional Same Same Same Different Different Small .03 .04 .00 .05 .99 .00 28.51 Functional Same Large .02 .02 .00 .00 .97 .01 42.15 Functional Same Small .09 .02 .04 .03 .99 .00 30.86 Functional Same Large .01 .14 .00 .01 .97 .01 34.57 Functional Different Same Same Different Different Small .22 .03 .01 .01 .98 .03 27.89 Functional Different Large .16 .00 .02 .00 .98 .04 44.63 Functional Different Small .07 .01 .01 .0 2 .97 .07 31.55 Functional Different Large .08 .03 .08 .00 .95 .02 34.57 Divisional Same Same Same Different Different Small .07 .04 .01 .01 1.00 .03 50.74 Divisional Same Large .06 .05 .03 .02 .99 .00 36.53 Divisional Same Small .13 .03 .00 .01 .98 .02 43.18 Divisional Same Large .01 .04 .00 .01 .99 .03 38.68 Divisional Different Same Same Different Different Small .02 .01 .01 .01 .99 .01 26.45 Divisional Different Large .15 .02 .01 .01 .99 .00 34.30 Divisional Differen t Small .15 .05 .00 .02 .98 .01 29.75 Divisional Different Large .11 .00 .01 .00 .99 .01 38.51 Disjunctive Same Same Same Different Different Small .00 .01 .03 .00 1.00 .01 40.50 Disjunctive Same Large .05 .03 .02 .01 1.00 .02 35.40 Disj unctive Same Small .04 .04 .04 .02 .99 .01 43.94 Disjunctive Same Large .07 .03 .01 .00 1.00 .02 39.53 Disjunctive Different Same Same Different Different Small .01 .02 .06 .05 1.00 .00 32.64 Disjunctive Different Large .01 .01 .03 .00 .9 8 .01 33.68 Disjunctive Different Small .02 .01 .09 .00 .99 .00 20.25 Disjunctive Different Large .03 .01 .02 .02 1.00 .03 30.25 Note. M 0 = model with team expectancy components and without individual expectancy components M 1 = model without team expe ctancy components and without individual expectancy components M 2 = model with team expectancy components and with individual expectancy components

PAGE 144

144 Figure 5 1. Predicted (baseline), predicted (calibrated), and observed choices over time. (Note. The tea m performance episode illustrated was from the condition of functional team task structure, same team member learning repre sents team focused actions.)

PAGE 145

145 CHAPTER 6 STUDY 3: HYPOTHESES TESTING USING A FIELD SAMPLE Study 3 aims to further test hypotheses derived from the model regarding how characteristics of the positions and members in the team and team task structure intera ct to influence team leadership behaviors . Specifically, Study 3 examines the moderation effect of task interdependence on the relationships between differentiation in team position characteristics (i.e., skill variety, job complexity, problem solving, and workflow centrality) and team member characteristics (i.e., learning ability and conscientiousness) on differentiated individual focused leadership behaviors, which replicates part of Study 2 in a field sample . In addition, Study 3 examines the moderation effect of task interdependence on the relationships between differentiation in team position and member characteristics, and average individual focused leadership behaviors and team performance. Study 3 extends Study 2 in two ways. First, Study 3 includes another two important outcomes of structural input factors to the team performance system, i.e., average individual focused leadership behaviors and team performance. Second, Study 3 examines more specific dimensions of team position and member characteri stics as compared to Study 2. The current model specifies that when task interdependence is high, differences among team positions and team members are positively related to the average level of individual focused leadership behaviors . Specifically, with a ll else being equal, when team positions va ry in knowledge characteristics (i.e., skill variety, job complexity, and problem solving; Morgeson & Humphrey, 2006) or workflow network centrality (Humphrey et al., 2009 ) , how fast individual positions progress toward the goals are likely to be different. When task interdependence is high, the performance state s of the positions with slowest task progress can influence the performance states of other positions and the overall performance of the team . Therefore, m ore individual -

PAGE 146

146 focused leadership behaviors are needed for both the positions with slowest progress as well as the related positions . When task interdependence is low, the performance state s of the positions with slowest progress are independent from the p erformance states of the other positions. More individual focused leadership behaviors are needed for the positions with slower progress whereas fewer individual focused leadership behaviors are needed for the positions with faster progress. Therefore, the average amount of individual focused leadership behaviors does not increase as compared to teams with identical positions. Similar to the effects of team positions, differentiation in team member characteristics can also influence the average level of ind ividual focused leadership behaviors when t ask interdependent is high . When team members have different levels of learning ability or conscientiousness , task progress of the member with lowest learning ability or conscientiousness can influence other team members when task interdependence is high. When task interdependence is low, the task states of the members are independent from each other. Therefore, the average of individual focused leadership behaviors increases as team members differentiate more from each other, when task interdependence is high. Taken together, task interdependence moderates the effect of team position and member characteristics on average individual focused leadership behaviors. Hypothesis 1: When task interdependence is high (vs. low) , the positive relationship between differentiation in (a) skill variety, ( b ) job complexity, ( c ) problem solving, (d) centrality in workflow, (e) learning ability, and (f) conscientiousness and average individual focused leadership behavior is stronge r . Consistent with hypotheses derived for Study 2, the current model specifies that when task interdependence is high, there are stronger positive relationships between differentiation

PAGE 147

147 among team positions in knowledge characteristics and workflow centrali ty, differentiation among team members in learning ability and conscientiousness, and differentiated individual focused leadership behaviors . Hypothesis 2: When task interdependence is high (vs. low) , the positive relationship between differentiation in (a ) skill variety, (b) job complexity, (c) problem solving, (d) centrality in workflow, (e) learning ability, and (f) conscientiousness and differentiated individual focused leadership behavior is stronger . As simulation results from Study 1 revealed, w ith a ll else being equal, differentiation among team members and positions is negatively related to team performance when task interdependence is high. This is partly because when task interdependence is high, team performance is more strongly influenced by the performance of the subordinate on the most difficult position or with the lowest motivation or learning ability whereas when task interdependence is lower, team performance is less influenced by the difference s between the highest and lowest performers. H ypothesis 3: When task interdependence is high (vs. low) , the negative relationship between differentiation in (a) skill variety, (b) job complexity, (c) problem solving, (d) centrality in workflow, (e) learning ability, and (f) conscientiousness and team performance is stronger . Method Sample and Procedure Employees working for the research and development (R & D) department of a large IT company located in South China were recruited for the present study. Employees in the R & D department worked in teams in charge of software development, data mining, product evaluation, etc. In some teams, the positions held by team members were in charge of similar responsibilities whereas in some teams the positions varied in terms of job responsibilities. Therefore, th e sample

PAGE 148

148 was suitable for studying how position characteristics vary within the teams. The teams also had varying levels of interdependence among different positions, which allowed the possibility of observing different levels of task interdependence. Each team had one external leader, who did not routinely engage in performing day to day team task and was mainly responsible for coaching, monitoring, and supporting the team, as well as representing the team to other organizational units. All team members in the same group had the same team leader as their direct report in the organization. Team members regularly interacted with the team leaders in meetings, discussions, and social events. The human resources department of the company solicited voluntary par ticipation from employees in the R & D department. Participants were assured of the confidentiality of their individual responses. Participants were also allowed to respond to the surveys during work hours. Responses to the measures were collected from dif ferent sources at different time points, in order to minimize common method variance in the data (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). At Time 1, subordinates of each group responded to questions about their demographic information and measures o f skill variety, job complexity, problem solving, learning ability, and conscientiousness. Subordinates of each team also responded to a social network measure about the task flow within the team. Team leader responded to the measure of leadership behavior s involved with each individual subordinate. At Time 2, external leader rated team performance. Time 1 and Time 2 measurement was monthly performance evaluation cycle, with about one month in between. Questionnaires were distributed to 313 subordinates and 62 supervisors working in all of the 62 teams in the department. Data from groups with lower than 80% response rate were not included in the analyses. Therefore, usable responses were collected from 281 subordinates

PAGE 149

149 (res ponse rate = 89.78%) and 58 leaders (response rate = 93.55%) from 58 teams. Among these subordinates, the average age was 29.15 years old ( SD = 4.42). 73.3% ( n = 206) of them were men. Average organizational tenure of the subordinates was 2.46 years ( SD = 2.06). The average age of the supervisors was 33.07 years old ( SD = 3.66). 81% ( n = 47) of them were men. Average organizational tenure of the supervisors was 5.15 years ( SD = 3.00). Team size ranged from 3 to 11 subordinates per team ( median = 5). Measure s All questionnaires were in Chinese. The Chinese versions of the scales were translated by back translation procedure. English versions of all the measures are listed in Appendix C. D ifferentiation in s kill variety, job complexity, and problem solving. Skill variety, job complexity, and problem solving were measured by three subscales (4 item each) from ing from 1 ( strongly disagree) to 5 ( strongly agree job complexity, and problem solving scale , respectively. Differentiation in skill variety, job complexity, and problem solving within teams were measured by the SD ratings. D ifferentiation in l earning ability. Learning ability was measured by five items adapted stay at the forefront

PAGE 150

150 (strongly disagree) to 5 (strongly agree). Differentiation in learning ability was measured by the SD D ifferen tiation in c onscientiousness. Conscientiousness was measured by the eight item item on a scale ranging from 1 ( extremely inaccurate) to 9 ( extremely accurate ) alpha for this scale was .84. Differentiation in conscientiousness was measured by the SD of the D ifferentiation in centrality in workflow (w orkflow network centralization ) . Workflow network was measured by the roster m ethod that is commonly used in social network research (Brass, 2012). Subordinates of each work group were provided with a list of names of team members in their work group. Subordinates were asked to recall their relations with peer teammates and place ch ecks to names of those whom were considered as workflow contacts. Following previous research (e.g., Mehra, Kilduff, & Brass, 2001), workflow network question centr ality in workflow was measured by betweenness centrality score. Betweenness centrality was calculated as the extent to which an actor was on the shortest path between two other actors in the workflow network (Wasserman & Faust, 1994). The betweenness centr ality scores were normalized to account for the different sizes of the teams, resulting in a score ranging from 0 to 1 (with 1 indicates larger betweenness centrality). Workflow network centralization was measured by the betwennness centralization of a net betweenness centrality in a network. Betweenness centralization was calculated as the sum of the

PAGE 151

151 and divided by the maximum possible value for the sum of deviation (Freeman, 1979; Wasserman & Faust, 1994). Task interdependence. Following previous research (e.g., Balkundi & Harrison, 2006; Oh, Chung, & Labianca, 2004), task interdependence was measured by the den sity of work network divided by the maximum number of possible ties among nodes (i.e., subordinates) in the network (Wasserman & Faust, 1994). Calculations of t he network indices were all conducted in software ORA 2013 (Carley, Pfeffer, Reming a, Storrick, & Columbus, 2013). Average and Differentiated i ndividual focused leadership behavior. Individual focused leadership behaviors were measured by items developed b y Zhou, Wang, Liu, Penn, and Shi (2013). Three items were used to measure information provided by leader to subordinates I provide this subordinate with knowledge, skills, and expertise to help him/her get the job done subordinate considerable opportunity for independence and freedom in how he/she does the This subordinate can rely on me for emotional support point Likert scale ranging from 1 ( strongly disagree) to 5 ( strongly agree interaction with each subordinate. on giving, autonomy giving, and socioemotional support scale, respectively. Average and differentiat ed individual focused leadership behavior were measured by the mean and SD of the team Team performance. External managers rated team perf ormance on the four item scale

PAGE 152

152 . A seven point Likert response scale ranging from 1 ( strongly disagree) t o 7 ( strongly agree ) was used scale was .90. Covariates. Previous research suggested that team size and team demographic Henderson et al. , 200 9 ). Thus, team size, average and standard deviation of supervisor subordinate time working together, gender diversity, and average and standard deviation of Leader education level and managerial experience were included as covariates as well, in order to control f or the effects of leader ability and experiences on differentiated leadership (Liden et al., 2006) . I also controlled for the means of skill variety, job complexity, proble m solving, learning ability, and conscientiousness, in order to provide a more accu rate test of the effects of the SD s of these factors (Harrison & Klein, 2007). Time working with the supervisor, gender, education level, team size , leader education level, and leader managerial experience were measured in Confirmatory Factor Analyses Skill variety, job complexity, problem solvin g, learning ability, and conscientiousness were measured from the same source at the same time. Therefore, I conducted confirmatory factor analyses (CFA) to examine whether the responses on these measures indicated five distinctive constructs. An expected five factor model was specified by loading observed indicators on their respective latent variables and having the correlations among the latent variables to be freely estimated. CFA results suggested that the five factor model fit the data well, 2 ( 265 , N = 281) = 552.73 , comparative fit index (CFI) = .9 1 , root mean square error of

PAGE 153

153 approximation (RMSEA) = .06, and standardized root mean square residual (SRMR) = .0 7 . All indicators significantly loaded on their respective latent factors. Considerin g that skill variety, job complexity, and problem solving items were all describing job characteristics, three alternative f our factor models were specified by having items from two scales loading on a single factor. The f our factor models fit the data sig nificantly worse than the five factor model, 2 ( 5 , N 222.05 , p < .01. These results suggested that the five measures responded by the subordinates at Time 1 captured five distinctive constructs. Another set of CFA was conducted to examine whether measures of the three types o f individual focused leadership behaviors captured distinctive constructs. An expected three factor model was specified by loading observed indicators on their respective latent variables and having the correlations among the latent variables to be freely estimated. CFA results suggested that the three factor model fit the data well, 2 (24, N = 281) = 43.7 3 , CFI = .99, RMSEA = .05, and SRMR = .03. All indicators significantly loaded on their respective latent factors. Three alternative two factor models were specified by having items from two scales loading on a single factor. The two factor models fit the data significantly worse than the three factor model, 2 (3, N p < .01. These results suggested that the three measures responded by the supervisors at Time 1 captured three distinctive types of leadership behaviors. Analytic Strategy In order to t est Hypotheses 1 3, hierarchical multiple regression analyses were conducted. All estimations were conducted using software Mplus 6.0 (Muthén & Muthén, 2010). The regression model was specified as: 0 1 2 (C2 ) + 3 (C3 4 (C4 5 (C5 6 (C6 7 (C7 8 (C8 ) + 9 (C9 1 0 (C1 0 ) 11 12 13 14 15 16 17 (X4) +

PAGE 154

154 18 19 20 21 22 23 24 25 (X5Z) + 26 (X6Z) where: = re gression coefficients representing average time working with gender representing differentiat ting task interdependence. Results Means, standard deviations, and bivariate correlations among the study variables are presented in Table 6 1. In order to test the hypotheses , I conducted regression analyses with average and differentiat ed information gi ving, autonomy giving, and socioemotional support, and team performance as the outcomes. All outcome variables were regressed on the covariates (i.e., leader education, leader managerial experience, team size, average and SD of time working with supervisor and subordinate education, gender diversity, and average skill variety, task complexity, problem solving, learning ability, and conscientiousness), main effects of task interdependence, differentiation in skill variety, task complexity, problem solving, l earning

PAGE 155

155 ability, conscientiousness, and workflow centralization, and the interaction terms between task interdependence and workflow centralization, differentiation in skill variety, task complexity, problem solving, learning ability, and conscientiousness . All predictor variables were mean centered to remove the nonessential collinearity between the predictor variables and to facilitate the interpretation of the findings (Cohen, Cohen, West, & Aiken, 2003). The interaction terms were formed by multiplying the mean centered scores. Analyses results are presented in Table 6 2. The interaction terms between task interdependence and workflow centralization, differentiation in skill variety, task complexity, problem solving, learning ability, and conscientiousn ess collectively accounted for 20 % of variance in average information giving, 23% of variance in average autonomy giving, 20% of variance in differentiat ed information giving, 19 % of variance in team performance, which was significantly above and beyond th e effects of the covariates and the main effects ( s[6, 3 1 ] = 2. 97 , 3.90 , 2.36 , and 2.61 , p s .05). Specifically, the interaction term between task interdependence and differentiation in skill variety was related to average information giving, average autonomy giving, average socioemotional support, di fferentiated information giving, and team performance ( B s = 7.14, 7.12, 5.68, 3.05, and 7.97 , p s < .05). The interaction term between task interdependence and differentiation in task complexity was related to team performance ( B = 10.33 , p < .05). The i nteraction term between task interdependence and differentiation in problem solving was related to differentiat ed information giving ( B = 3.64 , p < .01). The interaction term between task interdependence and workflow network centralization was related to d ifferentiat ed information giving and team performance ( B s = 9. 26 and 26.04 , p s < .05). The interaction term between task interdependence and differentiation in learning ability was related to average autonomy giving ( B

PAGE 156

156 = 3.34 , p < .05). The interaction te rm between task interdependence and differentiation in conscientiousness was related to differentiat ed information giving ( B = 4.74 , p < .05). I plotted these interactions following procedures recommended by Cohen et al. (2003). As shown in Figures 6 1, 6 2, and 6 3, when task interdependence was high (one SD above the mean), differentiation in skill variety was positively related to average information giving ( B = 1.37 , p < .01) and average autonomy giving ( B = 1.38 , p < .05), and was not significantly re lated to average socioemotional support; when task interdependence was low (one SD below the mean), differentiation in skill variety was not significantly related to average information giving and average autonomy giving, and was negatively related to aver age socioemotional support ( B = .9 6 , p = . 05 ). This finding suggested that when task interdependence was high (vs. low) , the positive relationship between differentiation in skill variety and average individual focused leadership behavior was stronger. A s shown in Figure 6 4, when task interdependence was high (one SD above the mean), differentiation in conscientiousness was positively related to average autonomy giving ( B = 1. 02 , p < .05); when task interdependence was low (one SD below the mean), differ entiation in conscientiousness was not significantly related to average autonomy giving. This finding suggested that when task interdependence was high (vs. low) , the positive relationship between differentiation in conscientiousness and average autonomy g iving was stronger. Therefore, Hypotheses 1a and 1 f were supported, but not Hypotheses 1b, 1c, 1d, and 1 e . As shown in Figure 6 5, when task interdependence was high (one SD above the mean), differentiation in skill variety was not significantly related to differentiat ed information giving; when task interdependence was low (one SD below the mean), differentiation in skill variety was positively related to differentiat ed information giving ( B = . 60 , p < .05). This finding suggested

PAGE 157

157 that when task interdepen dence was high (vs. low) , the positive relationship between differentiation in skill variety and differentiat ed information giving was weaker. As shown in Figure 6 6 , when task interdependence was high (one SD above the mean), differentiation in problem so lving was positively related to differentiat ed information giving ( B = . 64 , p < .05); when task interdependence was low (one SD below the mean), differentiation in problem solving was not significantly related to differentiat ed information giving. This fin ding suggested that when task interdependence was high (vs. low) , the positive relationship between differentiation in problem solving and differentiat ed information giving was stronger. As shown in Figure 6 7, when task interdependence was high (one SD a bove the mean), workflow network centralization was not significantly related to differentiated information giving; when task interdependence was low (one SD below the mean), workflow network centralization was positively related to differentiated informat ion giving ( B = 1.94 , p = . 07 ). This finding suggested that when task interdependence was high (vs. low), the positive relationship between workflow network centralization and differentiated information giving was weaker. As shown in Figure 6 8, when task interdependence was high (one SD above the mean), differentiation in learning ability was positively related to differentiat ed information giving ( B = . 51 , p < .05); when task interdependence was low (one SD below the mean), differentiation in learning ab ility was not significantly related to differentiat ed information giving. This finding suggested that when task interdependence was high (vs. low) , the positive relationship between differentiation in learning ability and differentiat ed information giving was stronger. Therefore, Hypotheses 2 c and 2 e were supported, not Hypotheses 2a, 2 b, 2d , and 2 f . As shown in Figure 6 9, when task interdependence was high (one SD above the mean), differentiation in skill variety was not significantly related to team perf ormance; when task

PAGE 158

158 interdependence was low (one SD below the mean), differentiation in skill variety was positively related to team performance ( B = 1.81 , p < .05). This finding suggested that when task interdependence was high (vs. low) , the negative rela tionship between differentiation in skill variety and team performance was stronger. As shown in Figure 6 10 , when task interdependence was high (one SD above the mean), differentiation in task complexity was negatively related to team performance ( B = 1 .38, p = .06) ; when task interdependence was low (one SD below the mean), differentiation in task complexity was positively related to team performance ( B = 1. 31 , p = . 07 ). This finding suggested that when task interdependence was high (vs. low) , the negat ive relationship between differentiation in task complexity and team performance was stronger. As shown in Figure 6 11, when task interdependence was high (one SD above the mean), workflow network centralization was not significantly related to team perfo rmance; when task interdependence was low (one SD below the mean), workflow network centralization was negatively related to team performance ( B = 8.62 , p < .05). This finding suggested that when task interdependence was high (vs. low) , the negative relat ionship between workflow network centralization and team performance was weaker. Therefore, Hypotheses 3a and 3b were supported, but not Hypotheses 3c, 3d, 3e, and 3f. Discussion The current study found that task interdependence moderated the effects of t eam position and team member characteristics on team leadership behaviors and team performance. Specific ally, this study found that task interdependence moderated the relationships between differentiation in skill variety and average information giving, au tonomy giving, and socioemotional support from team leaders, such that the positive relationship s were stronger when task interdependence was higher. Task interdependence also moderated the relationship

PAGE 159

159 between differentiation in conscientiousness and aver age autonomy giving, such that the positive relationship was stronger when task interdependence was higher. In addition, the current study found that t he positive relationships between differentiation in skill variety and differentiated information giving and between workflow network centralization and differentiated information giving were weaker when task interdependence was higher. The positive relationships between differentiation in problem solving and differentiated information giving and between diff erentiation in learning ability and differentiated information giving were stronger when task interdependence was higher. This study also found that t he negative relationships between differentiation in skill variety and team performance and between differ entiation in task complexity and team performance were stronger when task interdependence was higher. The negative relationship between workflow network centralization and team performance was weaker when task interdependence was higher . Findings about the se joint effects of team design factors and team composition on team leadership and team performance revealed several theoretical and practical implications. Theoretical Implications First, findings of the current study suggested that team task structure i s an important boundary condition on the relationship between team member characteristics and team leadership behaviors. This study found that when task interdependence was higher, the positive relationship between differentiation of team members in consci entiousness and average autonomy giving was stronger, and the positive relationship between differentiation of team members in learning ability and differentiat ed information giving was stronger. These findings are consistent with previous research which f ound that team design factors could change the relationship s between team member characteristics and team outcomes (e.g., LePine, 2005; LePine et al., 1997). The current study extended the previous research by showing that team task

PAGE 160

160 structure could also in fluence the relationship s between team composition characteristics and team leadership behaviors, which were in turn related to team performance state. Second, the study found that differences across the team positions in skill variety was positively rel ated to the average amount of team leaders hip behaviors toward the team members as well as the differentiated leadership in the teams , with the strength of the relationship contingent on team task structure. This finding supports the idea that teams with h ighly specialized positions and high task interdependence among the positions is more likely to be influenced by uncertainty in teak environment (Hollenbeck et al., 2002). Accordingly, in order to counter the influence of external disturbance and meet the task requirement s , more intervention is required from team leadership to ensure team functioning. This finding is also consistent with ( i.e., sub system s such as leadership play important su pporting role in the organizational system when the jobs are highly functional and specialized ) . Third, the current study suggested that differentiated leadership in teams was under the joint influence of team task structure and characteristics of position s in the team. Supporting the hypotheses derived from the theoretical model, the positive relationship between differentiation in problem solving and differentiat ed leadership was stronger when task interdependence was higher. This finding advances existin g research on position characteristics in the team, which noticed that different positions in the team have different amount of problem solving responsibilities (e.g., Humphrey et al., 2009). The current study further demonstrated that such differences acr oss positions could influence how team leaders treat team members in different positions.

PAGE 161

161 Interestingly, this study found that the positive relationships between differentiation in skill variety and workflow network centralization , and differentiat ed indiv idual focused leadership were weaker when task interdependence was higher. The study also found that the negative relationship between workflow network centralization and team performance was weaker when task interdependence was higher. One possible explan at ion for this finding is that when task interdependence was higher, team members had more opportunities to influence each other. Therefore, when a special task exceeds the competency of the members with less skill variety, they could receive help from tho se members with more skill variety. When a task demand exceeds the competency or workload of the members in more peripheral positions, they could receive support from those in the central of workflow directly or indirectly through other members connected b y the central position. Future research could examine this possibility by incorporating inter member regulatory mechanisms into the model. Findings of the current study also suggested that the observed correlation between differentiat ed individual focused leadership and team performance might be caused by differentiation in team position characteristics. T his study found that when task interdependence was low, workflow network centralization was positively related to differentiated information giving and n egatively related to team performance. In a supplementary analysis, when controlling for the effects of workflow network centralization on differentiated information giving and team performance, the correlation between differentiated information giving and team performance was not significant . This finding suggested that correlated team states (e.g., differentiat ed leadership behaviors and team performance state) as found in previous research (e.g., Wu et al., 2010) might be outcomes that both resulted from the same input factors (e.g., differentiation in team position characteristics) .

PAGE 162

162 Practical Implications The current study provided some suggestions for selecting team members. The study found that when task interdependence was high, differentiation amon g team members in conscientiousness was positively related to average individual focused leadership behaviors and differentiation among team members in learning ability was positively related to differentiated individual focused leadership behaviors. There fore, when using highly interdependent structure to organize tasks in work teams, to reduce the amount of differential treatment among subordinates or decrease the amount of workload of team leaders, employees that are relatively similar in learning abilit y and conscientiousness can be composed into the same team. In addition, it is important for higher level managers to realize that differentiation in individual focused leadership behaviors might result from existing differences of team members. Although t here might be unwanted effects associated with differentiat ed individual focused leadershi p (e.g., perceived exclusion or lower positive attitude toward the leader), such differentiation is one way through which leader s respond to team composition characte ristics. The current study further suggested that when organizations design the interdependent structure of team tasks, they should also be aware of characteristics of the positions. This study found that when task interdependence was high, differentiati on in task complexity across positions could be negatively related to team performance. However, when task interdependence was low, differentiation in skill variety could be positively related to team performance. Therefore, organizations should take multi ple aspects of team design factors into consideration in order to maximize team performance. Limitations and Future Research Directions The current study had several limitations. First, the current study measured team design ubjective ratings. Piccolo and Colquitt (2006) found that

PAGE 163

163 (i.e., task variety, identity, significance, autonomy, and feedback from job), which in turn influenced goal commitment and task performance. The current study focused on knowledge characteristics of the jobs (i.e., skill variety, job complexity, and problem solving), which have been shown to be conceptually distinctive from task characteristics perceived b y the employees (Morgeson & Humphrey, 2006). Although both types of characteristics could be assessed by eristics are less likely to be influenced by leadership behaviors. Nevertheless, future studies should test the hypotheses using objective ratings of the job character istics from job analyses . Second, the current study did not directly examine the mediatio n processes between team design and team composition factors , and team leadership and team performance. Team performance was measured by evaluations from higher level managers, which did not allow examining the trajectory of team performance state from sta rting a team task to reaching team goal. Future studies should use longitudinal design to collect multiple observations of team leadership and team performance state over time, and estimate how team design and team composition factors influence the dynamic relationships between team leadership and team performance. the relationships between team composition in learning ability and conscientiousness and team leadership. Although I controlled for demographic characteristics of the team s and leaders in the preferences for subordinates (Bauer & Green , 1996). Future research sh ould directly examine

PAGE 164

164 whether affective (e.g., mutual liking) and social structural (e.g., interpersonal connections) factors could eliminate the effects of team composition in learning ability and conscientiousness on team leadership.

PAGE 165

165 Table 6 1. Means, standard deviations, and bivariate correlations among study variables . Variable M SD 1 2 3 4 5 6 7 8 9 10 11 12 1. Leader education 2.28 .72 2. Leader managerial experience (year) 1.88 2.11 .12 3. Team size 4.95 1.56 .10 .12 4. Average time working with supervisor 21.87 16.39 .29* .16 .12 5. SD of time working with supervisor 10.84 8.44 .12 .1 7 .09 .70** 6. Gender diversity .23 .20 .27* .08 .11 .04 .07 7. Average education 2.28 .58 .69** .15 .14 .40** .19 .25 8. SD of education .36 .29 .04 .01 .05 .01 .09 .01 .02 9. Average skill variety 3.57 .50 .35 ** .11 .02 .13 .17 .07 .44 ** .05 10. Average task complexity 3.57 .49 .39** .02 .09 .14 .06 .07 .26 * .06 .03 11. Average problem solving 3.63 .38 .23 .09 .06 .05 .08 .02 .34 ** .02 .45 ** .04 12. Average learning ability 3.89 .34 . 48* .00 .17 .14 .02 .18 .52 ** .08 .33 * .08 .22 13. Average conscientiousness 6.70 .59 .27* .04 .10 .01 .06 .20 .41 ** .01 .47 ** .04 .26 .30 * 14. Differentiation in skill variety .62 .29 .05 .05 .02 .05 .11 .23 .21 .17 .30 * .01 .03 .25 15 . Differentiation in task complexity .67 .30 .25 .16 .08 .09 .42** .33 * .12 .09 .29 * .26 * .10 .07 16. Differentiation in problem solving .60 .28 .05 .24 .01 .04 .12 .08 .05 .06 .05 .03 .31 * .02 17. Workflow network centralization .05 .12 .02 .0 5 .04 .03 .06 .01 .07 .12 .08 .04 .11 .11 18. Differentiation in learning ability .53 .21 .04 .10 .11 .04 .06 .22 .19 .13 .13 .15 .07 .27 * 19. Differentiation in conscientiousness .99 .37 .37** .14 .07 .16 .17 .31 * .48 ** .19 .03 .27 .0 5 .20 20. Task interdependence .16 .13 .22 .13 .17 .06 .03 .05 .26 .04 .18 .18 .03 .29 * 21. Average information giving 4.23 .53 .21 .06 .03 .09 .10 .29 * .06 .11 .06 .22 .26 .02 22. Average autonomy giving 4.17 .68 .40** .13 .13 .25 .03 .22 .27 * .14 .12 .25 .13 .23 23. Average socioemotional support 3.88 .64 .06 .06 .14 .22 .43** .01 .00 .18 .13 .08 .16 .04 24. Differentiated information giving .43 .31 .07 .16 .10 .09 .06 .00 .09 .13 .07 .18 .27 * .05 25. Differentiated autonomy givi ng .50 .35 .12 .09 .31* .05 .06 .20 .15 .14 .10 .05 .16 .17 26. Differentiated socioemotional support .39 .27 .11 .10 .32* .11 .09 .09 .10 .08 .20 .06 .03 .08 27. Team performance 3.83 1.10 .31* .11 .17 .34** .22 .25 .41 ** .02 .12 .29 * .13 .02 Note. N = 58. * p < .05, ** p < .01, two tail tests.

PAGE 166

166 Table 6 1. Continued. Variable 13 14 15 16 17 18 19 20 21 22 23 24 25 26 2 7 1. Leader education 2. Leader managerial experience 3. Team size 4. Average time working with supervisor 5. SD of time working with supervisor 6. Gender diversity 7. Average education 8. SD of education 9. Average skill variety 10. Average task complexity 11. Average problem solving 12. Average learning ability 1 3. Average conscientiousness 14. Differentiation in skill variety .23 15. Differentiation in task complexity .29* .04 16. Differentiation in problem solving .04 .31* .02 17. Workflow network ce ntralization .13 .06 .00 .06 18. Differentiation in learning ability .06 .20 .09 .05 .16 19. Differentiation in conscientiousness .16 .04 .07 .09 .00 .09 20. Task interdependence .10 .30* .08 .07 .57** .27* .17 21. Average information giving .02 .21 .03 .24 .14 .13 .04 .05 22. Average autonomy giving .08 .33* .00 .28* .09 .05 .04 .08 .61** 23. Average socioemotional support .06 .05 .22 .05 .28* .09 .08 .09 .42** .22 24. Differentiated information giving .23 .24 .12 .05 .09 .03 .02 .06 .21 .07 .18 25. Differentiated autonomy giving .18 .07 .26 .08 .03 .12 .30* .13 .14 .16 .04 .21 26. Differentiated socioemotional support .33* .16 .18 .0 7 .04 .04 .20 .02 .05 .06 .14 .54** .49** 27. Team performance .17 .20 .03 .04 .08 .05 .21 .09 .03 .15 .06 .06 .03 .11

PAGE 167

167 Table 6 2. Hierarchical multiple r egression r esults. Average information giving Average autonomy giving Predict ors Step 1 Step 2 Step 3 Step 1 Step 2 Step 3 Step 1: Covariates Intercept 4.23** 4.23** 4.22** 4.17** 4.17** 4.20** Leader education .16 .19 .19 .30* .38* .29* Leader managerial experience .02 .02 .04 .06 .06 .10** Team size .00 .00 .01 .02 .02 .02 Average time working with supervisor .00 .00 .00 .01 .01 .02** SD of time working with supervisor .01 .01 .00 .01 .02 .03* Gender diversity .91** .94* .82* .48 .74 .99* Average education .20 .12 .26 .15 .10 .22 SD of education .08 .18 .49* .34 .11 .27 Average skill variety .21 .33* .27 .01 .09 .08 Average task complexity .21 .26 .20 .15 .30 .28 Average problem solving .60** .65** .61** .18 .23 .01 Average learning ability .16 .20 .05 .07 .05 .55 Average conscientiousness .03 .05 .00 .04 .01 .07 Step 2: Main effects Differentiation in skill variety .66* .44 .72* .45 Differentiation in task complexity .05 .31 .62 1.04** Differentiation in prob lem solving .04 .09 .00 .14 Workflow network centralization .22 1.02 .07 1.28 Differentiation in learning ability .51 .11 .42 .03 Differentiation in conscientiousness .24 .18 .43 .41 Task interdependence (T) .01 .31 .03 .92 Step 3: Interaction Differentiation in skill variety × T 7.14** 7.12** Differentiation in task complexity × T 2.61 3.18 Differentiation in problem solving × T 2.64 1.31 Workflow network centralization× T 4.22 8.29 Differentiation in learning ability × T 2.86 2.19 Differentiation in conscientiousness × T .37 4.74* R 2 .32 .46 .66 .29 .48 .70 F 1.58 1.61 2.31* 1.37 1.68 ( p = .08) 2.80** R 2 .15 .20 .19 .23 F 1.45 2.97* 1.89 ( p = .10) 3.90** Note. N = 58. * p < .05, ** p < .01, one tail tests.

PAGE 168

168 Table 6 2. Continued Average socioemotional support Differentiated information giving Predictors Step 1 Step 2 Step 3 Step 1 Step 2 Step 3 Step 1: Covariates Intercept 3 .88** 3.88** 3.87** .43** .43** .51** Leader education .12 .18 .18 .03 .03 .06 Leader managerial experience .01 .00 .01 .02 .02 .02 Team size .05 .02 .00 .03 .04 .03 Average time working with supervisor .01 .01 .01 .00 .00 .00 SD of time working with supervisor .04** .05** .04** .01 .00 .01 Gender diversity .31 .63 .53 .17 .20 .25 Average education .11 .16 .09 .01 .06 .00 SD of education .56* .47* .25 .12 .08 .06 Average skill variety .18 .24 .24 . 07 .04 .15 Average task complexity .21 .22 .20 .13 .11 .08 Average problem solving .60** .54* .55* .33** .38** .49** Average learning ability .03 .22 .15 .02 .04 .09 Average conscientiousness .01 .06 .09 .18* .19* .16* Step 2 : Main effects Differentiation in skill variety .06 .22 .14 .20 Differentiation in task complexity .02 .17 .08 .03 Differentiation in problem solving .17 .08 .09 .16 Workflow network centralization 2.07* 1.43 .66 .74 Differentiation in learning ability .39 .17 .09 .08 Differentiation in conscientiousness .41 .41 .07 .01 Task interdependence (T) 1.30 1.21 .29 .20 Step 3: Interaction Differentiation in skill variety × T 5.68* 3.0 5* Differentiation in task complexity × T 2.12 1.98 Differentiation in problem solving × T 3.61 3.64** Workflow network centralization× T 1.07 9.26** Differentiation in learning ability × T 2.89 3.34* Differentiation in conscientiousness × T .19 .08 R 2 .37 .49 .58 .31 .38 .57 F 2.02* 1.81 ( p = .06) 1.61 ( p = .10) 1.53 1.12 1.60 ( p = .10) R 2 .12 .08 .07 .20 F 1.26 .97 .56 2.36 ( p = .05)

PAGE 169

169 Table 6 2. Continued. Differentiated autonomy giving Differentiated socioemotional support Predictors Step 1 Step 2 Step 3 Step 1 Step 2 Step 3 Step 1: Covariates Intercept .50** .5 0** .57** .39** .39** .40** Leader education .03 .02 .00 .05 .06 .02 Leader managerial experience .00 .01 .01 .01 .01 .00 Team size .07* .07* .05 .06** .05* .06* Average time working with supervisor .00 .00 .00 .00 .00 .00 SD of time working with supervisor .00 .01 .00 .00 .00 .00 Gender diversity .35 .24 .03 .04 .11 .11 Average education .14 .15 .11 .04 .02 .08 SD of education .11 .03 .07 .05 .04 .16 Average skill variety .03 .03 .02 .07 .05 .10 Aver age task complexity .02 .02 .07 .04 .09 .07 Average problem solving .10 .11 .09 .15 .17 .20 Average learning ability .35* .31 .36 .17 .20 .14 Average conscientiousness .09 .06 .02 .16** .14* .14* Step 2: Main effects Differentiation in skill variety .04 .02 .12 .13 Differentiation in task complexity .30 .29 .17 .14 Differentiation in problem solving .06 .01 .03 .03 Workflow network centralization .24 1.00 .26 .46 Differentiation in le arning ability .17 .01 .05 .16 Differentiation in conscientiousness .15 .21 .10 .09 Task interdependence (T) .04 .11 .03 .04 Step 3: Interaction Differentiation in skill variety × T 2.21 .01 Differentiation in task complexity × T 2.11 1.60 Differentiation in problem solving × T .59 .43 Workflow network centralization× T 6.36 .62 Differentiation in learning ability × T .56 1.15 Differentiation in conscientiousness × T 1.83 .06 R 2 .30 .38 .47 .32 .37 .44 F 1.44 1.12 1.04 1.62 1.10 .94 R 2 .08 .09 .05 .07 F .67 .84 .41 .64

PAGE 170

170 Table 6 2. Continued. Team performance Predictors Step 1 Step 2 Step 3 Step 1: Covariates Intercept 3.83** 3.83** 3.65** Leader education .05 .14 .18 Leader managerial experience .00 .00 .04 Team size .06 .07 .07 Average time working with supervisor .01 .01 .01 SD of time working with supervisor .00 .01 .01 Gender diversity .78 .84 1.04 Average educ ation .75* .70 1.03* SD of education .01 .22 .29 Average skill variety .31 .31 .17 Average task complexity .32 .40 .49 Average problem solving .77* .80* .55 Average learning ability 1.06* 1.17* 1.86** Average conscientiousness .1 4 .19 .19 Step 2: Main effects Differentiation in skill variety .50 .78 Differentiation in task complexity .56 .03 Differentiation in problem solving .12 .38 Workflow network centralization .10 5.24* Differentiation in learning ability .10 .59 Differentiation in conscientiousness .15 .22 Task interdependence (T) .44 1.08 Step 3: Interaction Differentiation in skill variety × T 7.97* Differentiation in task complexity × T 10.33* Differentiation in problem solving × T 3.19 Workflow network centralization× T 26.04* Differentiation in learning ability × T 4.56 Differentiation in conscientiousness × T 3.34 R 2 .40 .44 .63 F 2.26* 1.45 2.01* R 2 .04 .19 F .37 2.61*

PAGE 171

171 Fig ure 6 1. Task interdependence moderates the relationship between differentiation in skill variety and average information giving. Figure 6 2. Task interdependence moderates the relatio nship between differentiation in skill variety and average autonomy g iving.

PAGE 172

172 Figure 6 3. Task interdependence moderates the relationship between dif ferentiation in skill variety and average socioemotional support. Figure 6 4. Task interdependence moderates the relatio nship between differentiation in conscientiousness an d average autonomy giving.

PAGE 173

173 Figure 6 5. Task interdependence moderates the relatio nship between differentiation in skil l variety and differentiated information giving. Figure 6 6. Task interdependence moderates the relationship between differentiation in problem solving and differentiated information giving.

PAGE 174

174 Figure 6 7. Task interdependence moderates the relationship between workflow network centralization and differentiated information giving. Figure 6 8. Task interdependence moderates the relatio nship between differentiation in learning ability and differentiated information giving.

PAGE 175

175 Figure 6 9. Task interdependence moderates the relatio nship between differentiation in skill variety and team performance. Figure 6 10. Task interdependence moder ates the relationship between differentiation in task complexity and team performance.

PAGE 176

176 Figure 6 11. Task interdependence moderates the relationship between workflow network centralization and team performance .

PAGE 177

177 CHAPTER 7 GENERAL DISCUSSION Summary of Ke y Findings This dissertation includes three studies that together proposed and tested a computational model about team leadership. Study 1 specified a computational model that formally represents the dynamic model. Qualitative model evaluation results sug gested that the model behaved consistently with relevant empirical findings reported in the literature. Specifically, computational modeling results suggested that magnitude and duration of negative disturbances were positively related to team leadership b ehaviors, which were consistent with research on the effects of unexpected negative events on team leader intervention (Morgeson, 2005; Morgeson & DeRue, 2006). Consistent with research on team leader individual differences and leadership styles (e.g., Che n et al., 2007) , computational modeling results suggested that leader reactivity was positively related to team leadership behavior s and team performance, and the positive relationship was stronger when team task structure was functional (vs. divisional an d disjunctive; i.e., high task interdependence). In addition, computational modeling results suggested the differentiation in team member learning ability was positive ly related to differentiated individual focused leadership behaviors and differentiation in individual expectancy. When team task structure wa s functional, differentiation in team member learning ability was negatively related to team performance whereas when team task structure was disjunctive, differentiation in team member learning ability was positively related to team performance. Study 2 conducted a lab experiment to empirically test a series of hypotheses derived theoretical model, Study 2 found (1) at the within task level, leader spent more time regulating a

PAGE 178

178 larger, and this relationship was stronger when team task structure was functional (vs. divisional and disjunctive); and (2) at the task level, differentiation in learning ability across members , differentiation in task difficulties across positions , and team task structure jointly influence d the differentiation in leader time allocation across subordinates . Specifically, when task difficulties were different (vs. same) across the positions, more time was allocated to the more difficult position and differentiation in leader time allocated among subordinates was larger. When team members had different (vs. s ame) learning abilities, more time was allocated to the subordinate with lower learning ability and differentiation in time allocation among subordinates was larger. In addition, the negative relationship between learning ability of a position and time spe nt by the leader was stronger when disturbance magnitude was smaller. The positive relationship between differentiation in learning ability and differentiation in leader time allocation was stronger when disturbance magnitude was smaller or when team task structure was functional (vs. divisional and disjunctive). Study 3 collected data from a field sample and tested how the team task structure (operationalized as task interdependence) interacted with knowledge characteristics of the positions and member di fferences in learning ability and conscientiousness to influenc e average and differentiated individual focused leadership. Study 3 found that task interdependence moderated t he effects of differentiation in skill variety across positions and differentiatio n in conscientiousness among members on average individual focused leadership behavior, such that the positive relationships were stronger when task interdependence was higher. This study also found that task interdependence moderated t he effects of differ entiation in skill variety, workflow network ce ntralization, differentiation in problem solving, and differentiation in learning ability on differentiat ed individual focused leadership. The positive relationships

PAGE 179

179 between differentiation in problem solving and differe ntiation in learning ability, and differentiated individual focused leadership were stronger when task interdependence was higher, which were consistent with Study 1 and Study 2 findings and supported the theoretical model. However, the study al so found that the positive relation ships between differentiation in skill variety and workflow network centra lization, and differentiated individual focused leadership were weaker when task interdependence was higher. In addition, it was found that task in terdependence moderated the relation ships between differentiation in skill variety, workflow network centra lization, and differentiation in task complexity, and team performance. The negative relation ships between differentiation in skill variety and task complexity, and team performance were stronger when task interdependence was higher, which supported the theoretical model. However, the study also found that the negative relationship between workflow network centralization and team performance was weaker when task interdependence was higher. Theoretical Implications First, the current research provides a coherent theoretical model on how team leadership behaviors change over time. The current model advances team leadership research by formally conceptuali zing team leadership as part of the team performance system, which provides a mechanism for how changes in team leadership behaviors unfold over time and how the changes in team leadership are interwoven with team performance. The present research suggeste d that team leadership behavior fluctuated over time, and the fluctuation is driven by relative environment, and task structure. Team leadership has rarely been ex amined as a dynamic phenomenon although qualitative research started to document the changes in team leadership behaviors over time within a performance episode (Klein et al., 2006). The present research

PAGE 180

180 explicitly addresses the dynamic issues in team lead ership, and provides an explanation for the dynamics observed. In addition, the present research helps integrate existing team leadership theories. A number of theories have emerged that examine the team leadership phenomenon from different perspectives. H owever, the existing theories are limited in understanding the dynamic issues involved . This limitation is partly due to the lack of a coherent theoretical framework to understand the interactions between subordinates and leaders. The present research prop oses a model that conceptualizes leadership as a regulatory agent that follows the discrepancy reduction rule in facilitating the team to strive toward team goal. This way the current model helps integrate existing team leadership theories by incorporating key motivational and leadership concepts into the same theoretical framework. Second, this research integrates two major theoretical perspectives in team, motivation, and leadership research (i.e., functional leadership theory and self regulation theory) . Through three studies, the integrated framework demonstrates that team leadership serves a regulatory function in organizational units, which follows the principle of discrepancy reduction. Based on this mechanism, the integrated model shows how team lea dership behaviors , as part of team regulatory processes , drive team s toward the collective goal . Through a lab experiment, the present research supports the basic assumption of self regulation theory and functional theory, latory efforts is driven by relative discrepancy between alternative choices of actions. In addition, this research suggested that team leadership regulates team states and processes in reaction to external disturbances in team performance environment. Thi s finding provides support to the functional view on leadership and extends previous theories

PAGE 181

181 by showin g that team leadership is under the joint influence of and characteristics of the task situation. Third, this research helps br idge the gaps between team leadership and team design literatures . The current model incorporates team composition and team design factors into a team leadership model (Hollenbe ck et al., 2004). Simulation and empirical studies in this research suggested t hat who are selected and composed into a team and how the team is designed can significantly affect team leadership behaviors and team performance. Through i ntegrat ing team and leadership literatures, this research also helps resolve a long existing debate in the l iterature regarding the beneficial versus detrimental effects of differentiated leadership on team performance (e.g., Graen & Uhl Bien, 1995; Liden et al. , 2006; Wu et al., 2010). The research suggests that differentiated leadership may have diffe rent effects on team performance depending on the specific team task structure . In addition, empirical f indings from Study 2 and Study 3 consistently suggested that differentiat ed le adership behavior is influenced by team task structure and team position a nd member characteristics. Furthermore, this research demonstrates that the structure and function of organizational units together with their leaders can be studied from a system based perspective and explained by a coherent set of mechanisms (Katz & Kah n, 1978). Through three studies, this dissertation demonstrates that leaders can strategically allocate resources and regulate team functioning under different demands and capacities of the team. F rom a broader perspective of organizational research, study ing the dynamic interaction between leadership and team in a given task environment help s extend our understanding of organizations as complex dynamic systems. According to the general systems theory, units in organization can be viewed as subsystems that are subject to the influence from higher level subsystems (e.g., leaders) and the task environment

PAGE 182

182 (Boulding, 1956; Katz & Kahn, 1978). However, few studies have directly examined how leaders dynamically interact with and regulate a collective set of indiv iduals while they work on a common goal, or how the interaction among multiple subsystems enacted by leader and followers is influenced by the team task environment (Kozlowski & Bell, 200 3 ; Kozlowski & Klein, 2000). The current research provide s so me insig hts on these questions about dynamic multisystem interaction s in organizations. Practical Implications This research provides some implications for human resource management practices. First, this research suggested that team leaders change their behavior s over time in reaction to selected into the team and how team tasks are designed can influence whom team leaders choose as the target of their regulation actions and the amount of efforts leaders spend on the regulatory actions. Therefore, in order to change the behaviors of team leaders, organizations should pay attention to how the team is designed and composed. Second, organizations should prepare and develop the c ompetencies of formal team leaders or team members in self managing teams, given the functional and dynamic nature of team leadership. The present research suggests that team leaders face the challenge of making dynamic decisions over time when team member s are different from each other and assume different duties, team positions are highly interdependent, and team task environment is dynamic. These challenges are not uncommon in modern workplaces. In order to ensure team leadership works effectively, organ izations should prepare individuals assigned to team leadership functions. Third, this research suggests that individual differences of team leaders, such as leader reactivity, leader sensitivity to relative discrepancy, and team effectiveness. These findings suggest that selection and training procedures should be

PAGE 183

183 taken to help individuals in leadership positions become more attentive and reactive to team memb regulated systems (e.g., on regulatory actions. Limitations and Future Research Directions This dis sertation research used a mixture of research methodologies, including computational modeling, lab experiment, and empirical survey research. Computational modeling allowed the research to precisely communicate the model and integrate previous theoretical and empirical studies. Lab experiment allows this research to test the theoretical model in a controlled environment and help establish causal relationships. The empirical survey Findings from these three studies using different methodologies generally supported the theoretical model. Despite this strength of the current research, it also had several limitations that can be addressed in future research. First, t he computational modeling study could not examine whether the theoretical model behaved consistently with repeated observations of team leadership behaviors over time. This was because there was no quantitative study that has documented team leadership beh avior using longitudinal research design. The lab study of the current research contributed to testing a dynamic theory on team leadership. However, it is not clear wh ether team leaders behave in similar ways in real world organizations. Therefore, more quantitative field studies are needed to document the changes of team leadership behaviors either within a single performance episode or across multiple performance epis odes.

PAGE 184

184 Second, the current model assumes that interactions among subordinates is relatively stable and thus does not specify the interactions among subordinates in team goal pursuit process. As Marks et al. (2001) suggested, interpersonal processes are in volved in both transition and action processes. Typical member member interactions include motivating each other, affect management, and conflict management. These member member interactions can also be conceptualized as other regulation processes. For exa mple, team members may perceive the coaching perceive a discrepancy between cow further developing the current theoretical model is to introduce variables and processes into the model to represent interactions among peer subordinates over time. Third, both empirical and theoretical research is needed to further understand team leadership over multiple team performance episodes. Although the current theoretical model include s components that represent how leaders regulate teams over multiple team performance episodes, the current research did not empirically examine team leadership over multiple team performance episodes. The current lab study simulation platform did not allow interactions between the team performance episodes . Future research should develop lab study plat forms that allow scenarios of multiple team performance episodes. From the theoretical standpoint, future research can extend the current model to account for multiple goal pursuit scenarios. Work groups and teams often work on multiple tasks simultaneousl y. For example, professional service teams might serve multiple clients at the same time. It is also conceivable that action teams (e.g., flight crew) strive toward goals with different priorities at the same time (e.g., safety, customer

PAGE 185

185 satisfaction). It is important to understand how team leadership regulates the team when the team is striving toward multiple goals simultaneously. Furthermore, future research is needed to understand how to organize the regulatory duties properly in the teams. Previous research suggested that shared (or distributive, collective) leadership structure can be beneficial for both team performance and team member job attitudes, particularly when the work is more complex (Wang, Waldman, & Zhang, 2014). However, research on mul ti team system, i.e., complicated team task system, suggests that team performs better when strategizing and coordination, two critical functions of team processes, are centralized rather than distributed in the multi team system (DeChurch & Marks, 2006; L anaj, Hollenbeck, Ilgen, Barnes, & Harmon, 2013). To resolve these conflicting findings, it is important to clarify the underlying mechanisms of the relationship between team leadership and team performance. Taking the regulatory perspective might be helpf ul. It is possible that leadership to be an effective leadership structure within certain complexity of team task. Beyond certain complexity level of the performance system, it is necessary to have subsystems be responsible for regulatory functions. In the extreme case, organizations as a large performance system often have specialized managing (or regulating) units, i.e., the management team (Katz & Kahn, 1978). Futur e research can integrate the current model with distributed leadership theories to further clarify this issue.

PAGE 186

186 APPENDIX A CODES FOR SPECIFYING THE COMPUTATIONAL MODEL IN SOFTWARE VENSIM PROFESSIONAL Mathematical e quations used to define the constructs are listed below the figures in Chapter 3. The following codes can directly be used to program a team/ system with four subordinates using the software Vensim Professional. The codes can be adapted to represent systems with 2, 3, 5 or more below represent subordinate 1, 2, 3, and 4. accepted deadline A= deadline accepted deadline B= deadline accepted deadline C= deadline accepted deadline D= deadline clear task state A=IF THEN ELSE(task number output=1, 0 task st ate A , 0 ) clear task state B=IF THEN ELSE(task number output=1, 0 task state B , 0 ) clear task state C=IF THEN ELSE(task number output=1, 0 task state C , 0 ) clear task state D=IF THEN ELSE(task number output=1, 0 task state D , 0 ) clear team task sta te=IF THEN ELSE(task number output=1, 0 team task state , 0 ) clear time spent on task team=IF THEN ELSE(task number output=1, 0 time spent on task team, 0 ) competency A= INTEG (learning ability A*(learning A+developmental leadership behavior A)+competen cy disturbance A*competency A, initial competency A) competency B= INTEG (learning ability B*(learning B+developmental leadership behavior B)+competency disturbance B*competency B, initial competency B) competency bias A=0 competency bias B=0 competency bi as C=0 competency bias D=0 competency C= INTEG (learning ability C*(learning C+developmental leadership behavior C)+competency disturbance C*competency C, initial competency C)

PAGE 187

187 competency comparator A=IF THEN ELSE(desired competency A competency perception A>0, desired competency A competency perception A , 0) competency comparator B=IF THEN ELSE(desired competency B competency perception B>0, desired competency B competency perception B , 0) competency comparator C=IF THEN ELSE(desired competency C compete ncy perception C>0, desired competency C competency perception C , 0) competency comparator D=IF THEN ELSE(desired competency D competency perception D>0, desired competency D competency perception D , 0) competency D= INTEG (learning ability D*(learning D +developmental leadership behavior D)+competency disturbance D*competency D, initial competency D) competency disturbance A=0 competency disturbance B=0 competency disturbance C=0 competency disturbance D=0 competency perception A=competency A+competency b ias A competency perception B=competency B+competency bias B competency perception C=competency C+competency bias C competency perception D=competency D+competency bias D conscientiousness A=1 conscientiousness B=1 conscientiousness C=1 conscientiousness D =1 deadlines((1,30)) desired competency A=0.1*skill variety A desired competency B=0.1*skill variety B desired competency C=0.1*skill variety C desired competency D=0.1*skill variety D

PAGE 188

188 developmental leadership behavior A=IF THEN ELSE(regulation target=5,le ader reactivity*position importance A*leader A LMX*leader competency comparator A/0.1/4, IF THEN ELSE(regulation target=1,leader reactivity*position importance A*leader A LMX* leader competency comparator A/0.1,0)) developmental leadership behavior B=IF TH EN ELSE(regulation target=5,leader reactivity*position importance B*leader B LMX*leader competency comparator B/0.1/4, IF THEN ELSE(regulation target=2,leader reactivity*position importance B*leader B LMX*leader competency comparator B/0.1,0)) developmenta l leadership behavior C=IF THEN ELSE(regulation target=5,leader reactivity*position importance C*leader C LMX*leader competency comparator C/0.1/4, IF THEN ELSE(regulation target=3,leader reactivity*position importance C*leader C LMX*leader competency comp arator C/0.1,0)) developmental leadership behavior D=IF THEN ELSE(regulation target=5,leader reactivity*position importance D*leader D LMX*leader competency comparator D/0.1/4, IF THEN ELSE(regulation target=4,leader reactivity*position importance D*leader D LMX*leader competency comparator D/0.1,0)) expectancy comparator A=max((accepted deadline A Time expectancy input A),0) expectancy comparator B=max((accepted deadline B Time expectancy input B),0) expectancy comparator C=max((accepted deadline C Time ex pectancy input C),0) expectancy comparator D=max((accepted deadline D Time expectancy input D),0) expectancy input A=task state comparator A/competency perception A expectancy input B=task state comparator B/competency perception B expectancy input C=task state comparator C/competency perception C expectancy input D=task state comparator D/competency perception D expectancy output A=expectancy comparator A expectancy output B=expectancy comparator B expectancy output C=expectancy comparator C expectancy out put D=expectancy comparator D FINAL TIME = 30 individual goal A="leader assigned individual goal A"+individual goal bias A individual goal B="leader assigned individual goal B"+individual goal bias B

PAGE 189

189 individual goal bias A=0 individual goal bias B=0 indiv idual goal bias C=0 individual goal bias D=0 individual goal C="leader assigned individual goal C"+individual goal bias C individual goal D="leader assigned individual goal D"+individual goal bias D initial competency A=5e 005 initial competency B=5e 005 i nitial competency C=5e 005 initial competency D=5e 005 INITIAL TIME = 0 job complexity A=1 job complexity B=1 job complexity C=1 job complexity D=1 k1=0 k2=0 leader A LMX=INTEG (k1*(developmental leadership behavior A+"task contingent leadership behavior A")+k2*(competency A+task state A),"leader A similarity") leader B LMX=INTEG (k1*(developmental leadership behavior B+"task contingent leadership behavior B")+k2*(competency B+task state B),"leader B similarity") leader C LMX=INTEG (k1*(developmental leade rship behavior C+"task contingent leadership behavior C")+k2*(competency C+task state C),"leader C similarity") leader D LMX=INTEG (k1*(developmental leadership behavior D+"task contingent leadership behavior D")+k2*(competency D+task state D),"leader D si milarity") leader A similarity=1 leader B similarity=1

PAGE 190

190 leader C similarity=1 leader D similarity=1 leader competency bias A=0 leader competency bias B=0 leader competency bias C=0 leader competency bias D=0 leader competency comparator A=IF THEN ELSE(desir ed competency A leader perceived competency A>0, desired competency A leader perceived competency A, 0 ) leader competency comparator B=IF THEN ELSE(desired competency B leader perceived competency B>0, desired competency B leader perceived competency B, 0 ) leader competency comparator C=IF THEN ELSE(desired competency C leader perceived competency C>0, desired competency C leader perceived competency C, 0 ) leader competency comparator D=IF THEN ELSE(desired competency D leader perceived competency D>0, d esired competency D leader perceived competency D, 0 ) leader perceived competency A=(competency A+leader competency bias A) leader perceived competency B=(competency B+leader competency bias B) leader perceived competency C=(competency C+leader competency bias C) leader perceived competency D=(competency D+leader competency bias D) leader perceived task state A=(task state A+leader task state bias A) leader perceived task state B=(task state B+leader task state bias B) leader perceived task state C=(task s tate C+leader task state bias C) leader perceived task state D=(task state D+leader task state bias D) leader perceived team competency=IF THEN ELSE(team task structure=2,min(min(competency A, competency B),min(competency C, competency D)) , IF THEN ELSE(t eam task structure=3, max(max(competency A, competency B),max(competency C, competency D)) , (competency A+competency B+competency C+competency D)/4)) leader perceived team expectancy=max((task deadline Time leader team task state comparator/leader perceiv ed team competency),1e 006) leader perceived team task state=team task state+leader team task state bias

PAGE 191

191 leader reactivity=1 leader sensitivity to relative discrepancy=0.01 leader task choice A=position importance A*leader reactivity*leader task state outp ut A*(1/leader perceived team expectancy) leader task choice B=position importance B*leader reactivity*leader task state output B*(1/leader perceived team expectancy) leader task choice C=position importance C*leader reactivity*leader task state output C*( 1/leader perceived team expectancy) leader task choice D=position importance D*leader reactivity*leader task state output D*(1/leader perceived team expectancy) leader task state bias A=0 leader task state bias B=0 leader task state bias C=0 leader task st ate bias D=0 leader task state comparator A=IF THEN ELSE("leader assigned individual goal A" leader perceived task state A>0,"leader assigned individual goal A" leader perceived task state A,0) leader task state comparator B=IF THEN ELSE("leader assigned i ndividual goal B" leader perceived task state B>0,"leader assigned individual goal B" leader perceived task state B,0) leader task state comparator C=IF THEN ELSE("leader assigned individual goal C" leader perceived task state C>0,"leader assigned individu al goal C" leader perceived task state C,0) leader task state comparator D=IF THEN ELSE("leader assigned individual goal D" leader perceived task state D>0,"leader assigned individual goal D" leader perceived task state D,0) leader task state output A=lead er task state comparator A leader task state output B=leader task state comparator B leader task state output C=leader task state comparator C leader task state output D=leader task state comparator D leader team task state bias=0 leader team task state co mparator=IF THEN ELSE("leader assigned team goal" leader perceived team task state>0, "leader assigned team goal" leader perceived team task state, 0 )

PAGE 192

192 leader team task state output=leader team task state comparator "leader assigned individual goal A"="lea der assigned team goal" "leader assigned individual goal B"="leader assigned team goal" "leader assigned individual goal C"="leader assigned team goal" "leader assigned individual goal D"="leader assigned team goal" "leader assigned team goal"=team goal le arning A=conscientiousness A*competency comparator A/0.1 learning ability A=0.01 learning ability B=0.01 learning ability C=0.01 learning ability D=0.01 learning B=conscientiousness B*competency comparator B/0.1 learning C=conscientiousness C*competency co mparator C/0.1 learning D=conscientiousness D*competency comparator D/0.1 position importance A=workflow centrality A position importance B=workflow centrality B position importance C=workflow centrality C position importance D=workflow centrality D proble m solving A=1 problem solving B=1 problem solving C=1 problem solving D=1 regulation target=IF THEN ELSE(leader task choice A leader task choice B<=leader sensitivity to relative discrepancy:AND:leader task choice A leader task choice C<=leader sensitivity to relative discrepancy:AND:leader task choice C leader task choice D<=leader sensitivity to relative discrepancy,5,IF THEN ELSE(team task structure=2:OR:team task structure=1, IF THEN ELSE(leader task choice A leader task choice B>=leader sensitivity to relative

PAGE 193

193 discrepancy:AND:leader task choice A leader task choice C>=leader sensitivity to relative discrepancy:AND:leader task choice A leader task choice D>=leader sensitivity to relative discrepancy,1,IF THEN ELSE(leader task choice B leader task choice C>=leader sensitivity to relative discrepancy:AND:leader task choice B leader task choice D>=leader sensitivity to relative discrepancy,2,IF THEN ELSE(leader task choice C leader task choice D>=leader sensitivity to relative discrepancy,3,4))),IF THEN ELSE (leader task choice A leader task choice B<=leader sensitivity to relative discrepancy:AND:leader task choice A leader task choice C<=leader sensitivity to relative discrepancy:AND:leader task choice A leader task choice D<=leader sensitivity to relative d iscrepancy,1,IF THEN ELSE(leader task choice B leader task choice C<=leader sensitivity to relative discrepancy:AND:leader task choice B leader task choice D<=leader sensitivity to relative discrepancy,2,IF THEN ELSE(leader task choice C leader task choice D<=leader sensitivity to relative discrepancy,3,4))) )) skill variety A=1 skill variety B=1 skill variety C=1 skill variety D=1 task choice comparator A=IF THEN ELSE(task choice input A "utility of off task activities A">0,task choice input A "utility of off task activities A",0) task choice comparator B=IF THEN ELSE(task choice input B "utility of off task activities B">0,task choice input B "utility of off task activities B",0) task choice comparator C=IF THEN ELSE(task choice input C "utility of off tas k activities C">0,task choice input C "utility of off task activities C",0) task choice comparator D=IF THEN ELSE(task choice input D "utility of off tasks activities D">0,task choice input D "utility of off tasks activities D",0) task choice input A=IF TH EN ELSE(expectancy output A>0,1,0)*IF THEN ELSE(team expectancy output A>0,1,0)*conscientiousness A*task state output A task choice input B=IF THEN ELSE(expectancy output B>0,1,0)*IF THEN ELSE(team expectancy output B>0,1,0)*conscientiousness B*task state output B task choice input C=IF THEN ELSE(expectancy output C>0,1,0)*IF THEN ELSE(team expectancy output C>0,1,0)*conscientiousness C*task state output C task choice input D=IF THEN ELSE(expectancy output D>0,1,0)*IF THEN ELSE(team expectancy output D>0,1, 0)*conscientiousness D*task state output D task choice output A=task choice comparator A task choice output B=task choice comparator B

PAGE 194

194 task choice output C=task choice comparator C task choice output D=task choice comparator D task deadline=deadlines(task number) task difficulty A=(job complexity A+problem solving A)/2 task difficulty B=(job complexity B+problem solving B)/2 task difficulty C=(job complexity C+problem solving C)/2 task difficulty D=(job complexity D+problem solving D)/2 task disturbance A=( 0.1)*PULSE(11,5)+( 0.1)*PULSE(21,5) task disturbance B=( 0.1)*PULSE(11,5)+( 0.1)*PULSE(21,5) task disturbance C=( 0.1)*PULSE(11,5)+( 0.1)*PULSE(21,5) task disturbance D=( 0.1)*PULSE(11,5)+( 0.1)*PULSE(21,5) task number= INTEG (task number output,1) task n umber comparator=IF THEN ELSE(total number of task task number input>0, total number of task task number input , 0 ) task number input=task number task number output=IF THEN ELSE(task number comparator>0:AND:(Time>=task deadline:OR:team task gain=1),1,0) t ask state A= INTEG (task disturbance A*task state A+(task choice output A+"task contingent leadership behavior A")*competency A*task difficulty A+clear task state A,0) task state B= INTEG (task disturbance B*task state B+(task choice output B+"task contin gent leadership behavior B")*competency B*task difficulty B+clear task state B,0) task state bias A=0 task state bias B=0 task state bias C=0 task state bias D=0 task state C= INTEG (task disturbance C*task state C+(task choice output C+"task contingent le adership behavior C")*competency C*task difficulty C+clear task state C,0)

PAGE 195

195 task state comparator A=IF THEN ELSE(individual goal A task state perception A>0, individual goal A task state perception A, 0) task state comparator B=IF THEN ELSE(individual goal B task state perception B>0, individual goal B task state perception B , 0) task state comparator C=IF THEN ELSE(individual goal C task state perception C>0, individual goal C task state perception C , 0) task state comparator D=IF THEN ELSE(individual goa l D task state perception D>0, individual goal D task state perception D , 0) task state D= INTEG (task disturbance D*task state D+(task choice output D+"task contingent leadership behavior D")*competency D*task difficulty D+clear task state D,0) task stat e output A=IF THEN ELSE(task state comparator A>0,task state comparator A,0) task state output B=IF THEN ELSE(task state comparator B>0,task state comparator B,0) task state output C=IF THEN ELSE(task state comparator C>0,task state comparator C,0) task st ate output D=IF THEN ELSE(task state comparator D>0,task state comparator D,0) task state perception A=task state A+task state bias A task state perception B=task state B+task state bias B task state perception C=task state C+task state bias C task state p erception D=task state D+task state bias D "task contingent leadership behavior A"=IF THEN ELSE(regulation target=5,position importance A*leader A LMX*leader reactivity*leader task state output A/4,IF THEN ELSE(regulation target=1,position importance A*lea der A LMX*leader reactivity*leader task state output A,0)) "task contingent leadership behavior B"=IF THEN ELSE(regulation target=5,position importance B*leader B LMX*leader reactivity*leader task state output B/4,IF THEN ELSE(regulation target=2,position importance B*leader B LMX*leader reactivity*leader task state output B,0)) "task contingent leadership behavior C"=IF THEN ELSE(regulation target=5,position importance C*leader C LMX*leader reactivity*leader task state output C/4,IF THEN ELSE(regulation ta rget=3,position importance C*leader C LMX*leader reactivity*leader task state output C,0)) "task contingent leadership behavior D"=IF THEN ELSE(regulation target=5,position importance D*leader D LMX*leader reactivity*leader task state output D/4,IF THEN

PAGE 196

196 EL SE(regulation target=4,position importance D*leader D LMX*leader reactivity*leader task state output D,0)) team competency bias A=0 team competency bias B=0 team competency bias C=0 team competency bias D=0 team competency perception A=IF THEN ELSE(team ta sk structure=2, min(min(competency perception A, competency B),min(competency C, competency D)) , IF THEN ELSE(team task structure=3, max(max(competency perception A, competency B),max(competency C, competency D)) , (competency perception A+competency B+co mpetency C+competency D)/4))+team competency bias A team competency perception B=IF THEN ELSE(team task structure=2, min(min(competency A, competency perception B),min(competency C, competency D)) , IF THEN ELSE(team task structure=3, max(max(competency A, competency perception B),max(competency C, competency D)) ,(competency A+competency perception B+competency C+competency D)/4))+team competency bias B team competency perception C=IF THEN ELSE(team task structure=2, min(min(competency A, competency B),min (competency perception C, competency D)),IF THEN ELSE(team task structure=3, max(max(competency A, competency B),max(competency perception C, competency D)),(competency A+competency B+competency perception C+competency D)/4))+team competency bias C team co mpetency perception D=IF THEN ELSE(team task structure=2, min(min(competency A, competency B),min(competency C, competency perception D)) , IF THEN ELSE(team task structure=3, max(max(competency A, competency B),max(competency C, competency perception D)) , (competency A+competency B+competency C+competency perception D)/4))+team competency bias D team expectancy comparator A=max((accepted deadline A Time team expectancy input A),0) team expectancy comparator B=max((accepted deadline B Time team expectancy input B),0) team expectancy comparator C=max((accepted deadline C Time team expectancy input C),0) team expectancy comparator D=max((accepted deadline D Time team expectancy input D),0) team expectancy input A=team task state output A/team competency perce ption A team expectancy input B=team task state output B/team competency perception B team expectancy input C=team task state output C/team competency perception C

PAGE 197

197 team expectancy input D=team task state output D/team competency perception D team expectanc y output A=team expectancy comparator A team expectancy output B=team expectancy comparator B team expectancy output C=team expectancy comparator C team expectancy output D=team expectancy comparator D team goal=1 team goal A="leader assigned team goal"+te am goal bias A team goal B="leader assigned team goal"+team goal bias B team goal bias A=0 team goal bias B=0 team goal bias C=0 team goal bias D=0 team goal C="leader assigned team goal"+team goal bias C team goal D="leader assigned team goal"+team goal b ias D team task gain=IF THEN ELSE(leader team task state output>0, 0 , 1 ) team task state= INTEG (IF THEN ELSE(team task structure=1, min(( team task state+(task state A+task state B+task state C+task state D)/4),( team task state+"leader assigned team go al")),IF THEN ELSE(team task structure=2, min(( team task state+min(min(task state A,task state B),min(task state C,task state D))),( team task state+"leader assigned team goal")), min(( team task state+max(max(task state A,task state B),max(task state C,t ask state D))),( team task state+"leader assigned team goal"))))+clear team task state,0) team task state bias A=0 team task state bias B=0 team task state bias C=0 team task state bias D=0 team task state comparator A=IF THEN ELSE(team goal A team task st ate perception A>0, team goal A team task state perception A, 0)

PAGE 198

198 team task state comparator B=IF THEN ELSE(team goal B team task state perception B>0, team goal B team task state perception B, 0) team task state comparator C=IF THEN ELSE(team goal C team t ask state perception C>0, team goal C team task state perception C, 0) team task state comparator D=IF THEN ELSE(team goal D team task state perception D>0, team goal D team task state perception D, 0) team task state output A=team task state comparator A team task state output B=team task state comparator B team task state output C=team task state comparator C team task state output D=team task state comparator D team task state perception A=team task state+team task state bias A team task state perception B=team task state+team task state bias B team task state perception C=team task state+team task state bias C team task state perception D=team task state+team task state bias D team task structure=1 time spent on task team= INTEG (TIME STEP+clear time spe nt on task team,0) total number of task=1 "utility of off task activities A"=0 "utility of off task activities B"=0 "utility of off task activities C"=0 "utility of off tasks activities D"=0 workflow centrality A=1 workflow centrality B=1 workflow centrali ty C=1 workflow centrality D=1

PAGE 199

199 APPENDIX B TASK AND MANIPULATIONS USED IN STUDY 2 Simulation Task Interface Descriptions Used in Team Task Structure Manipulation Functional Condition Remember, the collective performance of the kitchen is determined by the interdependent work among the four team members. An individual team member cannot accomplish his/her task without information or materials from other members of the team. To function effectively as a group, each individual has to effectively perform th e specific duty he/she is assigned to. The worst performer will significantly hurt the performance of the group.

PAGE 200

200 Divisional Condition Remember, the collective performance of the kitchen is determined by the average performance of the four team members. A n individual team member can accomplish his/her task independently without information or materials from other members of the team. To function effectively as a team, each individual has to effectively perform the specific duty he/she is assigned to. The a mount of food delivered is the simple average output. Disjunctive Condition Remember, the collective performance of the kitchen is determined by the best performer among the four team members. An individual team member can accom plish his/her task independently without information or materials from other members of the team. To function effectively as a team, one team member has to effectively deliver the food ordered by the guests. Only the amount of food delivered by the best pe rformer matters. Descriptions Used in Differentiation in Learning Ability Manipulation Same Learning Ability Condition Andrew Position: sous chef Age: 30 Gender: Male Ethnicity: Caucasian Marital status: Single Education: High school diploma Experience as a sous chef: 6 years Time working in the restaurant: 2 years Responsibilities: manages and coordinates line cooks in daily operation, distributes orders Brian Position: pastry chef Age: 30 Gender: Male Ethnicity: Caucasian Marital status: Single Educati on: High school diploma Experience as a pastry chef: 6 years Time working in the restaurant: 2 years Responsibilities: orders baking ingredients, prepares baked goods and desserts for guests Christopher Position: prep chef Age: 30 Gender: Male

PAGE 201

201 Ethnicity: Caucasian Marital status: Single Education: High school diploma Experience as a prep chef: 6 years Time working in the restaurant: 2 years Responsibilities: oversees pantry, manages and coordinates preparation for cooking Daniel Position: wine steward Age : 30 Gender: Male Ethnicity: Caucasian Marital status: Single Education: High school diploma Experience as a wine steward: 6 years Time working in the restaurant: 2 years Responsibilities: orders wine, keeps wine inventory, educates and suggests wine for g uests Different Learning Ability Condition Andrew Position: sous chef Age: 30 Gender: Male Ethnicity: Caucasian Marital status: Single Education: Middle school diploma Experience as a sous chef: 6 years Time working in the restaurant: 2 years Responsibili ties: manages and coordinates line cooks in daily operation, distributes orders Brian Position: pastry chef Age: 30 Gender: Female Ethnicity: Caucasian Marital status: Single Education: High school diploma Experience as a pastry chef: 6 years Time workin g in the restaurant: 2 years Responsibilities: orders baking ingredients, prepares baked goods and desserts for guests Christopher Position: prep chef Age: 30 Gender: Female Ethnicity: Caucasian

PAGE 202

202 Marital status: Single Education: High school diploma Exper ience as a prep chef: 6 years Time working in the restaurant: 2 years Responsibilities: oversees pantry, manages and coordinates preparation for cooking Daniel Position: wine steward Age: 30 Gender: Male Ethnicity: Caucasian Marital status: Single Educati on: Associate degree Experience as a wine steward: 6 years Time working in the restaurant: 2 years Responsibilities: orders wine, keeps wine inventory, educates and suggests wine for guests Descriptions Used for Each Team Performance Episode Same Task Dif ficulty, Small Disturbance Magnitude This Saturday, the restaurant is reserved to host a social event for a local business inviting 12 guests . The guests are expecting a three course dinner with all of the dishes specially designed for the event. You, the executive chef, have designed the recipes and the presentation of the dishes. Now you need to train your team members to make sure they prepare sufficient high quality food for the event. Sous chef, pastry chef, prep chef, and wine steward will face the sa me workload during the event. The guests did not suggest any special preference for any aspect of the meal. The four chefs will have to coordinate with each other. Same Task Difficulty, Large Disturbance Magnitude This Saturday, the restaurant is hosting a recruiting event for a local organization inviting 12 guests . The guests are expecting a three course dinner with all of the dishes specially designed for the event. You, the executive chef, have designed the recipes and the presentation of the dishes. No w you need to train your team members to make sure they prepare sufficient high quality food for the event. Sous chef, pastry chef, prep chef, and wine steward will face the same workload during the event. The guests did not suggest any special preference for any aspect of the meal. The four chefs will have to coordinate with each other.

PAGE 203

203 Different Task Difficulty, Small Disturbance Magnitude tasting for 12 gu ests . The guests are expecting a three course dinner with all of the dishes specially designed for the event. You, the executive chef, have designed the recipes and the presentation of the dishes. Now you need to train your team members to make sure they p repare sufficient high quality food for the event. pastry chef will face the largest workload and have to solve most of the unexpected problems during the event. The sous chef, prep chef, and wine steward will have to co ordinate with the pastry chef. Different Task Difficulty, Large Disturbance Magnitude This Saturday, the restaurant is reserved to host a professional networking event for a group of 12 guests invited by a local organization. The guests are expecting a th ree course dinner with all of the dishes specially designed for the event. You, the executive chef, have designed the recipes and the presentation of the dishes. Now you need to train your team members to make sure they prepare sufficient high quality food for the event. the wine steward will face the largest workload and have to solve most of the unexpected problems during the event. The sous chef, pastry chef, and prep chef will have to coordinate with the wine steward.

PAGE 204

204 APPENDIX C MEASURES USED IN STUDY 3 Time 1 Subordinate Survey Questions Demographic Information 1. Date of birth (yyyy mm dd) 2. Gender: Male, Female 3. Highest degree earned: Elementary school, middle school, vocational school, high school, vocational training college, bachelor, master, doctoral 4. Organizational tenure (year) 5. How long have you been working with your direct team supervisor? ( month ) Skill Va riety, Job Complexity, and Problem Solving Scale 1 = strongly disagree 2 = moderately disagree 3 = neutral 4 = moderately agree 5 = strongly agree 1. The job requires that I only do one task or activity at a time. 2. The tasks on the job are simple and uncompli cated. 3. The job comprises relatively uncomplicated tasks. 4. The job involves performing relatively simple tasks. 5. The job involves solving problems that have no obvious correct answer. 6. The job requires me to be creative. 7. The job often involves dealing with pro blems that I have not met before. 8. The job requires unique ideas or solutions to problems. 9. The job requires a variety of skill. 10. The job requires me to utilize a variety of different skills in order to complete the work. 11. The job requires me to use a number o f complex or high level skills. 12. The job requires the use of a number of skills. Learning Ability Scale 1 = strongly disagree 2 = moderately disagree 3 = neutral 4 = moderately agree 5 = strongly agree 1. I enjoy learning new approaches for conducting work .

PAGE 205

205 2. I often learn new information and skills to stay at the forefront of my profession. 3. I train to keep my work skills and knowledge current. 4. I am continually learning new skills for my job. 5. I take responsibility for staying current in my profession. Conscien tiousness Scale 1 = Extremely Inaccurate 2 = Very Inaccurate 3 = Moderately Inaccurate 4 = Slightly Iaccurate 5 = Neutral 6 = Slightly Accurate 7 = Moderately Accurate 8 = Very Accurate 9 = Extremely Accurate 1. Careless 2. Disorganized 3. Efficient 4. Inefficient 5. O rganized 6. Practical 7. Sloppy 8. Systematic Time 1 Team Leader Survey Questions Individual focused Leadership Behaviors Scale 1 = strongly disagree 2 = moderately disagree 3 = neutral 4 = moderately agree 5 = strongly agree 1. I provide this subordinate with know ledge, skills, and expertise to help him/her get the job done. 2. I provide this subordinate with useful advice on his/her job. 3. I provide this subordinate with necessary information about what to do or how to do it about his/her job . 4. I give this subordinate c onsiderable opportunity for independence and freedom in how he/she does the work. 5. I allow this subordinate to decide on his/her own how to go about doing his/her work. 6. I provide this subordinate with significant autonomy in making decisions .

PAGE 206

206 7. I can enjoy he aring about what this subordinate thinks and feels . 8. This subordinate can rely on me for emotional support. 9. There is me whom this subordinate could go to if he/she were just feeling down, without feeling funny about it later. Time 1 Network Survey Question s Please consider the workflow among positions in your work group. Here is a list of employees in your work group. To whom do you receive inputs from in the team work processes? Please check those names. Time 2 Team Manager Survey Questions Team Performan ce Scale 1 = strongly disagree 2 = moderately disagree 3 = slightly disagree 4 = neutral 5 = slightly agree 6 = moderately agree 7 = strongly agree 1. Team members often implement new ideas to improve the quality of our products and services. 2. This team pays attention to using new work methods and procedures. 3. This team often creates new products, techniques, services, work methods, and work procedures. 4. This is a creative team.

PAGE 207

207 LIST OF REFERENCES Agle, B. R., Nagarajan, N. J., Sonnenfeld, J. A., & Srinivasan , D. (2006). Does CEO charisma matter? An empirical analysis of the relationships among organizational performance, environmental uncertainty, and top management team perceptions of CEO charisma. Academy of Management Journal , 49 , 161 174. Anand, S., Vidya rthi, P. R., Liden, R. C., & Rousseau, D. M. (2010). Good citizens in poor quality relationship: Idiosyncratic deals as a substitute for relationship quality. Academy of Management Journal, 53, 970 988. Anderson, N. R., & West, M. A. (1998). Measuring clim ate for work group innovation: development and validation of the team climate inventory. Journal of Organizational Behavior , 19 , 235 258. Balkundi, P., & Harrison, D. A. (2006). Ties, leaders, and time in teams: Strong inference about ffects on team viability and performance. Academy of Management Journal , 49 , 49 68. Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice Hall. Barrick, M. R., & Mount, M. K. (1991). The Big Five personality dimensions and job performance: A meta analysis. Personnel Psychology, 44, 1 26. Barrick, M. R., Stewart, G. L., Neubert, M. J., & Mount, M. K. (1998). Relating member ability and personality to work team processes and team effectiveness. Journal of Applied Psychology, 83 , 377 391. Barry, B., & Stewart, G. L. (1997). Composition, process, and performance in self managed groups: the role of personality. Journal of Applied P sychology , 82 , 62 78 . Bass, B. M. (1985). Leadership and performance beyond expectations. New York: Free Press. Bass, B. M. (1990). From transactional to transformational leadership: Learning to share the vision. Organizational Dynamics, 18, 19 31. Bass, B. M., Avolio, B. J., Jung, D. I., & Berson, Y. (2003). Predicting un it performance by assessing transformational and transactional leadership. Journal of Applied Psychology , 88 , 207 217. Bauer, T. N., & Green, S. G. (1996). Development of leader member exchange: A longitudinal test. Academy of Management Journal, 39, 1538 1567. Bell, S. T. (2007). Deep level composition variables as predictors of team performance: a meta analysis. Journal of Applied Psychology , 92 , 595 615.

PAGE 208

208 Bell, B. S., & Kozlowski, S. W. J. (2002). A typology of virtual teams: Implications for effective le adership. Group & Organization Management, 27, 14 49. Blau, P. M. (1964). Exchange and power in social life . New York: John Wiley. Bloom, M. (1999). The performance effects of pay dispersion on individuals and organizations. Academy of Management Journal , 42 , 25 40. Boies, K., & Howell, J. M. (2006). Leader member exchange in teams: An examination of the interaction between relationship differentiation and mean LMX in explaining team level outcomes. The Leadership Quarterly, 17, 246 257. Boulding, K. E. (1 956). General systems theory the skeleton of science. Management S cience , 2 , 197 208. Bowen, D. E., & Ostroff, C. (2004). Understanding HRM firm performance linkages: The role of the" strength" of the HRM system. Academy of M anagement R eview , 29 , 203 221. Brass, D. J. (1985). Men's and women's networks: a study of interaction patterns and influence in an organization. Academy of Management Journal , 28 , 327 343. Brass, D. J. (2012). A social network perspective on organizational psychology. In S. W. J. Kozlo wski (Eds.), Oxford handbook of organizational psychology , (pp. 667 695). NY: Oxford University Press. Brislin, R. W. (1980). Translation and content analysis of oral and written material. In H. C. Triandis & J. W. Berry (Eds.), Handbook of crosscultural p sychology (pp. 398 444). Boston, MA: Allyn & Bacon. Burns, T., & Stalker, G. M. (1961). The management of innovation. University of Illinois at Urbana Champaign's Academy for Entrepreneurial Leadership Historical Research Reference in Entrepreneurship . Ca nnella, A. A., Park, J. H., & Lee, H. U. (2008). Top management team functional background diversity and firm performance: Examining the roles of team member colocation and environmental uncertainty. Academy of Management Journal , 51 , 768 784. Carley, K. M ., Pfeffer, J., Reminga, J., Storrick, J., & Columbus, D. (2013). ORA user's guide 2013 . Pittsburgh, PA: Carnegie Mellon University. Carmeli, A., Schaubroeck, J., & Tishler, A. (2011). How CEO empowering leadership shapes top management team processes: Imp lications for firm performance. The Leadership Quarterly , 22 , 399 411. Carpenter, M. A., & Fredrickson, J. W. (2001). Top management teams, global strategic posture, and the moderating role of uncertainty. Academy of Management Journal , 44 , 533 545.

PAGE 209

209 Carver , C. S., & Scheier, M. F. (1998). On the Self regulation of Behavior . New York, NY: Cambridge University Press. Chen, G., & Kanfer, R. (2006). Toward a systems theory of motivated behavior in work teams. Research in O rganizational B ehavior , 27 , 223 267. Ch en, G., Kirkman, B. L., Kanfer, R., Allen, D., & Rosen, B. (2007). A multilevel study of leadership, empowerment, and performance in teams. Journal of Applied Psychology, 92, 331 346. Chen, G., Sharma, P. N., Edinger, S., Shapiro, D. L., & Farh, J. L. (20 11). Motivating and de motivating forces in teams: Cross level influences of empowering leadership and relationship conflict. Journal of Applied Psychology, 96, 541 557. Colbert, A. E., Kristof Brown, A. L., Bradley, B. H., & Barrick, M. R. (2008). CEO tr ansformational leadership: The role of goal importance congruence in top management teams. Academy of Management Journal , 51 , 81 96. Cole, M. S., Bedeian, A. G., & Bruch, H. (2011). Linking leader behavior and leadership consensus to team performance: Inte grating direct consensus and dispersion models of group composition. The Leadership Quarterly , 22 , 383 398. Colquitt, J. A., LePine, J. A., & Noe, R. A. (2000). Toward an integrative theory of training motivation: a meta analytic path analysis of 20 years of research. Journal of Applied Psychology , 85 , 678 707. Cordery, J. L., Morrison, D., Wright, B. M., & Wall, T. D. (2010). The impact of autonomy and task uncertainty on team performance: A longitudinal field study. Journal of Organizational Behavior , 31 , 240 258. Cropanzano, R., & Mitchell, M. S. (2005). Social exchange theory: An interdisciplinary review. Journal of Management, 31, 874 900.Day, D. V., Gronn, P., & Salas, E. (2006). Leadership in team based organizations: On the threshold of a new era. Th e Leadership Quarterly , 17 , 211 216. DeChurch, L. A., & Marks, M. A. (2006). Leadership in multiteam systems. Journal of Applied Psychology , 91 , 311 329. De Dreu, C. K. W. (2002). Team innovation and team effectiveness: The importance of minority dissent a nd reflexivity. European Journal of Work and Organizational Psychology , 11 , 285 298. Delery, J. E., & Doty, D. H. (1996). Modes of theorizing in strategic human resource management: Tests of universalistic, contingency, and configurations. performance pred ictions. Academy of management Journal , 39 , 802 835.

PAGE 210

210 DeShon, R. P., Kozlowski, S. W., Schmidt, A. M., Milner, K. R., & Wiechmann, D. (2004). A multiple goal, multilevel model of feedback effects on the regulation of individual and team performance. Journal of Applied Psychology , 89 , 1035 1056. psychological collectivism on team performance over time. Journal of A pplied P sychology , 96 , 247 262. Druskat, V. U., & Wheeler, J . V. (2003). Managing from the boundary: The effective leadership of self managing work teams. Academy of Management Journal , 46, 435 457. Edmondson, A. C. (2003). Speaking up in the operating room: How team leaders promote learning in interdisciplinary ac tion teams. Journal of Management Studies , 40, 1419 1452. Eisenbeiss, S. A., van Knippenberg, D., & Boerner, S. (2008). Transformational leadership and team innovation: Integrating team climate principles. Journal of Applied Psychology , 93, 1438 1446. Fish er, D. M., Bell, S. T., Dierdorff, E. C., & Belohlav, J. A. (2012). Facet personality and surface level diversity as team mental model antecedents: implications for implicit coordination. Journal of Applied Psychology , 97 , 825 841 . Fleishman, E. A., Mumfor d, M. D., Zaccaro, S. J., Levin, K. Y., Korotkin, A. L., & Hein, M. B. (1991). Taxonomic efforts in the description of leader behavior: A synthesis and functional interpretation. The Leadership Quarterly , 2, 245 287. Foa, U. G., & Foa, E. B. (1980). Resour ce theory: Interpersonal behavior as exchange. In K. J. Gergen & M. S. Greenberg & R. H. Willis (Eds.), Social exchange: Advances in theory and research. New York: Plenum. Freeman, L. C. (1979). Centrality in social networks: Conceptual clarification. Soci al Networks, 1, 215 239. Gerstner, C. R., & Day, D. V. (1997). Meta analytic review of leader member exchange theory: Correlates and construct issues. Journal of Applied Psychology , 82 , 827 843. Gibson, C. B. (1999). Do they do what they believe they can? Group efficacy and group effectiveness across tasks and cultures. Academy of Management Journal , 42 , 138 152. Goodman, P. S., Devadas, R., & Griffith Hughson, T. L. (1988). Groups and productivity; Analyzing the effectiveness of self managing teams. In J . P. Campbell & R . J. Campbell (Eds.), Productivity in Organizations : New Perspectives from Industrial and Organizational Psychology (pp. 295 327) . San Francisco , CA : Jossey Bass. 4). What makes a good team player? Development of a conditional reasoning test of team orientation. In S. Gustafson (Chair), Making conditional reasoning tests work: Reports from the frontier.

PAGE 211

211 Symposium conducted at the 19th Annual Conference of the Societ y for Industrial and Organizational Psychology, Chicago . Graen, G. B., & Scandura, T. A. (1987). Toward a psychology of dyadic organizing. Research in Organizational Behavior , 9, 175 208. Graen, G. B., & Uhl Bien, M. (1995). Relationship based approach to leadership: Development of leader member exchange (LMX) theory of leadership over 25 years: Applying a multi level multi domain perspective. The Leadership Quarterly, 6, 219 247. Gundersen, G., Hellesøy, B. T., & Raeder, S. (2012). Leading International P roject Teams The Effectiveness of Transformational Leadership in Dynamic Work Environments. Journal of Leadership & Organizational Studies , 19 , 46 57. Hackman, J. R., & Wageman, R. (2005). A theory of team coaching. Academy of Management Review , 30 , 269 28 7. Hackman, J. R., & Walton, R. E. (1986). Leading groups in organizations. In P. S. Goodman (eds.), Designing Effective Work Groups . San Francisco, CA: Jossey Bass. Haleblian, J., & Finkelstein, S. (1993). Top management team size, CEO dominance, and firm performance: the moderating roles of environmental turbulence and discretion. Academy of Management Journal , 36 , 844 863. Harrison, D. A., & Humphrey, S. E. (2010). Designing for diversity or diversity for design? Tasks, interdependence, and within unit d ifferences at work. Journal of Organizational Behavior , 31 , 328 337. Harrison, D. A., & Klein, K. J. (2007). What's the difference? Diversity constructs as separation, variety, or disparity in organizations. Academy of Management Review , 32 , 1199 1228. Hen derson, D. J., Liden, R. C., Glibkowski, B. C., & Chaudhry, A. (2009). LMX differentiation: A multilevel review and examination of its antecedents and outcomes. The leadership Q uarterly , 20 , 517 534. Hersey, P., & Blanchard, K. (1977). Management of organi zation behavior: Utilizing human resources (3 rd ed.). Englewood Cliffs, NJ: Prentice Hall. Hmieleski, K. M., & Ensley, M. D. (2007). A contextual examination of new venture performance: entrepreneur leadership behavior, top management team heterogeneity, a nd environmental dynamism. Journal of Organizational Behavior , 28 , 865 889. Hoffman, B. J., Bynum, B. H., Piccolo, R. F., & Sutton, A. W. (2011). Person organization value congruence: How transformational leaders influence work group effectiveness. Academy of Management Journal , 54 , 779 796.

PAGE 212

212 Hollenbeck, J. R., DeRue, D. S., & Guzzo, R. (2004). Bridging the gap between I/O research and HR practice: Improving team composition, team training, and team task design. Human Resource Management , 43 , 353 366. Hollen beck, J. R., Ellis, A. P., Humphrey, S. E., Garza, A. S., & Ilgen, D. R. (2011). Asymmetry in structural adaptation: The differential impact of centralizing versus decentralizing team decision making structures. Organizational Behavior and Human Decision P rocesses , 114 , 64 74. Hollenbeck, J. R., Moon, H., Ellis, A. P., West, B. J., Ilgen, D. R., Sheppard, L., Porter, C. O. L. H., & Wagner III, J. A. (2002). Structural contingency theory and individual differences: Examination of external and internal person team fit. Journal of Applied Psychology , 87 , 599 606. Hollenbeck, J. R., & Spitzmuller, M. (2012). Team structure: Tight versus loose coupling in task oriented groups. In S. W. J. Kozlowski (eds.), The Oxford Handbook of Organizational Psychology (pp. 733 766). New York, NY: Oxford Press. Holman, D., Clegg, C., & Waterson, P. (2002). Navigating the territory of job design. Applied Ergonomics , 33 , 197 205. House, R. J. (1971). A path goal theory of leadership effectiveness. Administrative Science Quarterly, 16, 321 339. Howell, J. P., Bowen, D. E., Dorfman, P. W., Kerr, S. , & Podsakoff, P. M. (1990). Substitutes for leadership: Effective alternatives to ineffective leadership. Organizational Dynamics, 19, 21 38. Humphrey, S. E., Morgeson, F. P., & Mannor, M . J. (2009). Developing a theory of the strategic core of teams: a role composition model of team performance. Journal of Applied Psychology , 94 , 48 61. Ilgen, D. R., & Hulin, C. L. (2000). Computational modeling of behavior in organizations: The third sci entific discipline . Washington , DC : American Psychological Association. Johns, G. (2006). The essential impact of context on organizational behavior. Academy of Management Review, 31, 386 408. Johnson, M. D., Hollenbeck, J. R., Humphrey, S. E., Ilgen, D . R., Jundt, D., & Meyer, C. J. (2006). Cutthroat Cooperation: Asymmetrical adaptation to changes in team reward structures. Academy of Management Journal , 49 , 103 119. Jordan, P. J., & Troth, A. C. (2004). Managing emotions during team problem solving: Em otional intelligence and conflict resolution. Human P erformance , 17 , 195 218. Kanfer, R., & Ackerman, P. L. (1989). Motivation and cognitive abilities: An integrative/aptitude treatment interaction approach to skill acquisition. Journal of Applied Psycholo gy, 74, 657 690.

PAGE 213

213 Kark, R., Shamir, B., & Chen, G. (2003). The two faces of transformational leadership: empowerment and dependency. Journal of A pplied P sychology , 88 , 246 255. Katz, D., & Kahn, R. L. (1978). The social psychology of organizations (2nd ed.) . New York: Wiley. Kearney, E., & Gebert, D. (2009). Managing diversity and enhancing team outcomes: the promise of transformational leadership. Journal of Applied Psychology , 94 , 77 89. Kerr, S., & Jermier, J. (1978). Substitutes for leadership: Their mea ning and measurement. Organizational Behavior & Human Performance, 22, 375 403. Kirkman, B. L., & Rosen, B. (1999). Beyond self management: Antecedents and consequences of team empowerment. Academy of Management Journal , 42 , 58 74. Kirkman, B. L., Rosen, B ., Tesluk, P. E., & Gibson, C. B. (2004). The Impact of Team Empowerment on Virtual Team Performance: The Moderating Role of Face to Face Interaction. Academy of Management Journal , 47 , 175 192. Klein, K. J., Ziegert, J. C., Knight, A. P., & Xiao, Y. (2006 ). Dynamic delegation: Shared, hierarchical, and deindividualized leadership in extreme action teams. Administrative Science Quarterly , 51 , 590 621. Kozlowski, S. W. J., & Bell, B. S. (2003). Work groups and teams in organizations. In W. C. Borman, D. R. I lgen, & R. Klimoski (Eds.), Handbook of psychology: Industrial and organizational psychology (Vol. 12 , pp. 333 375). London: Wiley. Kozlowski, S. W. J., Gully, S. M., Salas, E., & Cannon Bowers, J. A. (1996). Team leadership and development: Theory, princi ples, and guidelines for training leaders and teams. In M. M. Beyerlein, D. A. Johnson, & S. T. Beyerlein (Eds.), Advances in interdisciplinary studies of work teams (Vol. 3, pp. 253 291). Greenwich, CT: JAI Press. Kozlowski, S. W. J., & Ilgen, D. R. (2006 ). Enhancing the effectiveness of work groups and teams. Psychological Science , 7, 77 124. Kozlowski, S. W. J., & Klein, K. J. (2000). A multilevel approach to theory and research in organizations: Contextual, temporal, and emergent processes. In K. J. Kle in & S. W. J. Kozlowski (Eds.), Multilevel theory, research, and methods in organizations: Foundations, extensions, and new directions (pp. 3 90). San Francisco, CA: Jossey Bass. Kuhn, T. S. (1962). The structure of scientific revolutions . Chicago, IL: Uni versity of Chicago Press. Lanaj, K., Hollenbeck, J. R., Ilgen, D. R., Barnes, C. M., & Hormon, S. J. (2013). The double edged sword of decentralized planning in multiteam systems. Academy of Management Journal, 56, 735 757.

PAGE 214

214 Le Blanc, P. M., & González Romá , V. (2012). A team level investigation of the relationship between Leader Member Exchange (LMX) differentiation, and commitment and performance. The Leadership Quarterly, 23, 534 544 . LePine, J. A. (2003). Team adaptation and postchange performance: Effec ts of team composition in terms of members' cognitive ability and personality. Journal of Applied Psychology , 88 , 27 39. LePine, J. A. (2005). Adaptation of teams in response to unforeseen change: effects of goal difficulty and team composition in terms of cognitive ability and goal orientation. Journal of Applied Psychology , 90 , 1153 1167. LePine, J. A., Colquitt, J. A., & Erez, A. (2000). Adaptability to changing task contexts: Effects of general cognitive ability, conscientiousness, and openness to exper ience. Personnel Psychology , 53 , 563 593. LePine, J. A., Hollenbeck, J. R., Ilgen, D. R., & Hedlund, J. (1997). Effects of individual differences on the performance of hierarchical decision making teams: Much more than g. Journal of Applied Psychology , 82 , 803 811. Liden, R. C., Erdogan, B., Wayne, S. J., & Sparrowe, R. T. (2006). Leader member exchange, differentiation, and task interdependence: implications for individual and group performance. Journal of Organizational Behavior , 27 , 723 746. Lim, B. C., & Ployhart, R. E. (2004). Transformational leadership: relations to the five factor model and team performance in typical and maximum contexts. Journal of Applied Psychology , 89 , 610 621. Locke, E. A., & Latham, G. P. (1990). A theory of goal setting and t ask performance . Prentice Hall, Inc. Lord, R. G., & Levy, P. E. (1994). Moving from cognition to action: A control theory perspective. Applied Psychology: An International Review, 43, 335 367. Mayer, D. M., Kuenzi, M., Greenbaum, R., Bardes, M., & Salvador, R. (2009). How low does ethical leadersh ip flow? Test of a trickle down model. Organizational Behavior and Human Decision Processes, 108, 1 13. Manz, C. C., & Sims, H. P. (1980). Self management as a substitute for leadership: A social learning theory per spective. Academy of Management Review , 5, 361 367. on leading selfmanaged groups. Human Relations , 37, 409 424. Manz, C. C., & Sims, H. P. (1987). Leading workers to lead themselves: The external leadership of self managing work teams. Administrative Science Quarterly , 32, 106 128. Mathieu, J. E., Gilson, L. L., & Ruddy, T. M. (2006). Empowerment and team effectiveness: an empirical test of an integrated model. Jou rnal of Applied Psychology , 91 , 97 108.

PAGE 215

215 Marks, M. A., Mathieu, J. E., & Zaccaro, S. J. (2001). A temporally based framework and taxonomy of team processes. Academy of Management Review , 26 , 356 376. Martin, A., & Bal, V. (2006). The state of teams: CCL res earch report . Greensboro, NC: Center for Creative Leadership. McGrath, J. E. (1962). Leadership behavior: Some requirements for leadership training . Washington, DC: U.S. Civil Service Commission, Office of Career Development. Mehra, A., Kilduff, M., & Bras s, D. J. (2001). The social networks of high and low self monitors: Implications for workplace performance. Administrative Science Quarterly , 46 , 121 146. Morgeson, F. P. (2005). The external leadership of self managing teams: Intervening in the context of novel and disruptive events. Journal of Applied Psychology , 90 , 497 508. Morgeson, F. P., & DeRue, D. S. (2006). Event criticality, urgency, and duration: Understanding how events disrupt teams and influence team leader intervention. The Leadership Quarte rly , 17 , 271 287. Morgeson, F. P., DeRue, D. S., & Karam, E. P. (2010). Leadership in teams: A functional approach to understanding leadership structures and processes. Journal of Management , 36 , 5 39. Morgeson, F. P., Humphrey, S. E. (2006). The work desi gn questionnaire (WDQ): Developing and validating a comprehensive measure for assessing job design and the nature of work. Journal of Applied Psychology, 91, 1321 1339. Morgeson, F. P., & Humphrey, S. E. (2008). Job and team design: Toward a more integrati ve conceptualization of work design. Research in P ersonnel and H uman R esources M anagement , 27 , 39 91. Mumford, T. V., Campion, M. A., & Morgeson, F. P. (2007). The leadership skills strataplex: Leadership skill requirements across organizational levels. Th e Leadership Quarterly , 18 , 154 166. Murphy, S. E., & Ensher, E. A. (2008). A qualitative analysis of charismatic leadership in creative teams: The case of television directors. The Leadership Quarterly , 19 , 335 352. Muth én, L. K., & Muthé n, B. O. (2010). (6th ed.) . Los Angeles, CA: Author. Myung, I. J. (2000). The importance of complexity in model selection. Journal of Mathematical Psychology, 44, 190 204. Myung, I. J. (2003). Tutorial on maximum likelihood estimation. Journal of Mathema tical Psychology, 47, 90 100.

PAGE 216

216 Naidoo, L. J., Scherbaum, C. A., Goldstein, H. W., & Graen, G. B. (2011). A longitudinal examination of the effects of LMX, ability, and differentiation on team performance. Journal of Business and Psychology , 26 , 347 357. Nem anich, L. A., & Vera, D. (2009). Transformational leadership and ambidexterity in the context of an acquisition. The Leadership Quarterly , 20 , 19 33. Oh, H., Chung, M. H., & Labianca, G. (2004). Group social capital and group effectiveness: The role of inf ormal socializing ties. Academy of Management Journal , 47 , 860 875. Oliva, R. (2003). Model calibration as a testing strategy for system dynamics models. European Journal of Operational Research, 151, 552 568. Parker, S. K., & Ohly, S. (2008). Designing m otivating jobs. In R. Kanfer, G. Chen, and R. D. Pritchard ( E ds.), Work Motivation: Past, Present, and Future (pp. 233 284) . SIOP Organizational Frontiers Series . Piccolo, R. F., & Colquitt, J. A. (2006). Transformational leadership and job behaviors: The mediating role of core job characteristics. Academy of Management Journal , 49 , 327 340. Ployhart, R. E., & Bliese, P. D. (2006). Individual adaptability (I ADAPT) theory: Conceptualizing the antecedents, consequences, and measurement of individual differen ces in adaptability. In C. S. Burke, L. G. Pierce, & E. Salas (Eds.), Understanding adaptability: A prerequisite for effective performance within complex environments. Advances in human performance and cognitive engineering research (Vol 6; pp. 3 39). Amst erdam, Netherlands: Elsevier Podsakoff, P. M., MacKenzie, S. B., Lee, J., & Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88, 879 903. Pods akoff, P. M., MacKenzie, S. B., Moorman, R. H., & Fetter, R. (1990). Transformational leader behaviors and their effects on followers' trust in leader, satisfaction, and organizational citizenship behaviors. The Leadership Quarterly , 1 , 107 142. Powers, W. T. (1973). Behavior: The Control of Perception. New York, NY: Hawthorne . Purvanova, R. K., & Bono, J. E. (2009). Transformational leadership in context: Face to face and virtual teams. The Leadership Quarterly , 20 , 343 357. Randall, K. R., Resick, C. J., & DeChurch, L. A. (2011). Building team adaptive capacity: The roles of sensegiving and team composition. Journal of Applied Psychology , 96 , 525 540. Rousseau, D. M., & Fried, Y. (2001). Location, location, location: contextualizing organizational research . Journal of O rganizational Behavior , 22 , 1 13. Saucier, G. (1994). Mini markers: A brief version of Goldberg's unipolar Big Five markers. Journal of Personality Assessment , 63 , 506 516.

PAGE 217

217 Schaubroeck, J., Lam, S. S., & Cha, S. E. (2007). Embracing transform ational leadership: tea values and the impact of leader behavior on team performance. Journal of Applied Psychology , 92 , 1020 1030. Schaubroeck, J., Lam, S. S., & Peng, A. (2011). Cognition based and affect based trust as mediators of leader behavior influ ences on team performance. Journal of Applied Psychology , 96 , 863 871. Schmidt, A. M., & DeShon, R. P. (2007). What to do? The effects of discrepancies, incentives, and time on dynamic goal prioritization. Journal of Applied Psychology, 92, 928 941. Shin, S. J., & Zhou, J. (2007). When is educational specialization heterogeneity related to creativity in research and development teams? Transformational leadership as a moderator. Journal of Applied Psychology , 92 , 1709 1721. Sivasubramaniam, N., Murry, W. D., Avolio, B. J., & Jung, D. I. (2002). A longitudinal model of the effects of team leadership and group potency on group performance. Group & Organization Management , 27 , 66 96. Soda, G., & Zaheer, A. (2012). A network perspective on organizational archite cture: performance effects of the interplay of formal and informal organization. Strategic Management Journal , 33 , 751 771. Sparrowe, R. T., & Liden, R. C. (1997). Process and structure in leader member exchange. Academy of M anagement Review , 22 , 522 552. Srivastava, A., Bartol, K. M., & Locke, E. A. (2006). Empowering Leadership in Management Teams: Effects on Knowledge Sharing, Efficacy, and Performance. Academy of Management Journal , 49 , 1239 1251. Steel, P., & König, C. J. (2006). Integrating theories o f motivation. Academy of Management Review, 31, 889 913. Steiner, I. D. (1972). Group process and productivity. New York, NY: Academic Press. Stewart, G. L. (2006). A meta analytic review of relationships between team design features and team performance. Journal of Management , 32 , 29 55. Stewart, G. L., & Barrick, M. R. (2000). Team structure and performance: Assessing the mediating role of intrateam process and the moderating role of task type. Academy of M anagement Journal , 43 , 135 148. Stogdill, R. M. ( 1959). Individual behavior and group achievement: A theory; the experimental evidence. Oxford, UK: Oxford Press. Taber, C. S., & Timpone, R. J. (1996). Computational modeling . Thousand Oaks, CA: Sage.

PAGE 218

218 Tangirala, S., Green, S. G., & Ramanujam, R. (2007). In Journal of Applied Psychology, 92, 309 320. Tesluk, P. E., & Mathieu, J. E. (1999). Overcoming roadblocks to effectiveness: Incorporating management of perf ormance barriers into models of work group effectiveness. Journal of A pplied Psychology , 84 , 200 217. Vancouver, J. B. (2000). Self regulation in organizational settings: A tale of two paradigms. Handbook of self regulation (pp. 303 341). San Diego, CA: Ac ademic Press. Vancouver, J. B. (2008). Integrating self regulation theories of work motivation into a dynamic process theory. Human Resource Management Review, 18, 1 18. Vancouver, J. B., Tamanini, K. B., & Yoder, R. J. (2010). Using dynamic computational models to reconnect theory and research: Socialization by the proactive newcomer as example. Journal of Management , 36 , 764 793. Vancouver, J. B., Weinhardt, J. M., & Schmidt, A. M. (2010). A formal, computational theory of multiple goal pursuit: Integrati ng goal choice and goal striving processes. Journal of Applied Psychology , 95 , 985 1008 . Venkataramani, V., Green, S. G., & Schleicher, D. J. (2010). Well connected leaders: The Jou rnal of Applied Psychology, 95, 1071 1084. Vroom, V. H., & Jago, A. G. (2007). The role of the situation in leadership. American Psychologist, 62, 17 24. Wageman, R. (1995). Interdependence and group effectiveness. Administrative Science Quarterly , 40, 145 180. Wageman, R. (1999). Task design, outcome interdependence, and individual differences: Joint effects on effort in task performing teams. Group Dynamics , 44 , 136 142. Wageman, R. (2001). How leaders foster self managing team effectiveness: Design choic es versus hands on coaching. Organization Science , 12 , 559 577. Wageman, R., Gardner, H., & Mortensen, M. (2012). The changing ecology of teams: New directions for teams research. Journal of Organizational Behavior , 33 , 301 315. Wall, T. D., Cordery, J. L. , & Clegg, C. W. (2002). Empowerment, performance, and operational uncertainty: A theoretical integration. Applied Psychology , 51 , 146 169. Wang, D., Waldman, D. A., & Zhang, Z. (2014). A meta analysis of shared leadership and team effectiveness. Journal o f Applied Psychology, 99, 181 198.

PAGE 219

219 Wang, X. H. F., & Howell, J. M. (2012). A multilevel study of transformational leadership, identification, and follower outcomes. The Leadership Quarterly, 23, 775 790. Wasserman, S., & Faust, K. (1994). Social network an alysis: Methods and applications. Cambridge, MA: Cambridge University Press. Wu, J. B., Tsui, A. S., & Kinicki, A. J. (2010). Consequences of differentiated leadership in groups. Academy of Management Journal , 53 , 90 106. Yun, S., Faraj, S., & Sims Jr, H. P. (2005). Contingent leadership and effectiveness of trauma resuscitation teams. Journal of Applied Psychology , 90 , 1288 1296. Zaccaro, S. J., Rittman, A. L., & Marks, M. A. (2001). Team leadership. The Leadership Quarterly , 12 , 451 483. Zhang, Z., & Pete rson, S. J. (2011). Advice networks in teams: The role of transformational leadership and members' core self evaluations. Journal of Applied Psychology , 96 , 1004 1017. Zhang, Z., Wang, M., & Shi, J. (2012). Leader follower congruence in proactive personali ty and work outcomes: The mediating role of leader member exchange. Academy of Management Journal , 55 , 111 130. subordinate outcomes: Testing the multilevel mediat ion role of empowerment. Journal of Applied Psychology, 97, 668 680. Zhou, L., Wang, M., Liu, Y., Penn, L. T., & Shi, J. (2013). Resources from supervisors to subordinates: Measurement development and theoretical integration. Poster presented at the 28 th A nnual Conference of Society for Industrial/Organizational Psychology, Houston, TX.

PAGE 220

220 BIOGRAPHICAL SKETCH Le Zhou received her Ph.D. from the University of Florida in the summer of 2014 . She earned her B.S. in psychology degree from the Peking University in China in 2009. She also earned a M.S. in Industrial Organizational Psychology from the University of Maryland, College Park in 2011.