xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001912433
007 cr mnu|||uuuuu
008 071009s2005 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001302
Stilson, Frederick R. B.
Does agreeableness help a team perform a problem solving task?
h [electronic resource] /
by Frederick R. B. Stilson.
[Tampa, Fla] :
b University of South Florida,
Thesis (M.A.)--University of South Florida, 2005.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 58 pages.
ABSTRACT: The relationship between mean team agreeableness and team performance has not been shown definitively. The present study was performed looking at archival data from a study that assessed team performance from 62 two person teams using the DDD and involving two types of training and two types of information probes during the computer task. In addition, each of the participants took a personality test based on the IPIP with an emphasis on agreeableness and its 6 facets. Using HLM analysis, it was determined that agreeableness does not have a significant effect on team performance for a problem solving tasks (delta chi square 2.04, p=n.s.), however it did significantly effect how an individual performed (delta chi square=18.06, p=.001) on the problem solving task. Intelligence had a significant effect on team performance (delta chi square=569.08, p=.001) and this may have washed out any personality effects. In addition, a linear regression indicated than none of the six facets of Agreeableness had a significant effect on team performance on a problem solving task.
Adviser: Michael D. Coovert, Ph.D.
t USF Electronic Theses and Dissertations.
Does Agreeableness Help a Team Perform a Problem Solving Task? by Frederick R. B. Stilson A thesis submitted in partial fulfillment of the requirements for the degree of Master of Arts Department of Psychology College of Arts and Sciences University of South Florida Major Professor: Michael D. Coovert, Ph.D. Carnot Nelson, Ph.D. Marcia A. Finkelstein, Ph.D. Date of Approval: September 9, 2005 Keywords: personality, big five, co mputer, simulation, critical thinking Copyright 2005, Frederick R. B. Stilson
Dedication This Masters Thesis is dedicated to my fam ily and friends, especially my parents, Fred and Sharon Stilson, who always told me that I could accomplish anything to which I put my mind.
Acknowledgments I would like to thank the follo wing people, without whom this thesis would have never been possible: Dr. Michael Coovert Dr. Carnot Nelson Dr. Marcia Finkelstein Tim Willis Matt Prewett Ashley Gray Rebecca Klein Katie Bush and Aptima
i Table of Contents List of Tables................................................................................................................. ....iii List of Figures....................................................................................................................iv Abstract....................................................................................................................... .........v Introduction................................................................................................................... .1-20 The Big Five Personality Factors.........................................................................2-3 Teams, Personality, and Comput er Simulated Tasks...........................................3-6 Relationship Between Team Learning and Team Performance..........................6-8 Agreeableness and a Shared Mental Model.......................................................8-10 Teams, Personality, and Back Up Behavior.................................................10-12 Why use a Computer Simulation for a Team Task?........................................12-16 The Current Study............................................................................................16-18 Hypotheses.......................................................................................................18-20 Method......................................................................................................................... 20-26 Participants.............................................................................................................20 Materials................................................................................................................20 Procedure.........................................................................................................2 1-27 Scoring...................................................................................................................25 Point Allocation.....................................................................................................25 Results........................................................................................................................ ..28-30
ii Data Analysis...................................................................................................2839 Power Analysis......................................................................................................39 Results for training conditions from the archival study.........................................40 Discussion....................................................................................................................4 2-45 Future Studies..................................................................................................44-45 References....................................................................................................................4 6-49 Appendices.........................................................................................................................50 Appendix A: Personality Questionnaire......................................................................51-53 Appendix B: Background Information..............................................................................54 Appendix C: Tactical Information...............................................................................55-58
iii List of Tables Table 1 Descriptive Statistics........................................................................................28 Table 2 ANOVA performed on the average team score and grouped by experimenter.............................................................................................. .29 Table 3 Mean and standard de viation of team score, level of Agreeableness and level of the six facets of Agreeableness ......................... 29 Table 4 Model Comparisons of da ta using HLM to predict team score........................31 Table 5 Model Comparisons of data using HLM to predict individual score...............31 Table 6 Beta weights associated with different variables in the HLM Model for team score........................................................................................32 Table 7 Summary of linear regression for Agreeableness Facets predicting team score ( N =62 teams).................................................................32 Table 8 ANOVA and Post Hoc test on CCT and Survival training conditions.......40-41
iv List of Figures Figure 1 Team learning orientation and performance......................................................7 Figure 2 Scatter plot of average team level trust and te am score with best fit line.......34 Figure 3 Scatter plot of average team level morality and team score with the best fit line....................................................................................................................35 Figure 4 Scatter plot of average team level altruism and team score with the best fit line....................................................................................................................36 Figure 5 Scatter plot of average team level cooperation and team score with the best fit line....................................................................................................................37 Figure 6 Scatter plot of average team level modesty and team score with the best fit line....................................................................................................................38 Figure 7 Scatter plot of average team level sympathy and team score with the best fit line....................................................................................................................39
v Does Agreeableness Help a Team Perform a Problem Solving Task? Frederick R. B. Stilson ABSTRACT The relationship between mean team Agreeableness and team performance has not been shown definitively. The present st udy was performed looking at archival data from a study that assessed team performance from 62 two person teams using the DDD and involving two types of training and two types of information probes during the computer task. In addition, each of the partic ipants took a personality test based on the IPIP with an emphasis on Agreeableness and its 6 facets. Using HLM analysis, it was determined that Agreeableness does not have a significant e ffect on team performance for a problem solving tasks ( =2.04, p =n.s.), however it did sign ificantly effect how an individual performed ( =18.06, p =.001) on the problem solving task. Intelligence had a significant effect on team performance ( =569.08, p=.001) and this may have washed out any personality effects. In addition, a lin ear regression indicated than none of the six facets of Agreeableness had a significant effect on team performance on a problem solving task.
1 In recent years, psychologists have used personality measurements like the International Personality Item Pool (IPIP) (Goldberg, 1999) to predict many different dependent variables. Psychologists in several different fields have studied personality using tools like the IPIP and then observing how personality relate s to everything from psychosis to job performance. Beginning in th e 1980s a structured approach to defining personality began to emerge: the Five Factor Model (FFM). This new formulation and the term FFM were coined by McCrae and Costa (1985). The FFM includes the personality dimensions of Openness to Experience, Conscientiousness, Extraversion, Agreeableness, and Neuroticism, commonly referred to as the acronym OCEAN. Conscientiousness, Extraversion and Neuroticism have been stud ied the most with fairly consistent and stable results emerging. In contrast to these three personality factors, Agreeableness and Openness to Experience are often over looked areas of the FFM. Specifically, Agreeableness, its facets, and team performance on a computer simulated task have been overlooked. To define team personality in th e current study, methods utilized in past studies, such as average sc ores (Neuman and Wright, 1999) and individual means (e.g. Heslin, 1964; Williams & Sternberg, 1988; Ba rrick, Stewart, Neubert, and Mount, 1998) were used. This study paired a personality test with a computer simulated task in order to help fill that knowledge gap in literature. A computer simulated task was chosen for this study because it allowed the participants to be presented with a novel situation. If one were to use a task one might encounter in the business world, a possible confound of having participants with speci fic training in that area of business may have arisen. In
2 addition to personality and computer simulati ons, the literature review for the current study will specifically cover the following topics: the Big Five Personality factors, findings involving team learning and personal ity and specifically how team learning and team performance are related, teams and back ing up behavior, why a computer simulated task is relevant, the relationship between t eam performance and team learning, and teams and the shared mental model. In addition, an overview of the ambiguous results involving Openness to Experience, Agreeableness and team performance is outlined and finally, the current study, one that examined individual and average team Agreeableness and its various facets and how they affected team performance on a computer simulated task, is discussed. The Big Five Personality Factors The concept of the Big Five Personal ity Factors was originally developed by Tupes and Christal (1961) and Norman (1963) This concept has been subsequently confirmed by Goldberg (1999) and McCrae an d Costa (1985). The Big Five Personality concepts consist of Openness to Ex perience, Conscientious, Extraversion, Agreeableness, and Neuroticism and together the five are commonly referred to as OCEAN. Openness to Experience involves being imaginative, curious, having broad interest and possibly going about life in an untraditional manner. Conscientious refers to one who is organized, punctual, ambitious and persevering. Extraversion is a trait marked by being sociable, talkative, f un-loving, and optimistic, which is in contrast to the fourth trait of Neuroticism. One who is classified as Neurotic is often worrying, nervous, and
3 insecure. Finally, Agreeableness is shown in someone who is softhearted, helpful, forgiving, and possibly gullible (Costa Busch, Zonderman, and McCrae, 1986). There have been several st udies that looked at some as pects of the Big Five and team performance. Very few, however, have looked at the effects of Agreeableness on team performance and none have looked at the relationship between these two variables as specifically and as in deta il as was done in the current study. The following is what is currently known about persona lity and team performance. Teams, Personality, and Computer Simulated Tasks A teams average personality score and th eir achieved results is an area of Industrial/Organizationa l (I/O) psychology that is gain ing momentum. This trend will continue as many corporations are increas ingly emphasizing teamwo rk. Psychologists are working on computer simulations that should be able to give generalizable feedback that will translate into real world success. Several studies have used team performance on a computer simulated task and a personality i ndex in order to look for relationship between the two. A portion of these studies like the one conducted by Colquitt, Hollenbeck, Ilgen, LePine, and Sheppard (2002), have met with success. The reason the Colquitt et al. (2002) study is being used as an example is that after an extensiv e computer literature search, no studies concentrating directly on Agreeableness and team performance on a computer simulated task could be found. Colquitt et al. (2002) conducted their study with a computer simulation called the Team Interactive Decision Exercise for T eams Incorporating Distributed Expertise
4 (TIDE). They also incorporated Costa and McCraes (1992) Revised Neuroticism, Extraversion, and Openness to Experience (N EO) personality inventory, with a specific emphasis on the aspect of Openness to Experi ence in order to look for a relationship between this personality dimension and team performance. They were successful in finding significant results between openness to experience and team performance ( r = .14, p <.05) using the computer simulated tas k. For additional information on the TIDE, refer to Hollenbeck, Ilgen, LePine, Colquitt, and Hedlund (1998) or Gigone and Hastie (1997). Another research team, Ellis, Hollenb eck, Ilgen, Porter, West, and Moon (2003) conducted a similar study to the one discu ssed above. Ellis et al (2003) utilized the Distributed Dynamic Decision-making (DDD) computer simulation. The TIDE discussed earlier is a derivative of the DDD. In their study, Ellis et al. (2003) focused on personality and team learning instead of team performance (The difference and relationship between team performance and t eam learning will be discussed later). For the Ellis et al. (2003) study, the experiment ers defined team learning as a relatively permanent change in the teams collective level of knowledge and skill produced by the shared experience of the team members (p. 822). Among the variables examined in their study were the effects of Openness to Experience and Agreeableness on team learning. Ellis et al. (2003) noted that agreeable i ndividuals tend to be compliant, self-effacing, modest, conforming, and non-confrontational, and though this may encourage team cohesion, but it may also detract from team l earning. Team learning may also be hindered because if the group as a whole scores high on Agreeableness they may be more likely to
5 reach a premature consensus on a course of action. If a group reaches consensus prematurely, this may lead the team to overlook several signifi cant steps of critical thinking that may have led to the team making a better deci sion. Their hypothesis on the relationship between Agreeableness and team learning is that: project teams with higher levels of Agreeableness will evidence lower levels of team learning (p. 823). Results of the Ellis et al (2003) study showed that Agreeableness correlated negatively with team learning on a computer simulation. This was consistent with Colquitt et al.s (2002) findings on team perf ormance mentioned earlier. In addition, Ellis et al. (2003) found that higher levels of Openness to Experience correlated with higher levels of team learning in a computer simula tion. Both of Ellis et al.s (2003) hypotheses regarding Openness to Experience and Agreeableness were supported. A reason given by Ellis et al. (2003) fo r the negative relationship between Agreeableness and team learning is that premature c onsensus, due to a lack of conflict, has a detrimental effect on both problem solv ing and decision making in groups. This is part of a phenomenon known as group think. In the current study, it is believed that a group high in Agreeableness may come to a premature c onsensus regarding task that require critical thought to be executed properl y. A definition used for group think is as follows: Group think is a phenomenon where alternative solutions to a problem are ignored due to an overassertive leader, an absence of diversity amongst opinions, or a group that has too much momentum in one direction (Janis & Mann, 1977). Group think often leads to a terrible final solution th at can end up causing anywhere from a minor inconvenience to death, as was the case with the Challenger Shuttle accident. Janis and
6 Mann (1977) proposed a model on the phenom enon of groupthink that supported the finding of Ellis et al. (2003) by including a facet of failure to re-examine preferred choice (p.132) as something that may eventually lead to group think. Moving from the concept of group think back to the concept of Agreeableness and teams, it is hypothesized in the current study that a curvilinear relationship exists between Agreeableness and team performance on a computer simulation. The reason for this hypothesis is that at the lower end of the Agreeableness spectrum teams will fail to agree on a solution where on the higher e nd, teams may agree too soon. In the next section, team learning and its relationship to team performance will be discussed since the concepts of team learning and team perfor mance are so closely re lated. A relationship must be established between the two in order to get a better idea of why Agreeableness seems to improve team performa nce, but hinder team learning. Relationship Between Team Learning and Team Performance Team learning is defined as a relative ly permanent change in the teams collective level of knowledge a nd skill produced by the shared experience of the team members (Ellis et. al. 2003, p. 822). Alternatively, team performance can have many definitions, but essentially refers to how we ll a team does on a given task. In theory it would seem that higher team learning would lead to better team perf ormance, but in order to clarify confusion in the field between the relationship of team learning and team performance, Bunderson and Sutcliffe (2003) did a study of Fortune 100 companies that looked at this specific concept and fo und some unexpected results. Employing
performance measures of actual profitability to targeted profitability (performance-to-plan) and actual profitability relative to units sold (profit-per-unit); Bunderson and Sutcliffe (2003) found that putting too much emphasis on learning may actually be deleterious to efficiency. 859095100105WeakModerateStrongTeam Learning OrientationPerformance/Plan HighPerformance-to-plan ModeratePerformance-to-plan LowPerformance-to-plan Figure 1. This is the predicted relationship between team learning orientation and business unit performance for different levels of team learning orientation. This is especially true if the team over emphasizing team learning had previously been performing well. Though counter-intuitive, when the results are graphed using performance-to-plan and three levels of team learning orientation (weak, moderate, and strong), a curvilinear relationship emerged with performance peaking around the moderate area of team learning orientation, as shown in Figure 1. These results suggest is that an overemphasis on team learning in a business setting may hinder performance. If one were to relate these findings to team Agreeableness, it might be hypothesized that an average level of Agreeableness for a team would foster both efficient team learning and team performance. If Agreeableness 7
8 levels were to fall substantially below or a bove the average level, there may be a drop off in team performance. This leads back to the hypothesis of the current study that was stated earlier of Agreeableness having a curvilinear relationship with team performance. In the next section Agreeableness and a shared mental model, which is important to successful team performance, will be disc ussed. A shared mental model essentially means that each team member has a similar picture in his or her mind of the information available and the best way to go about solving the given task. Agreeableness and a Shared Mental Model In a 1999 study by Neuman and Wright, it was determined that Agreeableness should help a group come to a consensus on a shared mental model (SMM), which they defined as a group conceptual ization of the environment a nd how to interpret it that transcends the cognitive approaches of th e individual (p. 379) Another interesting finding of the Neuman and Wright (1999) study was that Agreeableness in teams and supervisor task ratings (also referred to as task performance in th e study) were positively correlated, with r = .36, (p<.01). This is in contrast to the findings of the Ellis et al. (2003) findings that Agreeableness is detrimental to team performance. Another finding in contrast to the study done by Ellis et al. (2003) was that Agreeable individuals were more likely to be effective in group activiti es requiring coordination between the group members. Some of the reasons given by Ne uman and Wright (1999) for their findings were that group members high in Agreeableness are better at avoi ding disruptions at work that might be brought about by interper sonal conflict. They also mentioned that
9 Agreeableness should help a group come to a consensus on a SMM. It may be harder to draw definite conclusions from this study due to the subjective nature of supervisor task rating; however, this rating system may be more applicable to team work in organizations. Neuman and Wright (1999) we nt on to discuss the different facets of Agreeableness in their study (i.e. trust, straig htforwardness, altruism, compliance, modesty, and tender-mindedness). They stated that one would expect tender-mindedness, altruism, and trust to enhance interpersonal skills, thus al lowing organization members to relate effectively to others. Neuman and Wright (1999) went on to mention that compliance and straightforwardness should indica te sincerity in an individual and also signify willingness to work towards produc tive information-seeking and negotiation tactics. They did not indicate how the facet of modesty would relate, but one might posit that modesty would facilitate a team performing well. Unfortunately, Neuman and Wright (1999) only tested Agreeableness as a construct in thei r study and not any of the individual facets that make up Agreeableness. In the current study, the individual facets of Agreeableness and their effects on team performance on a computer simulated task are hypothesized and tested. Other studies that had similar results to Neuman and Wright (1999) are Bennet and Carbonari (1976) and Kilmann and Thom as (1975). These studies found that teams high in Agreeableness will have an easier time agreeing upon a shared mental model. In contrast to this opinion, the e xperimenters of the current stud y would argue that if a team is too collectively agreeable, then there ma y be an absence of the conflict that might stimulate the formation of the most optimal SMM. In the opposite di rection, a team very
10 low in collective Agreeableness may also struggle to form the optimal SMM because if one or more members of the team are unwilli ng to share their knowledge or information on a particular subject area important to the team, the team will be unable to make the best and most accurate decisions regarding a course of action (Klimoski & Mohammed, 1994). Teams, Personality, and Back Up Behavior McIntyre and Salas (1995) through their studies on teamwork, determined that there were four essential aspects to teamwork. These four aspects are Backing Up Behaviors, Closed-loop Communication, Performance Monitoring, and Feedback. We will concentrate on the concept of B ack Up Behavior in this section. Porter, Hollenbeck, Ilgen, Ellis, West, and Moon (2003) revisited the concept of backing up behaviors and included the Five Factor Model (FFM) personal ity types in a more recent study. Porter et al. (2003) returned to this concept of back ing up behavior because they felt backing up behavior may be the most critical aspect of teamwork given by McIntyre and Salas (1995). Backing Up Behaviors, as defined by Porter et al. (2003) means that team members will help each other to perform the task on which they are currently working. Some examples given are correct ing the mistakes of a fellow team member, or if a team member is unable to perform a certain task as signed to him or her, another team member will step in and fulfill the dut y. The specific definition used for the Porter et al. (2003) study was the discretionary provision of res ources and task-related effort to another member of ones team that is intended to help that team member obtain the goals as
11 defined by his or her role when it is apparent that the team member is failing to reach those goals (p. 391). The task that the participan ts were asked to perform in the Porter et al. (2003) study was set up using the DDD. This part icular variation of the DDD involved coordination of several military vehicles us ing Airborne Warning and Control Systems (AWACS). In the Porter et al. (2003) study, team members we re stationed in a common room in close proximity and used networke d computers, which is comparable to the current study. Teams in the study consisted of four individuals and in total there were 71 teams. In order to facilitate backing up be haviors, one team member out of the four (designated DM2, or Decision Maker 2), was gi ven a disproportionately heavy workload compared to the other three t eam members. Porter et al. ( 2003) did not specify whether or not this person was assigned randomly. As wa s done in the current study, Porter et al. (2003) administered the personality test to the participants before the DDD task was performed. They looked at all five of th e FFM personality type s in the study, but specifically and pertinent to the present study was their hypothe sis that states, in teams, provider Agreeableness and legitimacy of need will inter act in determining the amount of back up behaviors.(p. 395) The result of the Porter et al. (2003) study regarding Agreeableness was not what was expected. They found no significant effects for Agreeableness and legitimacy of need, which seems counter-intuitive. They did test for all six of Agreeableness facets, but only one, altruism, showed any effect regardi ng legitimacy of need. The relationship was in the expected direction, with higher altruism being a ssociated with higher back up
12 behaviors. A possible explanation give n by the authors is that perhaps the Agreeableness was indiscriminate in nature a nd applied in more of a blanke t approach, rather than being directed at the team member who truly n eeded the most help by being backed up, but upon post hoc examination, no evidence for this theory was provided. We do not specifically look at Back Up Be havior in this study, but futu re studies we perform on this subject will investigate this. Why use a Computer Simulation for a Team Task? The following section is dedicated specifi cally to computer simulations and their use for assessing team performance. Speci fically, the current study used a computer simulation called the Distributed Dynamic Decision-making (DDD) task. This allowed the experimenters to place the participants in an environment that they were probably not familiar with and test how well they did. Usi ng a task or environment with which most people are not familiar is usually chosen in orde r to test specific abilities in a laboratory. This is done so that there is little chance of outside prac tice effects becoming a nuisance variable. It should be noted th at it is possible to have co mputer simulations that are designed to mimic a work place such as an offi ce or a warehouse in order to better train employees. It should be noted that because a computer simulation is being used for experimental purposes and not training in the current study, the locati on of the Arctic was chosen instead of a familiar locale. This served to put the participants in an equally unfamiliar environment.
13 Specifically, the current study falls into a category referred to as Computer Simulations for Team Research (CSTR). A ccording to Rogleberg (1999), only about 10% of the studies done between 1996 and 1998 on team work utilized computer simulations. This will increase in the coming years as tec hnology allows more complex situations to be accurately modeled. In addition, the interf aces will become easier for participants to use and will only require a minimum amount of experience and training to be used. Marks (2000) pointed out that there are two types of computer simulations being used in teamwork research today. Simulations can be defined as situations created in order to place individuals in complex, dynamic, and malleable situations not easily created (p. 655). Simulations are critical for modeling real world team problems because they allow real life situations to be recrea ted without creating dange r to participants or equipment (Schiflett, Elliot, Salas, and C oovert, 2004). The first type of computer simulation is a task modeled on a real world si tuation. Examples of th is could be a flight simulation for pilots, a tank simulation for a ta nk team, or a stock market simulation for stock brokers. With these simulations, one can introduce predicaments such as an engine flame out on a jet, loss of night vision in a tank, or a stock market crash, without violating the principles of ethics or endangering a nyone. These specific simulations also allow a realistic crisis to be created without spe nding exorbitant amounts of money on the actual hardware that would be require d to obtain the same level of training results without a simulation. There is another type of computer simulation that differs from the first in that it does create a realistic e nvironment; however it is supposed to be unfamiliar to
14 participants in order to remove practice effects or other type of effects due to previous experience. The second type of computer simulation, and the type that will be used for the current study, is referred to as a Hypothes ized Nomological Net (M arks, 2000). This type of simulation is used to test the relationships that exist wi thin a team under a variety of situations. In the current study, Arctic Survival was chosen because it will most likely be unfamiliar to the participants, but still be able to stimulate critical thinking and allow assessment of how they work together as a team. For a computer simulation to be a valid way of assessing team performance, Raser (1969) laid out the following four criteria. The simulation must have psychological reality, process validity, stru ctural validity, and predictiv e validity. For psychological reality to be present the participants must believe that they are part of a real team: additionally, for the task(s) to be complete d successfully, the team must depend on each other and work together. If the participants do not believe that they are a part of a real team and that success requires collaborati on, then the results will not be valid. We attempted to induce this into the study by training the participants together and offering a cash prize to the team who obtained the highe st score on the task. The other types of validity process, structural, and predictiv e validity, must be induced differently. Process validity is achieved when the process one is attempting to test is present in the simulation. It is not possible to simula te all the real world details in a simulation but whichever facets are being tested must be present in the simulation. In our study we are looking at critical thinking and colla boration. These are both necessary to
15 successfully complete the Arctic survival computer simulation. This leads to structural validity. Structural validity is made possible by accurately representing the configuration of the real world or Nomologi cal-net model as it w ould be in reality. An example of the real world model having struct ural validity would be as fo llows: if one is trying to simulate an actual cockpit crew in a commercia l jetliner, one would have to simulate the responsibilities of the pilot, co -pilot, and navigator. If all three participants have control of the planes control surfaces (i.e. rudders, elevators, ailerons, etc.) and navigation equipment, then structural validity has been lost. If, in the case of a Nomological-net simulation, one is looking at le adership decisions it is necess ary to create a hierarchical team-member structure in addition to ma king the proper information and resources accessible to the leader (Marks, 2000). In the simulation used in the current study, each member only had access to a certain amount of limited resources, thus guaranteeing interdependence and cooperation in order to suc ceed. This leaves us needing to satisfy the criteria for predictive validity in order to realistically asses team performance. Predictive validity alludes to the simulation predicting the relationship that occurs in the reference system, which means that if a relationship is known to occur in reality, then it should also occur in the simulation (Marks, 2000). An example given by Marks (2000) is that if a business simulation indicates that stra tegic planning enhances team confidence levels, and one knows that these constructs are related through existing research, then this would present evidence fo r the predictive validity of that particular simulation. It is very difficult to generali ze simulations of this specificity and it is
16 important to validate each indi vidual simulation to a particul ar situation. In the current study, participants must cooperate and communi cate in order to be successful and obtain a high score. Therefore, the criteria for predictive validity are satisfied. In addition to the findings in the Marks (2000) study, a study done by Thompson and Coovert (2002) mentions that using com puters for a team task can influence many team decision processes, such as conforming to team norms (which working on the computer lessens) and equalizing the amount of participation by each group member. The Current Study So far we have discussed the history of personality a nd how we have arrived at the current number of five factors. We have also discussed what the five factors are and of what each one is comprised in order to distinguish it from the other factors. The specific focus of the personality discussion was based on Agreeableness because it and its six factors will be a main focus of the current study. Team performance and team learning were compared and contrasted and then two studies which had looked into both team performance and team learning using a computer simulation similar to the one being utilized in the current study were mentioned. In addition, the history and differe nt types of computer simulations were discussed and analyzed. The two different t ypes of computer simulations and the four types of validity that a computer simulation must possess in order to be considered valid were presented.
17 Also discussed were previous studies that involved teams, personality, and computers and their findings. Gaps in the literature were noted and c onflicting results of similar studies were brought to light. The curre nt study attempts to fill in some of the gaps in the literature and resolve the conflic ting results of previous studies or perhaps push the theory towards a certain direction. The following paragraphs discuss the current studys theory and methods in detail. Data for the current study were archival. The original data came from a study that looked at how either collaborativ e critical thinking (CCT) or survival training and either a presence or absence of information probes dur ing the task affected team performance. The experimenters also had the participant ta ke a personality survey based on Goldbergs (1999) International Personality Inventory Pool (IPIP) with 60 of the 100 items coming from the item pool for Agreeableness. It was the administration of the Agreeableness survey in the previous study that made the current study possible. In the current study we tried to de termine the relationship between the personality trait of Agreeableness and its effect on team perf ormance. Previous studies found mixed results, which is why the current study, with a specific focus on Agreeableness and team performance, has been proposed. Participants were placed into teams of three (two participants and one ad ministrator) and given a computer simulated task to perform as a team and their results we re objectively scored to determine how they performed. Specifically, the computer simula tion was the Distributed Dynamic Decisionmaking (DDD) task, which has been shown to be valid in several previous studies (Colquitt et al., 2002 & Ellis et al., 2003). This study utilized the Arctic Survival version
18 of the DDD. The Agreeableness of the participants was m easured using the IPIP. In addition, the six facets of Agreeableness were individually assessed. Our hypotheses for the study are listed in the following section. Hypotheses Following the discussion of how Agreeableness and its specific facets will affect team performance, these are th e hypotheses of the current study. Hypothesis 1. Teams whose average score is higher or lower than the mean on Agreeableness will perform worse in terms of cumulative team points scored than teams who score around the mean on the Agreeablen ess scale, therefore resulting in a curvilinear relationship. Due to the nature of personality an exact mean number for the Agreeableness score will not be known until after the comp letion of the study. Once the mean of the Agreeableness scores for all the participants is calculated, then a mean number for Agreeableness can be assigned. It is surmised that teams with a higher or lower level of Agreeableness than the mean will have a lower perf ormance score on the task than teams who are closer to the mean, because members of teams who are more agreeable, while being less combative may also be less critical of each other. Because they are less critical of each other, this may lead to more errors or a failure to perform to their optimal abilities. For the teams scoring lower than the mean on the Agreeableness scale, an inability to coordinate or agree will re sult in poorer performance on the task.
19 In addition, it is hypothesized that th e relationship with the six facets of Agreeableness will be as follows. These hypotheses are based on Ellis et al.s (2003) study in which a negative relationship between team learning and Agreeableness was found. In addition, these hypotheses are base d on Neuman and Wrights (1999) study where a positive relationship between Agreeableness and task performance was found and Bunderson and Sutcliffes (2003) study wher e more team learni ng did not lead to better team performance. Th e hypotheses on the facets of Agreeableness are as follows: Hypothesis 2 Trust will correlate positively with team performance because if the team members trust each other they will be more likely to cooperate and collaborate, therefore scoring higher. Hypothesis 3 Morality will correlate positivel y with team performance since treating ones teammate fairly and not wit hholding information or resources will be advantageous to the team. Hypothesis 4 Altruism will correlate positively w ith team performance because if one tries to do all the tasks alone then the score will suffer. Teams must be willing to share in the tasks equally or some will be left untended. Hypothesis 5 Cooperation will correlate positi vely with team performance because working together is part of being a good team. Hypothesis 6 Modesty will correlate negatively w ith team performance because if one or both team members are satisfied w ith scoring only a modest amount, then time and resources may be wasted tending to unimportant tasks.
20 Hypothesis 7. Sympathy will correlate negat ively with team performance because if one or both team members are too con cerned with giving orders that may lead to important tasks being uncomp leted, and time and resources may be wasted tending to unimportant tasks. These hypotheses are partially based upon Neuman and Wrights (1999) untested speculations on how the six facets of Agreeableness would individually affect a team. No studies could be found wher e the individual facets of Agreeableness were examined to determine if any or all of them enhanced team performance. In addition to the hypotheses on Agreeableness, the final hypothesis will be based on general mental ability (GMA) and team pe rformance. This was included to try and replicate the results of previous studie s that found a link between GMA and team performance such as Neuman and Wright (1999) among others. Our hypothesis is as follows: Hypothesis 8. Teams with higher GMA will score higher than teams with lower GMA.
21 Method Participants In this study we had 144 undergraduate psychology students from a large university in the southeast divided into 2 pe rson teams (72 total teams), who performed a task on the Distributed Dyna mic Decision-making (DDD) computer simulation. Of these 72 teams, 62 teams provided us with usable da ta. Age and ethnicity should not be a factor as our sample is assumed to be representa tive of the overall u ndergraduate population. Gender, however, may be a problem as our sample was comprised of 81% females. Although the exact effects of gender and teams are currently unknown, the gender makeup of the participant pool did not repr esent the overall populat ion. However, it did represent the population of the psychology de partment at the university where the experiment was conducted. Materials Two computers configured to run the DDD we re used to administer the task to the participants. Before participants began the task they viewed a se ries of Audio Video Interleaves (AVIs) running on a computer that demonstrated how to utilize the tools of the DDD. Several handouts that will be men tioned more specifically in the procedure were also utilized.
22 Procedure Each experimental session had a running time of three and a half hours with the actual DDD Arctic Survival task the participants were sc ored on running for 75 minutes. When the participants came into the lab, they were instructed to s it at the round table in the center of the lab. After brief introductions participants filled out informed consent forms and demographic sheets. Once this was finished, participants were instructed to draw a plastic tab out of a bow l held above the participants head. The tabs were marked either A or B and were used to determine whic h station the participant would sit at for the DDD task. After the participants drew the tabs we had them sit in front of a computer where the administrator gave them a power poi nt presentation on eith er Survival Training or Collaborative Critical Thinking (CCT) training. After the training presentations were finished we had the participants return to the round table in the middle of the lab and we in structed them to build a tower out of the tinker toys we provided. This exercise was done in order to facilita te the participants becoming more familiar with each other. The tink er toy task lasted for ten minutes and in order to make this task more interdependent for the participants, we instructed person A to only touch the joiners and person B to onl y touch the sticks. This way a tower could not be built by one person alone. Once the to wer building phase of the experiment was over, the tinker toys were put away and the pa rticipants then filled out the personality questionnaire which will be discussed later (Appendix A). After each of the participants had fini shed the personality questionnaire, we instructed them to sit at a computer located in the middle of the room where they viewed
23 the AVIs we had prepared. The AVI session was divided into thirds and encompassed all of the necessary training needed for the DDD. After each of the thre e sessions was over, we would give each team member five to seve n minutes to utilize what they had learned on a practice DDD scenario. This scenario wa s similar to the DDD scenario they would do for the experiment. In total, each team memb er received fifteen to twenty one minutes of actual time on the DDD in addition to the forty five minutes of AVI training. Once the last session of hands on training was complete and all of the participants questions about the DDD had been answered, the participants went back to the center table and the computers were set up to run the experi mental DDD scenario. Teams were given a background information page (Appendix B) to br ief them on what had theoretically taken place in the Arctic Survival scenario before th ey arrived. The participants were instructed that the goal of the task was to locate an unmanned aerial vehicle (UAV) and a lost team who was stranded. They were given a sheet with basic strategies and the point values for various tasks they could perform in the scenar io. This was titled tactical information and is located in Appendix C. The DDD Arctic Survival scenario is explained in more detail in the next section. In the Arctic Survival scenario ther e are four separately color coded team members. They are the red snow cat, the pur ple snow cat, an obser ver coded as green, and the blue fuel cat. The administrator played the blue fuel cat. Th e fuel cat only reacted to orders given by the participants. The green team member had the ability to see the entire map and everything that was going on in real time on the map. The reason for including the green team member in a scenar io that only required three people was that
24 the DDD Arctic Survival scenario was origin ally programmed to us e four people, three snow cats and one fuel cat. Since we we re only using two participants and one administrator, the green team member was reduc ed to an observer. If there had been an option to omit the green team member from the scenario that is the course of action we would have taken. To reiterate, the green team member did not have a way to participate in the game and was solely an observer, the DDD has the ability to run with three snow cats and a fuel cat, but for this scenario th e teams consisted of onl y two snow cats and a fuel cat. Green was fixed as an observer a nd no green snow cat icon was visible during the scenario. The red and purple participants did not know about the green team member. The blue team member or fuel cat, whic h was controlled by the administrator, only had the ability to refuel the red and purple team members. The administrator only used the blue fuel cat to refuel the red a nd purple team member when either the red or purple team member indicated that they need ed refueling by emailing the blue fuel cat and enabling their refueling icon. The refue ling icon could be seen by the blue team member if the blue fuel cat was within sensor range of the current snow cat in need of refueling. In addition, the partic ipants were instructed to direct any questions they had during the running of the DDD to the blue fuel cat. The red and purple team members, who were controlled by participants, were the only snow cats able to complete the different tasks of the scenario. Each red and purple team consisted of a snow cat, a medic, a tech nician, a scout, and a m echanic. All four of the personnel could be put onto a snow cat and transported to various locations. In addition, each of the personnel (medic, technici an, scout, and mechanic ) starts out with a
25 certain amount of usable units. For example, th e medic starts out with 15 medic units. If a task requires 3 medical units to complete a nd the red medic is assigned to a task that indicates it needs 3 medical units then after completing the ta sk, the red medic will have 12 medical units. This is similar for each of the other four color coded team members who each start with a certain amount of points. Each team, red and purple, had the same personnel and each of these personnel ha d the same amount of points as their counterparts on the other team (i.e. red techni cian had 15 units, so purple technician had 15 units). Communication between the red a nd purple participants and the blue administrator was only done electronically through a messenger system built into the DDD. Scoring for the DDD was recorded automatically by the computer and is explained in more detail below. In addition to receiving points for finding the crashed UAV or the lost team as they had been inst ructed to do, the part icipants could score points by completing such tasks as fixing a rusty drill, administering non-emergency medical assistance, or an a ssortment of other activities. Scoring: Scoring was divided into thr ee different areas. Objective scoring was accomplished by simply looking at the scores each team receives in accordance with the DDD. Each team had an opportunity to score points on their tasks. The point system is as follows: Point Allocation: 300 points: Find the unmanned aerial vehicle (UAV) or the lost team. 100 points: Render Emergency assistance (-100 from both te am members if emergency assistance is not rendered in the allotted time period) 50 points: Assist with re pair or medical requests 10 80 points: Process seismic monitors
26 Some of these tasks were available at the beginning of the scenario and others popped up at predetermined intervals and were relayed to the red and purple snow cats via the blue fuel cat played by the administ rator. The participants were told by the satellite messages relayed from the blue fuel cat where certain tasks were located. If the participants successfully co mpleted a task, they were awarded the above amount of points, depending on the task. There was no tim e limit to completing the tasks, with one key exception, which were emergencies. Emergency assistance had to be render ed within 15 minutes of when the participants received the e-mail alerting th em to the emergency. If the participants successfully administered emergency assistance in the given time, then they scored 100 points. If however, they were unsuccessful, they lost 100 points each. An anomaly in the programming for the DDD was that an emergenc y could be neutralized by processing it with only part of the needed resources (i.e the emergency requires 3 medical units, but the participant who attends to the emergency on ly has 2 medical units). In this case the team was neither penalized nor rewarded for attending to that part icular emergency and received 0 points. To clarify, if a team had a combined 300 points and one of the teammates attended to an emergency that re quired 3 medical units with only 2 medical units, they would still have 300 combined points afterwards instead of the 200 they would have had if they missed the emergency or the 400 they would have had if they had successfully attended to the emergency. Fort unately, this anomaly is believed to have happened only one time.
27 In order to determine a teams average level of Agreeableness each individual was administered the Agreeableness portion of the International Personality Inventory Pool (IPIP) (Goldberg, 1999) before his or her training on the DDD (Appendix A). The IPIP consists of 60 Agreeableness questions with 10 questions from each of the six facets of Agreeableness The other 40 questions were comprise d of the other four areas of the IPIP. The questions from the different fact ors were randomized in order to prevent priming for Agreeableness, as the IPIP was administered before the task. The two individuals scores on the IPIP were then combined and averaged to form an average team Agreeableness score. In addition to analyzing the average Agreeableness score, the scores of the individuals of each team were analyzed in order to determine if the variance within a team had any effect on team performance. To assess GMA, an area for SAT/ACT score was included on the demographics sheet. If a participant was unsure of his or he r SAT/ACT score, they were instructed to estimate it. In the case of estimation by a participant, a note was made on the demographics sheet. If the participant had neve r taken either of the tests or was unable to estimate their score, the space was left bl ank. After the study was concluded a formula was used to transform all SAT scores into an equivalent ACT score and all data used in the results section is based of the transformed scores.
28 Results Data Analysis All tables referred to in this section can be located at the beginning of the document or located through the table of cont ents. In our population of 62 teams, there were 24 males (19%) and 100 females (81%). Ta ble 1 lists the descriptive statistics for the study. Team scores ranged from -400 point s (teams who missed all four emergencies and completed no tasks) to 1420 points. Table 1 Descriptive statistics Variable Mean Standard Error Age Male Female 20.46 21.69 .643 .382 SAT/ACT conversion to ACT Male Female 24.43 24.37 .628 .460 An ANOVA was performed on the experimenters and the scores of the teams they administered the task to and was not significant ( F (4,61) = 1.030, p=.400) meaning that there was no experimenter effect (Table 2)
29 Table 2 ANOVA performed on the average team score and grouped by experimenter Sum of Squares df Mean Square F Sig. Between Groups 659371.7 03 4 164842.926 1.030 .400 Within Groups 9123049. 265 57 160053.496 Total 9782420. 968 61 Table 3 shows average team score, average level of Agreeableness and the average level of the six facets of Agreeableness. Table 3 Mean and standard deviation of team score, leve l of Agreeableness, and level of the six facets of Agreeableness Variable Mean Standard Error Team score 311.129 394.051 Agreeableness 2.91 .18 Trust 3.11 .15 Morality 2.29 .2 Altruism 3.45 .22 Cooperation 2.35 .25 Modesty 3.45 .15 Sympathy 2.90 .1
30 The original Hypothesis 1 stated that a curvilinear relationship between team average team Agreeableness and team score would be explored. The mean of average team Agreeableness was 2.91 ( SE =.18) and the range was only 1.483 with a small standard error. This indicated that there wa s a possible range restriction for the responses to the questions due to the Likert scale ra nging from only one to five. Due to the small amount of variability and small range in the team Agreeableness scores, a curvilinear analysis was deemed unfeasible. In addition, a linear regression analysis was also deemed unfeasible because we had individuals nested within teams. Therefore, the program MLWin and the method of Hierarchical Li near Modeling (HLM) were used for the analysis of Hypothesis 1. As noted before, individuals we re nested within teams and in addition to Agreeableness, some of the other level-1 vari ables were used to serve as predictors in the analysis were Intelligence Quotient (IQ), which was ascertained via SAT and ACT scores. Age, individual score on the task, gender and year in college were the other variables included. The le vel-2 variable was the team to which the individuals belonged. Table 4 shows the different models an alyzed using HLM to predict team score.
31 Table 4 Model Comparisons of data usi ng HLM to predict team score Model Variables Overall Deviance ( ) Change in Deviance Prob. A. Null 1836.05 B. Agreeableness 1834.01 2.04 Ns C. IQ 1266.97 569.08 .001 D. Individual score 1680.7 155.35 .001 E. IQ & Agreeableness Compare to agr. Compare to IQ 1265.3 570.75 568.71 1.67 .001 .001 ns F. Intercept and Condition 1816.88 19.17 .001 Table 5 includes HLM models predicting individual score and was included for comparison. Table 5 Model Comparisons of data using HL M to predict individual score Model Variables Overall Deviance ( ) Change in Deviance Prob. A. Null 1751.83 B. Agreeableness 1733.77 18.06 .001 C. IQ 1201.59 550.24 .001 D. IQ & Agreeableness 1200.38 551.45 .001 E. Intercept and Condition 1720.96 30.87 .001 Table 6 includes significant weights from the HLM equations.
32 Table 6 Beta weights associated with different va riables in the HLM m odel for team score Variable / (Sig. if / >2) Agreeableness 206.6 144.270 1.43 IQ 22.93 10.97 2.09 For Hypotheses 2 through 7, linear regression was used to analyze the data and the results along with those for Hypothesis 1 are summarized below. Test of Hypotheses Table 7 summarizes the results of the linear regression test of the hypotheses on the different facets of agreeableness. Table 7 Summary of linear regression for Agree ableness Facets predicting team score (N =62 teams) Variable Adjusted r B SE B Sig. Average Trust -.008 152.007 213.473 .092 .479 Average Morality -.003 -154.731 173.765 -.114 .377 Average Altruism .011 -297.186 229.384 -.165 .200 Average Cooperation -.008 -114.116 158.268 -.093 .474 Average Modesty .000 232.458 230.936 .129 .318 Average Sympathy -.017 -4.571 207.829 -.003 .983 Discussed below are the general findings on the facets.
33 Hypothesis 1. This hypothesis stated that teams whose average score is higher or lower than the mean on Agreeableness will perform worse in terms of cumulative team points scored than teams who score around the mean on the Agreeableness scale. Our results showed that Agreeableness does not affect how a team performs in terms of points scored on a computer simulated task ( =2.04, p >.05)., but Agreeableness does affect how an individual performs on the task ( = 18.06, p=.001). Hypothesis 2. Trust will correlate positively with team performance because if the team members trust each other they will be more likely to cooperate and collaborate and therefore score higher. The re lationship between average leve l of trust and team score was not significant ( F 1, 61=.021, p=.884 ).
5.004.003.002.001.00 avgtrust 150010005000-500 team score R Sq Linear = 0.008 Figure 2. Scatter plot of average team level trust and team score with the best fit line. Hypothesis 3. Morality will correlate positively with team performance since communication is limited to e-mail only and this being a timed task, succinctness and being direct will be advantageous to the team. The relationship between average level of morality and team score was not significant (F 1, 61=1.747, p=.191). 34
5.004.003.002.001.00 avgmoral 150010005000-500 team score R Sq Linear = 0.013 Figure 3. Scatter plot of average team level morality and team score with the best fit line. Hypothesis 4. Altruism will correlate positively with team performance because if one tries to do all the tasks alone then the score will suffer. Teams must be willing to share in the tasks equally or some will be left untended. The relationship between average level of altruism and team score was not significant (F 1, 61=2.947, p=.091). 35
5.004.003.002.001.00 avgalt 150010005000-500 team score R Sq Linear = 0.027 Figure 4. Scatter plot of average team level altruism and team score with the best fit line. Hypothesis 5. Cooperation will correlate positively with team performance because working together is part of being a good team. The relationship between average level of cooperation and team score was not significant (F 1, 61=1.71, p=.196). 36
5.004.003.002.001.00 avgcoop 150010005000-500 team score R Sq Linear = 0.009 Figure 5. Scatter plot of average team level cooperation and team score with the best fit line. Hypothesis 6. Modesty will correlate negatively with team performance because if one or both team members are satisfied with scoring only a modest amount then time and resources may be wasted tending to unimportant tasks. The relationship between average level of modesty and team score was not significant (F 1, 61=.000, p=.985). 37
5.004.003.002.001.00 avgmodest 150010005000-500 team score R Sq Linear = 0.017 Figure 6. Scatter plot of average team level modesty and team score with the best fit line. Hypothesis 7. Sympathy will correlate negatively with team performance because if one or both team members are too concerned with giving orders that may lead to important tasks being completed, then time and resources may be wasted tending to unimportant tasks. The relationship between average level of sympathy and team score was not significant (F 1, 61=.488, p=.488). 38
5.004.003.002.001.00 avgsymp 150010005000-500 team score R Sq Linear = 8.061E-6 Figure 7. Scatter plot of average team level sympathy and team score with the best fit line. Hypothesis 8. Teams with higher GMA will score higher than teams with lower GMA. This, like team and individual Agreeableness, was computing using HLM and this hypothesis was supported ( =569.08, p=.001). Power Analysis Determining power for a multilevel model is a complex process that is still lacking a definitive method of determination. However, there is a general consensus 39
40 among the researchers using multilevel models that even with a sample size of around 30, an estimate of residual error at the lowest level (level-1), is still very accurate (Hox & Maas, 2002, Barcikowski, 1981). Recall for this experiment, that our sample size was n=124. Results for training conditions from the archival study The data from the type of training we we re trying to prime, CCT or Survival, was not directly related to this st udy, but was a main focus of the archival study that we drew the data from to do our analysis. There were no hypotheses directly re lated to the training data in this study, but we deci ded to present the results in Table 8 in order for other researchers to draw their own conclusions a bout how the training data may relate to the hypotheses in this study. In Table 8, it is eviden t that there is a signi ficant difference in scores between the five different conditions, and the post hoc test indicates that this significant difference is between condition 1 (no CCT or Survival training, no probes) and condition 5 (survival training, probes.) Table 8 ANOVA and Post Hoc test on CCT and Survival training conditions Sum of Squares df Mean Square F Sig. Between Groups 1531226. 126 4 382806.532 2.644 .043 Within Groups 8251194. 841 57 144757.804 Total 9782420. 968 61 Dependent Variable: Score Tukey HSD
41 (I) Condition (J) Condition Mean Difference (I-J) Std. Error Sig. 95% Confidence Interval Lower Bound Upper Bound 1 2 -42.714 146.181 .998 -454.51 369.08 3 -214.714 146.181 .587 -626.51 197.08 4 -94.603 151.583 .971 -521.61 332.41 5 -421.548(*) 137.682 .027 -809.40 -33.70 2 1 42.714 146.181 .998 -369.08 454.51 3 -172.000 170.152 .849 -651.32 307.32 4 -51.889 174.814 .998 -544.34 440.56 5 -378.833 162.908 .152 -837.75 80.08 3 1 214.714 146.181 .587 -197.08 626.51 2 172.000 170.152 .849 -307.32 651.32 4 120.111 174.814 .958 -372.34 612.56 5 -206.833 162.908 .711 -665.75 252.08 4 1 94.603 151.583 .971 -332.41 521.61 2 51.889 174.814 .998 -440.56 544.34 3 -120.111 174.814 .958 -612.56 372.34 5 -326.944 167.772 .304 -799.56 145.67 5 1 421.548(*) 137.682 .027 33.70 809.40 2 378.833 162.908 .152 -80.08 837.75 3 206.833 162.908 .711 -252.08 665.75 4 326.944 167.772 .304 -145.67 799.56 The mean difference is si gnificant at the .05 level.
42 Discussion The purpose of this study was to determine if Agreeableness and its individual facets had a specific effect on team performa nce. What we have determined from our experiment is that there was not support for Agreeableness predicting team score, however there was support for Agreeableness predicting individual score. Our finding that Agreeableness predicts individual sc ore and not team score is perplexing, but this may be due to the DDD task being designe d to keep track of team scores, but inaccurately reporting individual scores. I am confident however that the team scores reported in the data are accurate as I witnessed some of the scores reported in the data being the administrator for several experiments. If there were any errors, I believe they took place in the reporting of the individual scores. Therefore the original hypothesis was not affected and the results showed that Agreeableness did not have a significant effect on team score. Another possibi lity for the lack of team Agreeableness level predicting team score may be due to communication only being allowed via the e-mail system and never occurring face-to-face during the task. Not surprisingly, results were also non-significant for all of the hypotheses on the individual Agreeableness facets. It should be noted th at the range of overall team Agreeableness was only between 2.3 and 3.783 with most scores falling around the middle of the scale ( M =2.91). This range restriction prevented using a curvilinear
43 regression and compelled us to use HLM in or der to extract more information from the data. Perhaps using a ten-point Likert scal e in future studies would allow for more variability and give more information about th e personality aspects of the participants. In contrast to this, it may reproduce the results found in this study and provide more evidence for the lack of a curvilinear relationship between Agreeableness and team performance. Information on age, gender, and ye ar in college were also used in the HLM analysis even though there were no specific hypotheses constructed for these variables and their effects can be found in Table 4. Our hypothesis for GMA leading to better team performance was supported. This finding was consistent with past studies lik e Hollenbeck et al. (2002) and Neuman and Wright (1999) where intelligence had a signifi cant effect on predicting team score. Our results may indicate that the task was in tellectually demanding and this may have affected our results, serving to wash out any personality aspects. It may also mean that the effects of personality pale in compar ison to how GMA affects team performance. Looking at the correlation between GMA and Agreeableness shows a slightly positive, but non significant rela tionship (r=.113, n.s). Some limitations of this study could have been that the task was fairly complex and that while some of the participants seem ed to pick up the task fairly easily, others clearly struggled. Better results might have been obtained by using a simpler task. Fatigue may have been a factor since the e xperiment took three a nd a half hours and the task lasted for 75 minutes. Apathy or lack of intrinsic motivation may have also played a part in the large number of par ticipants who failed to score hi ghly. It should also be noted
44 that even given the chance many participants would not ask questions to the administrator or consult their quick reference guide for assistance. In addition, being unfamiliar with a computer may have hindered a pa rticipants score. This is unlikely with the prevalence of computers in the classroom t oday, but still a possibility. Training differences were not a fact or as an ANOVA was performed on team scores and the experimenters who administ ered the task to them and it was not significant. Future studies For future studies, having more teams ma y be helpful to determine if there are really no specific eff ects for the facets of Agreeableness. In addition, specific hypotheses about general intelligence, e xperience with computer simula tions, and other aspects of personality may be looked into and tested. Also, adding additional members to the team should lead to more variance amongst the Agreeableness within the team. This may lead to a clearer picture of why Agreeableness seems to affect individual performance, but not the overall team performance. Regarding the participants who failed to ask for help or search for aid in their quick reference guide, it may actually be more beneficial to have the quick reference guide present during tr aining. All questions could be redirected towards the quick reference guide in order to prime more self reliance in the participants. However, this priming of self reliance may be detrimental to team performance due to a shift of focus to the individual. Perhaps a study using the two different training methods could be of interest. It would also be interesting to test teams who are currently working
45 together in the work place and perhaps l ook at the most successful teams and work backwards. In conclusion, this study was another important step in determining the relationship between an individuals personal ity and overall team performance. While we did not succeed in finding support for our mean Agreeableness hypotheses, we were able to test the individual facets of Agreeableness and their effects on team performance. Unfortunately, the hypothesis involving the in dividual facets were also not significant. We did find more support for higher levels of intelligence lead ing to better team performance. This may mean that intelligence is the one factor that must be present for team success.
46 relative-ga 34, 1127-1 Business U Colquitt, J. A., Hollenbeck, J. R., Ilgen, D. R, Lepine, J. A., & Sheppard, L. (2002). he Moderating y, 87(2), 402Costa, P. T., Bu 1986). Correlations of References Barcikowski, R. S.(1981). Statistical power with group means as th e unit of analysis. Journal of Educational Statistics, 63, 267-285. 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(3), 377-391. Bennet, R. P., & Carbonari, J. P. (1976). Pe rsonality patterns related to own-, joint-, and in maximizing behaviors. Journal of Personality and Social Psychology, 134. Bunderson, J. S. & Sutcliffe, K. M, (2003). Management Team Learning Orientation and nit Performance. Journal of Applied Psychology, 88(3), 552-560. Computer-Assisted Communication and T eam Decision-Making Performance: T Effect of Op enness to Experience. Journal of Applied Psycholog 410. sch, C. M., Zonderm an, A. B., & McCrae, R. R. ( MMPI Factor Scales With Measures of the Five Factor Model of Personality. Journal of Personality Assessment, 50(4), 640-650.
47 Psychologi measuring Deary, F. D l. 7 (pp. 7-28 Heslin, R. (1964 r personality and its effects on group problemHollenbeck, J. R nd expertise in Hox, J. J. & Ma Costa, P. T., & McCrae, R. R., (1992). Normal personality assessment in clinical practice: The NEO Personality Inventory. Psychological Assessment 4(1), 5-13. Ellis, A. P. J., Hollenbeck, J. R, Ilgen, D. R ., Porter, C. O. L. H, West, B. J., & Moon, H (2003). Team Learning: Collectively Connecting the Dots. Journal of Applied Psychology, 88(5) 821-835. Gigone, D., & Hastie, R. (1997). Proper anal ysis of the accuracy of group judgments. cal Bulletin, 121, 149-167. Goldberg, L. R. (1999). A broad-bandwi dth, public domain, personality inventory the lower-level facets of several five-factor models. In I. Mervielde, I e Fruyt, & F. Ostendorf (Eds.), Personality Psychology in Europe Vo ). Tilburg, The Netherland s: Tilburg University Press. ). Homogeneity of membe solving. Journal of Abnormal and Social Psychology, 62 248-256. ., Ilgen, D. R., LePine, J. A., Colquitt, J. A., & Hedlund, J. (1998). Extending the multilevel theory of team d ecision making: Effects of feedback a hierarchical teams. Academy of Management Journal, 41, 269-282. as, C. J. M. (2002). Sample sizes for multilevel modeling. In: J.Blasius, J Mann, L. (1977). Decis Janis, I., & ion Making: A Psychol ogical Analysis of Kilm r as orts, 37, 971-980. Conflict, Choice, and Commitment. New York: The Free Press. ann, R. H., & Thomas, K. W. (1975). In terpersonal conflict ha ndling behavio reflections of Jungian personality dimensions. Psychological Rep
48 Journal of Marks, M. A., (2000) A Critical Analysis of Computer Simulations for Conducting Team McCrae, R.R, & Adequate Taxonomy: Journal of Per Ability. Jo Norman, W. T., Abnormal oon, H. (2003) B imacy of Need. Jo Rogelberg, S. (1 meeting of Klimoski, R. & Mohammed, S. (1994). Team Mental Model: Construct or Metaphor? Management, 20(2), 403-437. Research. Small Group Research, Vol. 31 (6), 653-675. Costa, P. T. (1985) U pdating Normans Intelligence and Personality Dimensions in Natural Language and in Questionnaires. sonality And Social Psychology. 49(3) 710-721. McIntyre, R. M., & Salas, E. (1995). Meas uring and managing for team performance: Lessons from complex environments. In R.A. Guzzo & E. Salas (Eds.), Team effectiveness and decision-m aking in organizations, 9-45. Neuman, G. A., & Wright, J. (1999). Team Effectiveness: Beyond Skills and Cognitive urnal of Applied Psychology, 84(3), 376-389. (1963). Toward and adequate taxonomy of pers onality attributes: Replicated factor structure in pe er nomination personality ratings. Journal of and Social Psychology, 66, 574-583. Porter, C. O. L. H., Hollenbeck, J. R., Ilgen, D. R., Ellis, A. P. J., West, B. J., and M acking Up Behaviors in Teams: The Role of Personality and Legit urnal of Applied Psychology, 88(3), 391-403. Raser, J. C., (1969). Simulations and society: An exp loration of scientific gaming. 999) Simulations for the st udy of teams. Symposium conducted at the the Society for Industrial and Organizational Psychology, Atlanta, GA
49 Developme Thompson, L. F. & Coovert, M. D. (2002) Stepping Up to the Challenge: A Critical Group Dyn Tupes, E. C., & Christal, R. E. (1961). Recurrent personality factors based on trait Williams, W. M better than Schiflett, S. G., Elliott, L. R., Salas, E., and Coovert, M. D. 2004. Scaled Worlds: nt, Validation, and Applications. Abingdon: Ashgate Publishing Ltd. Examination of Face-to-Face and Com puter-Mediated Team Decision Making amics: Theory, Research, and Practice, 6(1) 52-64. ratings (USAF ASD Tech. Rep. No. 61-97). San Antonio, TX: Lakeland AFB. ., & Sternberg, R. J. (1988) Group intelligence: Why some groups are others. Intelligence, 12, 351-377.
51 Appendix A Personality Questionnaire Session ID: ___________ A B Listed below are phrases that describe people's be havirs. Please use the rating scale to describe how o accurately each statement describes you Describe yoself as you generally are now, not as you wish to be ur in the future. Describe yourself as you honestly see yourself, in relation to other people you know of the same sex as you are, and roughly your same age. So t you can describe yourself in an honest manner, hat your responses will be kept in absolute confidence. Pl ease read each statement carefully, and then circle the number to the right of the question. Veraccurate y In Moderately Inaccurate Neither Inaccurate nor Accurate Moderately Accurate Very Accurate 1 I trust other s. 1 2 3 4 5 2 I believe tha t others have good intentions. 1 2 3 4 5 3 I trust what people say. 1 2 3 4 5 4 I believe tha t people are basically moral. 1 2 3 4 5 5 I get angry e asily. 1 2 3 4 5 6 I like to solv e complex problems. 1 2 3 4 5 7 I distrust pe ople. 1 2 3 4 5 8 I suspect hidden motives in others. 1 2 3 4 5 9 I am wary ofothers. 1 2 3 4 5 10 I believe that people are essentially evil. 1 2 3 4 5 1 1 I w ould never cheat on my taxes. 1 2 3 4 5 12 I worry about things. 1 2 3 4 5 13 I fear for the worst. 1 2 3 4 5 14 I am afraid of many thin gs. 1 2 3 4 5 15 I know how to get around the rules. 1 2 3 4 5 16 I cheat to get ahead 1 2 3 4 5 17 I feel comfortable with myself 1 2 3 4 5 18 I prefer to stick w ith things that I know. 1 2 3 4 5 19 I can handle a lot of information 1 2 3 4 5 20 I obstruct others' plan s. 1 2 3 4 5 21 I make people feel welcome. 1 2 3 4 5 22 I anticipate the needs of others. 1 2 3 4 5 23 I love to help others. 1 2 3 4 5 24 I love action. 1 2 3 4 5 25 I believe in human goodness 1 2 3 4 5 26 I think that all will be well. 1 2 3 4 5 27 I act wild and crazy. 1 2 3 4 5 28 I am hard to get to know. 1 2 3 4 5 29 I have a lot of fun. 1 2 3 4 5 30 I take no time for others. 1 2 3 4 5 31 I would never go hang gliding or bungee jumping. 1 2 3 4 5 32 I go on binges 1 2 3 4 5
52 Appendix A (Continued) Ve ry Inaccurate Mo derately Inaccu rate Ne ither Inaccu rate nor Accura te Mo derately Ac curate Ve ry Accura te 33 I hate to seem pushy. 1 2 3 4 5 34 I have a sharp tongue. 1 2 3 4 5 35 I contradict others. 1 2 3 4 5 36 I amuse my friends 1 2 3 4 5 37 I stick to the rules. 1 2 3 4 5 38 I use flattery to get ahead. 1 2 3 4 5 39 I use others for my own ends. 1 2 3 4 5 40 I hold a gr udge 1 2 3 4 5 41 I do not have a g ood imagination. 1 2 3 4 5 42 I talk to a lot of different people at parties 1 2 3 4 5 43 I laugh aloud. 1 2 3 4 5 44 I take advantage of others. 1 2 3 4 5 4 5 I b elieve that I am better than others. 1 2 3 4 5 46 I think highly of myse lf. 1 2 3 4 5 47 I have a high opinion o f myself. 1 2 3 4 5 48 I seldom joke aroun d 1 2 3 4 5 49 I find it difficult to a pproach others 1 2 3 4 5 50 I make myself the c enter of attention. 1 2 3 4 5 51 I like a leisurely lifestyle 1 2 3 4 5 52 I don't like crowded events. 1 2 3 4 5 53 I am concerned a bout others. 1 2 3 4 5 54 I have a good word for everyone. 1 2 3 4 5 55 I look down on others. 1 2 3 4 5 56 I take charge. 1 2 3 4 5 57 I love a good fight. 1 2 3 4 5 58 I yell at people. 1 2 3 4 5 59 I insult people. 1 2 3 4 5 60 I get back at others. 1 2 3 4 5 61 I like to tidy up. 1 2 3 4 5 62 I have a vivid imagination 1 2 3 4 5 63 I complete tasks successfully. 1 2 3 4 5 64 I excel in what I do. 1 2 3 4 5 65 I dislike talking about myself 1 2 3 4 5 66 I consider myself an average pers on. 1 2 3 4 5 67 I seldom toot my own horn. 1 2 3 4 5 68 I am not intereste d in th eoretical discussions 1 2 3 4 5 69 I like to visit new pla ces. 1 2 3 4 5 70 I can manage man y things at the same time 1 2 3 4 5 71 I have difficul ty understanding abstract ideas. 1 2 3 4 5 72 I know the answers to many questions. 1 2 3 4 5 73 I boast about my virtues. 1 2 3 4 5
53 Appendix A (Continued) Ve ry Inaccurate Mo derately Inaccu rate Ne ither Inaccu rate nor Accura te Mo derately Ac curate Ve ry Ac curate 74 I avoid philosophical discussions 1 2 3 4 5 75 I am indiffere nt to the feelings of others. 1 2 3 4 5 76 I make people feel uncom fortable. 1 2 3 4 5 77 I turn my back on others. 1 2 3 4 5 78 I liste n to my conscience. 1 2 3 4 5 79 I dislike being the center of attention. 1 2 3 4 5 80 I put pe ople under pressure. 1 2 3 4 5 81 I pretend to be concerned for others. 1 2 3 4 5 82 I am not bothered by mess y people. 1 2 3 4 5 83 I am easy to satisfy. 1 2 3 4 5 84 I can't stand confrontations. 1 2 3 4 5 85 I plunge into tasks with all my heart. 1 2 3 4 5 86 I s eldom daydream. 1 2 3 4 5 87 I go straight for the goal. 1 2 3 4 5 88 I sympathize with the homeless. 1 2 3 4 5 89 I feel sympathy for those who are wors e off than myself. 1 2 3 4 5 90 I feel others' emotions 1 2 3 4 5 91 I get others to do my duties 1 2 3 4 5 92 I value cooperation over competition. 1 2 3 4 5 93 I suffer from others' sorrows. 1 2 3 4 5 94 I am not interested in other peopl e's 1 2 3 4 5 95 I tend to dislike softhearted people. 1 2 3 4 5 96 I do not enjoy going t o art m useums 1 2 3 4 5 97 I believe in an eye for an eye 1 2 3 4 5 98 I try not to think about the needy. 1 2 3 4 5 99 I believe people should fend for themselves. 1 2 3 4 5 100 I can't stand weak people. 1 2 3 4 5
54 Appendix B Background Information The U.S. Military has been testing a new Unmanned Aerial Vehi cles (UAV). During a snow storm the UAV was blown o ff course and contact was lost It was determined that your team is the closest to the last known c oordinates of the UAV. One team has already been sent out, but is now missing. You have b een asked to help locate the lost team and recover the UAV. Good luck. Antarctica is a continent covered with nearly 14 million square miles of ice. You and the other teams are at a political ly neutral a ffectionately known as Station Blue, a research site speciali zing in geological activity and in the Earths ozone layer. There are at most a few hours of ime temperature is minus 13 egrees Fahrenheit, a 6-25 mph wind blows snow constantly, and visibility is usually less an 0.6 miles. y effect in some places the wind can blow a sheet of ow and ice over crevices. Vehicles and individuals can be trapped or lost when they of supplies, has a top speed of 6 mph, but at its cruising speed of 4.8 ion ve Antarctic Site daylight, the average dayt d th Station Blue is located 30 miles inland on an ice sheet at an altitude of 4,600 feet. Eighteen miles further inland, northeast of the station, is a mountain range with peaks as high as 11,000 feet. The terrain around the sta tion is low undulating hi lls. To the east, there are canyons and bluffs formed by huge cracks and displacements of the ice sheet. The Antarctic wind has a deadl sn break through the sheet. In other places, the wind can fo rm a natural bridge strong enough to support vehicles. Lastly, the wind can suddenly create bli nding storms that can reduce visibility to zero and lower the temp erature to minus 103 degrees Fahrenheit. Station Blue has given your team permission to use two of its three snow cats. Specially designed to navigate Antarcticas terrain, th e snow cat can carry your team members and a limited amount mph they have a range of only 30 miles. Snow cats are equipped with communicat and navigation equipment and a set of special sensors and probes. In addition, the snow cats are connected to geostationa ry satellites that can provi de information about weather and geological events. Unfortunately, weather and terrain may interfere with satellite transmissions and disrupt radio communication with Station Blue. You decide to use the snow cats to try to replicate the path of the lost team. The researchers have placed seismic monitors around the area these m onitors would ha recorded the vibrations of a snow cat driving pa st it. There is a team at Station Blue to help access the supply depots and gas depots, and to provide guidance on the terrain. Since they have continual access to the sate llite, they might be able to give you new information.
55 Appendix C Tactical Information STATION BLUE AND ITS RESOURCES Seismic monitors: Seismic monitors usually examine the ice sheet for geologic activity, but they can indicate when a vehicle has pa ssed and in which direc tion it was traveling. To obtain this information, you must process the seismic monitor with the correct leve and combination of resources. You can appl y more than the required resources but n less. If the m l ot onitor requires more supplie s than you have, you can request another team join you and combine resources to obtain the message. Seismic monitors take one to three steps to process. For example, three-st ep monitors need to be repaired, prepared, and then analyzed for information. Tw rs need to be prepared and analyzed. One-step monitors only need to be analyzed. aypoints: Certain seismic monitors yield key information, which needs to be read for ic monitors in each scenario are designated a drilling machine. Technical resources nable the repair of circu itry in a computer or a dig ital computer chip. Scouting on u. It also returns formation about sectors where no clues were detected. Naturally, it provides weather to o-step monito W the mission to be successful. Five or six seis m as waypoints because they lie on the path taken by the fourth team. You will need to process these seismic monitors in order to uncover the path to the lost team. The closer a monitor is to the path of th e fourth team, the higher the po int value of a monitor. A monitor directly on the path of the fourth team will earn you 80 points; a monitor furthest away from the path of the fourth team will yield 10 points. Processing the Monitors: Monitors require the following resources: mechanical, technical, and scouting. Mechan ical ability refers to the materials needed to repair moving parts in a machine, like a vehicle or on e resources assist with the interpretation of th e encrypted data on the monitors or at open locations. Resources can be depleted, but th ey can be replaced at specific locations. Some tasks may require more resources th an you may presently possess. In those instances, you may have to reque st assistance from another team and pool your resources. Satellite: The satellite provides relatively hi gh-quality information and may indicati the location of man-made objects or other geol ogical clues to assist yo in and terrain information. Clues: Clues provide information that can help you in your search and usually lay on the ice. You do not need to apply any re sources to get a message from a clue. TASKS
56 errain Tasks, Rough Terrain and Clearing Blocked Terrain. You are operating in a will be able to clear the blockages; in other cases you must ke a different path. Each vehicle has multiple units of terrain-clearing resources. Those us areas. sks You may encounter people who need m echanical help to repair equipment or machinery that has broken do l consume resources, they may ation e.g. ng youre searching in the rong area. s Appendix C (Continued) T hostile environment that requires you to make decisions about which path to take. Hazardous terrain includes crevasses, m ountains, and passes blocked by snow or avalanches. Hazardous terrain will slow your progress. Blocked terrain will stop your progress. Sometimes you ta resources are depleted with use but can be re stocked at specific supply depots. Blue will receive information from the satellite concerning the locati on of the hazardo Repair ta wn. While th ese tasks wil Its Cold Out Here meani yield useful inform w Time-critical Emergency Tasks Occasionally Station Blue will call you to render critical emergency assistance to another team. These are time-sensitive emergency task that may have life-threatening consequences if you do not help. If you dont respond to Station Blues requests for help, your level of communication with Station Blue may be affected. You will lose 100 points each if you do not attend to these emergency tasks within the allotted time N on-critical medical tasks: Other medical requests may occur that are not critical or time sensitive, but your assistance may be rewarded with information. These tasks require medical personnel, which you may need anyway. I advise you to carry a medic for your own health. Refueling: If your vehicle runs ou t of fuel everyone on your ve hicle perishes! You can refuel at several fuel depots or via the m ovable fuel tanker. To do this you need to communicate with Blue and request refuel assistance. Your remaining fuel can be monitored. Restocking: Your vehicl onsumed by processing: seism e starts out with a fini te number of resour ces. Resources are ic monitors, me dical, repair, and emergency tasks. These unicate with Blue and request assistance. c re sources can be replenished at one of several resource depots or at home base. Note that each REMOTE resource depot has resources to restock only one vehicle and is only accessible at certain times. The home base is always available for replenishing resources and has no limit on restocking capabilities. In order to replenish your supplies, you will need to comm
57 Appendix C (Continued) SCORING core Displ S sn ay: You will be able to see your indivi dual score and the scores of the other 2. Attending to Emergency ta You earn 100 points if you attend to specified time; you lose 100 points if you dont. ow cats, if they are out there. BLUE will see the team score, which will be the sum of the scores from RED and PURPLE. Your score reflects four factors: 1. Points received after pro cessing a seismic monitor, The number of points you receive from processing a seismic monitor indicates how close you are to the lost partys path (10 points = far away, 80 points = very near). If you receive 0 points after processing a seismic monitor either someone has already pr ocessed it or you made an error in processing. sks an Emergency task within the 3. Proce ssing repair and medical tasks If any task is processed by two snow cats simultaneously both players will receive the points for that task. Remembera al units and 3 c If, for instance medic and a te task will disap n 4. Recovering the UAV or the Lost Party. tht to receive points on a task re quiring coordinated e fforts (e.g. 3 medic tehnician units), you must have them both process the task simultaneously. the medic finishes processi ng the task that requi res the use of both a chnician before the technician is instructed to process the task, then the pear and you will receive no points. All necessary personnel must be begi processing the task before the first person fi nishes processing or you will not be given credit f or completing the task. Point Allocatio n : 30 0 points: Recover the UAV. 300 point s: Recovering the lost team. 100 points: e Render emergency assistanc -1 erg ency task within the allotted time 00 points: Failure to responded to em 50 points: Assist with re pair or medical requests 10 80 points: Process seismic monitors (h igh points = close to lost partys path) Important points to remember: Make sure you attend to the emergencies within the time allotted! It is important to maintain f uel levels and to refuel when necessary. Your snow cat will be immobile if fuel drops be low 150 pounds. If your vehicle runs out of
58 veryo ill perish and you will be out of the game. You lso m re sources on your snow cat to be able to s the s ber t ea and medical/repair tasks. Remember, if the task requires more expertise than you have on board your cat, you can request an other team to help by combining their rces (i.e. medic fuel, e ne on your vehicle w must a aintain sufficient personnel proces eismic monitors and other tasks Remem o load personnel onto your snow cat, as this will allow them to travel faster Apply at l st the required amount of res ources to complete the seismic monitor resources with yours. Do not forget to fill up at the supply de pots if you run low on resou units, technician units, mechanic units, etc.)