Master thesis - University of Tilburg [PDF]

Master thesis How team-efficacy and collective regulatory focus influence team risk taking

Name:

Mirelle Vink

Supervisor: R. Leenders ANR:

309112

Date:

15-08-2014

Background information Details of student Name:

Mirelle Vink

ANR:

309112

Supervisors First reader:

R. Leenders

Second reader:

H. van Dijk

Title of the Master Thesis Circle:

Who Dares?

Title of the thesis:

How team-efficacy and collective regulatory focus influence team risk taking

Definitions Team-efficacy (TE)

People’s shared beliefs in their collective power to produce desired results

Collective

Regulatory How groups of people approach pleasure and avoid pain in

focus (CRF)

different ways.

Risk

The extent to which there is uncertainty about whether potentially significant and/or disappointing outcomes will be realized

Team risk taking (TRT)

A team’s deliberate decision of which the expected outcomes are uncertain and the goal difficult to achieve

Precision of information The specificityy of information about a risk’s pervasiveness, (PI)

potency, or effects

Abstract This research studies the effects of team-efficacy, collective regulatory focus and preciseness of information on team risk taking. Nowadays organizational teams are getting more and more autonomous and are therefore expected to cope with quite some responsibilities. With these responsibilities, often complex decisions have to be taken. But is every team up to this? Risk taking as shown by individuals is a topic widely researched. Risk taking shown by teams however is a more uncharted territory. The competencies of a team depend on several characteristics of the individual members and their dynamics among each other. This same is expected to happen with team risk taking. With help of 51 organizations that were willing to participate, this thesis describes an experimental research that was devoted to unravel parts of team risk taking. A precisely developed decision game was designed by seven students. The game consists out of four cases with organizational dilemmas all trying to challenge the participating teams to show their level of risk taking. The four cases where divided in two categories, two cases with precise information and two cases with imprecise information. The scores are finally presented to the team in the form of a fictive performance that is linked to a donation to KWF Kankerbestrijding. All data collected with help of these cases and three questionnaires that where completed by the participants, was analyzed using SPSS. The research shows that there is a positive effect of collective regulatory focus on team risk taking, meaning that teams with higher collective regulatory focus scores are more willing to take risks. Also a mediating effect of collective regulatory focus is shown between the relation of team-efficacy and team risk taking. The other four hypotheses show a nonsignificant effect.

Preface With this preface I want to thank everyone that made it possible to write this thesis. First of all I would like to thank all of the organizations that were willing to join in on our research. Without these teams the research would have been impossible to perform. Next, I would like to thank Roger Leenders and Hans van Dijk for their help and feedback on the thesis. Last but not least, I would like to thank the other circle members, as what we all deliver is actually quite a good team performance on its own. Everyone did their utmost best to perform for the team and to make it to an enjoyable activity. The complete process of writing this thesis was big lesson to me. I think that when performing an experiment again, there will be a lot of things that I will remember from this first time. All together I think it was a good experience!

Keywords Team-efficacy, collective regulatory focus, team risk taking, precision of information, experiment, organizational teams, scenario game, organizational science

Index 1. Introduction ............................................................................................................................ 1 1.1 Research problem ............................................................................................................. 1 1.2 Research question ............................................................................................................. 2 1.3 Relevance.......................................................................................................................... 2 2. Theoretical framework ........................................................................................................... 3 2.1 Risk and Team Risk Taking ............................................................................................. 3 2.2 Team-efficacy ................................................................................................................... 3 2.3 Collective Regulatory focus ............................................................................................. 4 2.4 Precision of information ................................................................................................... 6 2.5 Conceptual model ............................................................................................................. 7 3. Methodological framework .................................................................................................... 8 3.1 Research design ................................................................................................................ 8 3.2 Sample .............................................................................................................................. 8 3.3 Data collection .................................................................................................................. 8 3.4 Preliminary data analysis ................................................................................................ 11 3.4.1 Control variables ...................................................................................................... 11 3.4.2 Team-efficacy .......................................................................................................... 12 3.4.3 Collective regulatory focus ...................................................................................... 12 3.4.4 Team risk taking ...................................................................................................... 14 3.4.5 Quality indicators ..................................................................................................... 14 3.4.6 Power analysis ......................................................................................................... 15 4. Results .................................................................................................................................. 17 4.1 Correlations .................................................................................................................... 17 4.2 Moderated mediation analysis ........................................................................................ 17 4.3 Evaluation of hypotheses ................................................................................................ 18 5. Conclusion and recommendations ....................................................................................... 21 6. Discussion and limitations ................................................................................................... 23 References ................................................................................................................................ 25 Appendices ............................................................................................................................... 27 Appendix 1: Questionnaire control variables ....................................................................... 27

Appendix 2: Questionnaire Team-efficacy........................................................................... 27 Appendix 3: Questionnaire Collective regulatory focus ...................................................... 28 Appendix 4: Experiment protocol ........................................................................................ 29 Appendix 5: Experiment logistics ........................................................................................ 39 Appendix 6: Vignettes .......................................................................................................... 40 Appendix 7: Calculations of experiment .............................................................................. 44 Appendix 8: Reliability analysis TE ..................................................................................... 48 Appendix 9: Principal component analysis CRF .................................................................. 49 Appendix 10: Reliability analysis CRF ................................................................................ 51 Appendix 11: Within group agreement CRF ........................................................................ 51 Appendix 12: Reliability analysis TRT ................................................................................ 52 Appendix 13: Output correlations ........................................................................................ 55 Appendix 14: Moderated mediation analysis ....................................................................... 56 Appendix 15: Syntax preliminary analysis TE ..................................................................... 58 Appendix 16: Syntax preliminary analysis CRF .................................................................. 58 Appendix 17: Syntax preliminary analysis TRT .................................................................. 60 Appendix 18: Syntax correlations ........................................................................................ 61 Appendix 19: Syntax multicollinearity analysis ................................................................... 61

1. Introduction 1.1 Research problem As organizations grow, more and more different departments with different responsibilities and goals arise. Also the financial crisis is pushing organizations into a situation where everything has to work as efficient as possible. By cutting out staff functions due to cuts in the organizations budget, organizations need to be able to rely on the teams within the organization. They often need to work as autonomous and independent as possible. This way of coping with the budget cuts might have big consequences if a team is not up to it or if it’s not being supported enough. One task that is often seen with many teams is that they should be able to take decisions together. These decisions can be about the dividing of tasks, projects, investments or about certain risks. Risk taking is an important element of decision making within organizations because of its possible consequences for performance (Knight, Durham, & Locke, 2001). Some organizations thrive on risks taken in the past, while others will do about everything to avoid taking risks. But what makes organizational teams take risks? This research tries to explain why some phenomena trigger team risk taking. Every individual or group of individuals will respond in a somewhat different way to risk. This is probably the reason why risk taking is a quite often researched concept. A lot of researchers have tried to unravel why individuals take risks, but risk taking at team level is an area that is relatively unexplored. This research will focus on team risk taking, as shown by existing teams in an organizational setting. But what characteristics of a team make it more or less risk taking? The answer to that might (partially) be found in the concept of team-efficacy. When a team thinks it can have a positive influence on a specific task, it is likely that they will be more risk taking because of their confidence in their decision making, or executive performance (Krueger & Dickson, 1994; Knight, Durham, & Locke, 2001). When looking at research about risk taking and team-efficacy a word often found is ‘goal seeking’. These concepts are to a certain extend related. A phenomenon related to goal seeking is regulatory focus. Regulatory focus theory provides a framework to examine different strategies for approaching positive outcomes and avoiding negative outcome (Florack & Hartmann, 2007). Regulatory focus can be divided in two forms, being; prevention-focus and promotion-focus. In research on individuals all concepts have been connected more than once. But are they also Page | 1

related when they are at team level? Another concept that is an important part of risk taking is the amount of information that is available to the team on the decision they should make. Dahlstrom, Dudo and Brossard (2012) show that the completeness of information can have quite and influence on how people respond to risks. Therefore this concept will also be incorporated into this research. All together this research tries to create a conclusion, based on theoretically developed hypotheses that are being tested in a situation made to be as realistic as possible. 1.2 Research question Based on this, the research question will be: To what extent does team-efficacy influence collective regulatory focus and how do both affect team risk taking? How does the precision of information affect the relationship between collective regulatory focus and team risk taking? 1.3 Relevance This research tries to contribute in such a way that it researches several phenomena in a setting that is aimed to be as much realistic as possible. The research is unique that sense that it investigates influences on team risk taking. Team risk taking is an area that is quite unknown in the present literature. Team risk taking is a difficult and dynamic phenomenon that can have serious influences on the performance of a team. Gaining insight on this specific phenomenon might give organizations a chance to anticipate on it. Next, self-efficacy is a quite well researched phenomenon, but team-efficacy (also known as collective-efficacy) is to a lesser extent. Because the big difference in the individual and team form of efficacy and the applied characteristics of this study, it might be a valuable contribution to literature. Also this research is unique in the way of testing team-efficacy as in most researches the level of teamefficacy is manipulated while this research is interested in real levels of team-efficacy. This also applies to regulatory focus that is often manipulated while this research measures an existing level. Next to theoretical relevance this research tries to give practical recommendations that might be helpful for organizations on how to best manage their teams concerning risk taking. Page | 2

2. Theoretical framework 2.1 Risk and Team Risk Taking Sitkin and Pablo (1992) define risk as "the extent to which there is uncertainty about whether potentially significant and/or disappointing outcomes will be realized” (p. 10). Therefore risk taking is always a tradeoff between the possible gains and the possible losses of the risk to be taken. Whether a person is willing to take a risk often depends on if he or she is risk seeking or risk averse. Risk is often seen as a negative thing. It appears that when asking people if a risk is a distribution of negative or positive outcomes, 80% of the respondents says it is negative (March & Shapira, 1987). This research tries to approach risk taking as much neutral as possible. When looking at a classical way of interpreting risk, it should be measured by the probability of a possible gain or loss. Though when looking at how risks are handled by e.g. managers, it shows that they often do not trust, understand or use probabilities connected to the risk (March & Shapira, 1987). Rather they seem to be basing they decision on the outcome of the risk, like the opportunity or worst case scenario. This research aims at risk taking as displayed in organizational teams and defines this team risk taking (TRT) as: A team’s deliberate decision of which the expected outcomes are uncertain and the goal difficult to achieve (Tjosvold & Yu, 2007; Valacich, Sarker, Pratt, & Groomer, 2009). Because a team exists out of several individuals, all with their own risk preference, TRT is a highly dynamic concept. Team members can strengthen each other (and therefore the team) by sharing knowledge, interacting and developing confidence, though conflict can also help a team to solve a problem. When a team has to decide upon a certain risky situation, their motivation toward reaching a certain goal will be one of the main motivators. Another team phenomenon closely related to motivation and to confidence in being able to influence outcomes is teamefficacy. This phenomenon will be explained next. 2.2 Team-efficacy Team-efficacy (TE) is defined by Bandura (2000) as people’s shared beliefs in their collective power to produce desired results. This phenomenon is actually the team version of selfefficacy. Just as an individual can have a certain believe in power to influence outcomes, groups can too. Some outcomes might not even be realistic for an individual to reach by himor herself. The use of a group of people (in this case a team) can bring a certain dynamic, enabling knowledge sharing, interaction and coordination. TE is not simply a sum of all the Page | 3

group members’ self-efficacy. It might be that team members perform well individually, but that they cannot work well together. When this leads to low performance by the team it might lower the TE beliefs. Bandura (1997) explains that team-efficacy influences future achievements strived for by the team, their consistency to a certain goal, the effort they are willing to take for a task and how they handle discouragements. TE reacts to feedback received on a certain task or achievement. This means that as a team receives positive feedback on a certain decision, the team-efficacy will be strengthened while failing a certain task might weaken it. Krueger and Dickson (1994) found that self-efficacy has an effect on risk taking of individuals. They state that as the self-efficacy of an individual goes up, their idea of understanding the possible outcomes increases, leading to more risk taking. Also research was executed on these phenomena on team level. Knight, Durham and Locke (2001) state that teams that have high confidence in their ability to achieve a certain goal will be more likely to take risk, because they believe that they can implement the task successfully. Therefore the following is hypothesized: Hypothesis 1: Team-efficacy has a positive effect on team risk taking When looking at the research of Knight Durham and Locke (2001) it shows that confidence might have a mediating role between TE and TRT. A phenomenon that is also expected to have this role is collective regulatory focus. Therefore this concept will be discussed next. 2.3 Collective Regulatory focus Collective regulatory focus (CRF) is the team form of regulatory focus (RF) which is a phenomenon that tries to answer why people prefer one action over the other or why they participate in a certain group action instead of a different one (Stekelenburg, 2006). The theory of RF says that people all act differently when approaching something positive or avoiding something negative. When having a promotion focus, one would for example try to pursue a goal that is advancing or that is a certain accomplishment. Prevention focus is more of a way to protect or safe-guard something (Florack & Hartmann, 2007). It can be said that a person having a promotion focus, sets different priorities than a prevention focused person will. Table 1 shows an overview of some characteristics of promotion- and prevention focus.

Promotion-focus 

Nurturance needs

Prevention-focus 

Security needs

Page | 4



Growth and advancements goals



Protection and responsibility goals



Self-direction and stimulation values



Tradition and conformity values



Eagerness



Vigilance



Openness to change



Need for stability



Success = gain, failure = no-gain



Success = no loss, failure = loss



Approach



Avoidance

Table 1: Promotion and prevention foci (Stekelenburg, 2006)

Just as a person can have a prevention or promotion focus, a team can too. The main difference between RF and CRF is that when a team should make a decision, it is subjected to several goals and values of different team members. CRF can also be seen as some kind of group norm. For example the group norm might be striving for advancement. An article by Keller (2006) shows an effect between RF and efficacy. This article states that as a person has a more promotion focus, he or she is more eager to reach the goal. This is exactly the other way around of how this research will approach the effect. This research tries to determine whether there is an influence of TE on CRF considering the task specific characteristics of TE. The theory is expected to work in such a way that as a team received positive feedback on certain actions, its TE rises, pushing the team into a more promotion focused intention. Or more simply said, as the teams’ confidence in a specific task rises, they will be more approaching towards reaching goals instead of avoiding them. Therefore the following is hypothesized: Hypothesis 2: Team-efficacy has a positive effect on collective regulatory focus In a paper by Higgins (2002) the influence of RF on risk taking is being discussed. In this article it is stated that people who have a promotion focus will see promotion relevant outcomes as more important. Thus in a risk situation they might be more focused on the possible gains of the risk. Florack and Hartmann (2007) in their research find that groups that are prevention focused are more risk averse. This research tries to find if the same phenomenon will occur at team level; meaning that as a team is more promotion focused it will pay more attention to the possible positive outcomes of a decision, making them take more risky decisions. Therefore the following is hypothesized: Hypothesis 3: Collective regulatory focus has a positive effect on team risk taking Based on the previous three hypotheses, a mediation effect is stated, saying that as the efficacy of a team rises, the team becomes more promotion fo cused, leading to more risky Page | 5

decisions taken by the team. This effect is expected to be a partial mediation effect because also a direct effect between TE and TRT is hypothesized. Therefore the following is hypothesized: Hypothesis 4: Team-efficacy has a positive effect on team risk taking, through its positive effect on collective regulatory focus. In hypothesis three it is stated that RF has an influence on TRT because teams with high RF will be focused on possible gains instead of losses. But what happens to this relation when the information needed to make this consideration is imprecise? Next the concept of information precision will be discussed. 2.4 Precision of information Precision of information (PI) in this research will be determined as the specificity of information about a risk’s pervasiveness, potency, or effects (Dahlstrom, Dudo, & Brossard, 2012). Dahlstrom, Dudo and Brossard show in their article that there are two ways people can respond to the difference in precise and imprecise information. On the one hand, people might have a higher risk perception when information is precise. This could lead to more risk taking because they have the idea that they understand the risk to a high level. On the other hand it might work the opposite way because people see all the components of the risk and therefore overestimate it. When looking at imprecise information, this contrast also holds. When having imprecise information a person might overestimate the risk because of vagueness leading to less risk taking but it might also start ignoring the risk because of the vagueness making it more risk taking. In hypothesis three it was explained that as teams have a promotion focus they will focus more on the positive side of the risk, making them more risk taking. When this effect is influenced by imprecise information it is expected to cause an overestimation of the risk, lowering the possible gains. Because the gains are the motives for the promotion focused teams to take the risk, the overestimation will probably cause the team to make a less risky choice in comparison to a precise information situation. Therefore the following is hypothesized: Hypothesis 5: The positive effect of collective regulatory focus on team risk taking will be weaker when imprecise information is applied.

Page | 6

When combining all of the five hypotheses as described above, there will be a moderated mediation effect, also known as a conditional indirect effect. This summary of all expectations can be stated as following: Hypothesis 6: The positive effect of team-efficacy on team risk taking is partially mediated by collective regulatory focus and the effect of collective regulatory focus on team risk taking is moderated by preciseness of information. 2.5 Conceptual model

Figure 1: Conceptual model with expected effects

Page | 7

3. Methodological framework 3.1 Research design The hypotheses in this research will be tested by a quantitative approach with data gathered via an experiment. The research has a deductive nature, meaning that it works from a theory to a confirmation. The data will be measured at one specific point in time making it a crosssectional analysis. In this research, both individual and team data was collected. Regarding the TE and TRT measure the unit of observation and analysis are both the organizational team. The RF measure uses the individuals as units of observation and the organizational team as its unit of analysis. All specific methods are further defined in chapter 3.3. 3.2 Sample The data collection was based on convenience sampling among the population of Dutch organizations. There were no requirements regarding the industry the organization is in because of an expected scarcity of organizations willing to participate. The experiment was designed in such a way that it is applicable for all kinds of industries. With regards to the teams, the requirement was that it was an existing team instead of a temporary team. Also the teams should exist out of three to 15 team members (Jones & Bearley, 2001). Due to problems with feasibility and limited resources the sample size of the teams was expected to be relatively low. Due to this possible low sample size, a post-hoc power analysis will be conducted to estimate the chance of false conclusions. The aim is to reach a power of at least 0.8 (Pallant, 2005). The final sample consists out of 218 individual participants and 49 organizational teams. The average team size is six team members with an average age of 38 (ranging from 18 to 64). The sample consists out of 44,5% (97) of male and 55,5% (121) of female participants. The average time the team existed in its current form was 22 months (as indicated by the teams). In the sample exists out of teams spread over 10 different industries of which most in the healthcare section. 3.3 Data collection In order to gather all the necessary data for this research an experiment was developed. Next to the experiment, three separate sets of questionnaires were given to each participant in order to gain a part of the data. The control variables, TE and CRF where measured with help of the questionnaires and TRT was measured with the experiment. The experiment was executed by Page | 8

the teams without knowing the exact research objectives. The information they were given was that the experiment was about decision making. For each team the set-up was as follows: 1. Thesis circle members arrive 30–60 min before start of experiment. 2. Introduction: Welcoming the team, explaining who we are and what we are going to do. 3. First questionnaire (A) with 59 items. 4. Explanation of the experiment. 5. First round of the experiment, with two of the four vignettes. 6. Second questionnaire (B) with 30 items. 7. Second round of the experiment, with the last two of the four vignettes. 8. Third questionnaire (C) with 47 items. 9. Presenting the results, thanking the team for participating and cleaning the room. The control variables were measured in questionnaire A, the TE was measured with 5 items in questionnaire B and the CRF was measured with 14 items in questionnaire C. All the questions relevant to this research can be found in appendices 1, 2 and 3. TRT was measured during the two rounds and four vignettes of the experiment, while differences in the vignettes allowed the manipulation of PI to be measured. The questionnaire The control variables, TE and CRF will be gathered through a short set of questions. For TE the decision has been made to let the team as a whole decide on their TE instead of using measures of the team members as individuals even though this way of measuring has been questioned by some (Bandura A. , 2000). The reason why this form is being questioned is because of fear for pressure for conformity and power differences among team members influencing the ‘true TE’ (Bandura A. , 2000). Though, when this is the case, this same effect is expected to be present when the team takes the decision on the risk to be taken. Therefore this consistency should make the measurement more reliable. TE will be measured in a task specific way. The task they are establishing their TE on will be the decision process as used in the experiment. First the team will play round one and two in order for them to gain insight on the task. Next the team is asked to fill in the questions (5 items) with regards to the upcoming two (comparable) rounds. The scale used to measure TE, is based on a scale designed by Bandura (2006) with regards to measuring collective efficacy to promote reading. Page | 9

For CRF the decision was made to use the existing shortened version of the RFQ by Stekelenburg (2006). This questionnaire is based on proverbs that either stand for promotion or for prevention focus (e.g. ‘Nothing ventured, nothing gained’ and ‘Act normal, that’s crazy enough’). The choice for the use of the shortened version was made with regards to the time available for the whole experiment. Because of time reasons it appeared at the test-teams that filling in the questions of the RFQ as a team was taking too long. Therefore the choice was made to make the questionnaire individual ones, were the team members had to answer the questions with regards to the team. This means that the team members were asked to answer to what extend they thought each proverb was applicable to their team. The experiment The thesis circle created an experiment in order to measure TRT. Though ultimately created by the circle, the game was also inspired by Shupp and Williams’ (2008) lottery evaluation experiment where participants allocated money between risky decisions, and by Wallach, Kogan, and Bem’s (1962) experiment where participants chose between different risk scenarios (like expansion and new projects). By providing teams with realistic, organizational decisions and conducting the experiment in-house, we try to facilitate a decision-making process and a level of risk-taking that is identical to their real-life situation. Additionally, we donate a monetary reward to the KWF Kankerbestrijding based on their performance in the experiment, hopefully increasing their dedication and seriousness to the experiment. Also the teams where offered an evaluative report based on our observations and the comparison with other teams. The experiment took approximately one hour to execute, and in every experiment a detailed word-for-word protocol was followed (appendix 4). The logistics in the form of necessary materials and room set-up is shown in appendix 5. The number of circle members present varied based on team size, but was between two and four persons, consisting of one game host and one or more observers (watching maximum three participants each). An excel sheet calculated the team results, and with that the basis for the KWF-donation. This program randomly decides the win/loss situations in each scenario and presents the final team winning/loss considering the allocations as make by the team (calculations can be found in appendix 7). Later this number was divided by 20.000 to settle each team’s contribution before averaging all the teams’ results to calculate the final donation – a total of 410 euros. Explanation of the experiment Page | 10

The experiment is based on a vignette study, where teams are presented with four consecutive decision-making scenarios; two providing them with imprecise information and two providing (more) precise information. In each scenario, teams were asked to allocate 1 million Euros between two investment options, where the risky one has a lower probability of success but a higher amount of gain/loss and the safe option has a higher probability of success but a lower amount of gain/loss. This separation is based on the definition of risk used in the thesis. For measuring the moderating variables, the “precise information” conditions have exact numbers for the probability percentages and gain/loss values (for instance 40% chance of success); whereas the “imprecise information” conditions have intervals for each (for instance 30–50% chance of success). The full vignettes are listed in appendix 6. The scenario-order was randomized from team to team, to rule out a sequence effect on the results. The alternatives also had equal expected value in each scenario, since “Virtually all theories of choice assume that decision-makers prefer larger expected returns to smaller ones” (March & Shapira, 1987, p. 1404). The calculations behind the numbers can be found in appendix 7. We consistently used a gain frame in the scenario texts. Then, through discussion, the teams must decide how to allocate the money between the two options. TRT is measured in the experiment as the proportion of one’s budget allocated to a risky option (i.e. an option with lower probability of success, but with the possibility of higher gains) rather than to a safer option. By taking the amount allocated to the risky option, a continuous risk score between 0 and 100 arises. Thus, if a team invests 500 000 (out of 1.000.000) in the risky option they get a risk score of 50. Subsequently, the average across the two scenarios of one precision-level was calculated in order to get team risk score. 3.4 Preliminary data analysis For the analysis of the data SPSS version 20, Excel 2010 and the fourth edition of the SPSS survival manual by Pallant (2005) was used. First all of the data from the questionnaires, observations sheets and allocation sheets from the experiment were entered into SPSS. This way a file was created containing all of the data for all of the members of the group (TEAM_DATA.SAV). Next a file was created containing only individual data (Individual_data.sav). In this individual file contains the measures of TE, CRF, TRT and several control variables. Because all variables have a different way of measurement the steps that are taken are described per variable.

3.4.1 Control variables Page | 11

Team age was established by taking the mean of the age as filled in by all of the participants within the team. Team tenure was established by taking the team’s indication of the existence of the team in its current composition. This measure has the limitation that it was difficult for the team to answer as it was sometimes difficult defining the boundaries of the team. The team size was determined by the team in years.

3.4.2 Team-efficacy Reliability analysis The TE measure is build up out of 5 team level items. The scale is analyzed with a reliability analysis which found that TE_5 has no variance at all and was there for removed. Cronbachs alpha is .743 which it above .7 and therefore acceptably reliable. When looking at the Cronbachs Alpha if Item Deleted (.807), it shows that removing TE_4 would heighten the reliability of the scale. Because of this the highest and lowest items are removed due to low variance, is was decided to remove both TE_5 and TE_4. A possible explanation for the low variance in TE_5 and TE_4 is that they are quite extremely negative. When a team would have said they were unsure to reach a result of 1-2 or 3-4 this might have felt like failing. The team cohesion might therefore withhold the members from giving this answer. Next the mean of the items TE_1, TE_2 and TE_3 was used to aggregate to one TE measure. The complete output of the reliability analysis can be found in appendix 8.

3.4.3 Collective regulatory focus CRF is measured with 14 items which can be divided into two sub-scales. No items where recoded because no negatively formed items where used. Principal component analysis The RFQ is one that, based on literature, should exist out of two sub-scales. With the use of a principal component analysis (PCA) it is possible to test whether these two sub-scales are also visible in the collected data. In order to see if the scales are suitable for PCA it is needed to assess if the Kaiser-Meyer-Olkin Measure of Sampling Adequacy (KMO) value is above .6 and that the Bartlett’s test of sphericity is significant. The KMO value is .811 and Bartlett’s test of sphericity is .000 meaning it is significant. With this the conclusion can be drawn that the scales a suitable for PCA. The rotation used for the PCA is Direct Oblimin. The Total Variance Explained shows that there are three components with an Eigenvalue of 1 or more explaining 49% of the variance. Also the scree plot shows a change in the shape of the plot Page | 12

after the third component. When looking at the loadings in the component matrix, it shows that not all items load as they are supposed to. The first seven items in the component matrix show to promotion focus and the last seven show the prevention focus. Also it shows that there are four items that load on a third component though two are negative and they all are quite low. Because of this discrepancy between the outcomes of the PCA and the theoretical construct of the scales a reliability analysis will be executed in order to see if items should be removed. All output can be found in appendix 9. Reliability analysis When analyzing the reliability of the items originally belonging to the promotion scale, it indicates no negative values in the Inter-Item Correlation Matrix. Cronbach’s alpha indicates .789 meaning the internal consistency reliability is acceptable. The Corrected Item-Total Correlations are all above .3 meaning the items are all measuring the scale they are supposed to. Also for none of the items the Cronbach’s alpha if item deleted is higher than the final alpha value meaning no item has to be removed based on this analysis. When checking the same for the prevention scale it shows there are two items (‘voorkomen is beter dan genezen’ and ‘onder zorgen gebukt gaan’) that have a Cronbach’s alpha if item deleted that is higher than the final alpha value (.707). Even though this reliability analysis shows it might be relevant to remove two items from the prevention scale, the decision is made to retain the original items and number of components. When deviating from it, it will not be possible to compare the findings of this research to findings of other researches using this scale. This is decided as more valuable. All output on the reliability analysis can be found in appendix 10. Aggregation For this research the unit of analysis is the organizational team, meaning the individual scores on the RFQ have to be aggregated to a team level. First the within group agreement should be measured in order to see if aggregation can be done. When testing all the scores on within group agreement it shows that all items are above .70 meaning suitable for aggregation (all scores are shown in appendix 11). The aggregation to team level is performed by taking the mean score of all team members per item (14 items total). The values are than aggregated to team level by taking the mean of all items per sub scale (seven items for the promotion scale and seven items for the prevention scale). These final scores represent the two final team scores on the prevention and promotion scale. Than the score on prevention focus was

Page | 13

subtracted from the score on promotion focus leaving one CRF measure. The outcomes of the normality test can be found in appendix 10

3.4.4 Team risk taking In order to analyze the risk taken by the teams, the allocation to the vignettes was used. This means there are four scores on risk taking per team. The allocation to the vignettes was divided by 10.000 in order to produce scores ranging from 0-100. The risk score is divided into two groups because of the manipulation of the experiment. There is one score consisting out of two allocations to the vignettes with the precise information and there is one score consisting out of two allocations of the vignettes with the imprecise information. For analyzing the risk score, a file was created with duplicated scores, meaning; one risk score for team one with precise information and on risk score for team one with imprecise information. The TE and CRF measures where duplicated for the team. More information on the normality analysis can be found in appendix 12

3.4.5 Quality indicators In order to warrant the construct validity of the research the measurements were based on existing scales. The TE measure was based on an existing scale developed by Bandura (2006). In order to make this measure as specific as possible the TE questions were aimed at the task the team had to perform in the experiment. This decision was made because otherwise it would be difficult finding a task that was recognizable to every team in the sample. On the level to measure TE the decision was made to use the ‘discussion approach’ as used by Guzzo, Yost, Campbell, & Shea (Guzzo, Yost, Campbell, & Shea, 1993). This way is questioned by some because when a group should come to a unanimous decision, social persuasion and power differences could interfere with the outcome. Though when taking into account that this, when occurring, would be likely to also occur when the team is coming to a unanimous decision about the allocation of the money to the vignettes, it makes sense to use this method. A limitation to this measure was the difficulty of answering it. Many of the teams at first did not understand how to start. Therefore the decision was made to include a small example. This could bias the results of the scale. The construct validity of the CRF measure was warranted by the use of an existing and tested scale by Stekelenburg (2006). To keep the time needed for the complete experiment as low as possible the decision was made to use the shortened scale (14 items). The big advantage of Page | 14

this scale was that it was developed in Dutch. Therefore the risk that the questions where translated in a wrong way was avoided. At first the measure of CRF had to be one at team level, though in practice this took too much time in the experiment. Therefore the scale was brought to individual level where the participants had to answer with regards to the team. TRT was measured with help of the experiment. The experiment is developed by the group of this master thesis circle with help of our tutor R. Leenders. The experiment exists out of two different sorts of cases (project and expansion) in order to make it suitable for all teams. After the development process the cases where tested on several individuals in order to check if they see the difference between the risky and the non-risky case. When these tests succeeded, the complete experiment was tested on teams of students. The order of the vignettes was constantly switched per team in order to control for a learning effect that would interfere with the manipulation. In order to warrant the internal validity as much as possible, several control variables where added to see if these would impact the risk taken by the team. Also the fact that the team was being observed and videotaped during the experiment might endanger the internal validity. Throughout the whole experiment a very detailed protocol was followed (appendix 4). This protocol was exactly the same for all experiments except for the order of the vignettes. Because this protocol was so strict, all of the teams where handled in the same manner making sure the experiment was not being influenced. Next to a protocol also a logistics file was used in order to make sure every team had everything they needed for every experiment (appendix 5). The fact that the sample is based on convenience sampling, could influence the external validity in such a way that the teams might have something in common driving them to participate in such an experiment.

3.4.6 Power analysis A power analysis was executed in order to test the statistical power of the conclusions. A power analysis is needed when sample sizes are low (Cohen, 1988). This analysis was done with help of an online test by D.S. Soper1. A general boundary of the power analysis is that the outcome should be above .8 saying that there is 80% or greater chance of finding a

1

* The test can be found on http://www.danielsoper.com/statcalc Page | 15

statistically significant difference. The analysis was run with two predictors (TE and CRF). The observed R2 (0.195) and the probability level (0.005) are retrieved from the second model in the moderated mediation analysis. The sample size for the analysis is the same as the sample size used for the moderated mediation analysis (98). The outcome of the analysis is that the observed statistical power is 0.950 meaning 95% chance. This shows that the power of the sample is big enough to test this model.

Page | 16

4. Results 4.1 Correlations First the relationship between the variables was tested using Pearson correlation. Table number two shows the correlations between the all the variables. With help of this table it is checked that the explanatory variables (TE and CRF) do not have a high correlation (r = .011). Also it shows the descriptives of the variables. Next also an analysis of possible multicollinearity is executed. This is done to see if there is variability between the two explanatory variables TE and CRF. The VIF value is 1 which is lower than the threshold of three meaning no multicollinearity is present. The output of the correlations can be found in appendix number 13. Variable

Descriptives Min

1. Average

Max

Correlation matrix Mean

Std. dv

1

2

3

4

5

6

18,00

52,5

37,598

10,421

-

Team

3,00

19

6,571

3,178

-.001

-

Team

0,25

144

21,148

36,463

.324*

-.271

-

4. TE

27,33

100

70,824

18,559

.342*

-.099

.275

-

5. CRF

-1,38

3,19

1,000

0,955

-.340*

.168

-.155

.011

-

6. TRT

0,00

100

38,849

19,062

.163

-.087

.084

-.037

.247*

age 2. size 3. tenure

*Correlation is significant at the 0,05 level (2-tailed) Table number 2

4.2 Moderated mediation analysis

Page | 17

-

In order to test all six hypotheses a moderated mediation analysis was executed with help of the SPSS Macro PROCESS by Hayes. The macro was downloaded2 and installed into SPSS. In order to make a dummy variable for the preciseness of information a new file called MMA.sav was created. The team variables for TE, CRF, TRT and the control variables where duplicated. Next the dummy variable (PI) was added. From the PDF file ‘Templates.pdf’ included with the macro it was found that the model to be used with the macro was model number 14. For the Y variable the risk score was used, for the X variable the TE score was used, for the M variable the CRF score was used and for the V variable the dummy score for the PI was used. Next the control variables average team member age, team size and team tenure where added. The complete SPSS output can be found in appendix 14. When analyzing the first model it shows that team age (β= -.035, p=.000) has a negative and significant effect on CRF, team tenure has a negative, but non-significant effect (β= -.000, p = .798) on CRF and team size has a positive, but non-significant effect (β=.053, p = .075) on CRF. The direct effect between TE and CRF is a slightly positive, but also non-significant one (β= .008, p = .101). Next, the second model includes all of the variables. Here the model summary shows that the F-Value is 3.12 meaning not all group averages are the same, the model is statistically significant (p = .005) and that the R2 is .195 meaning the model predicts 19,54% of the variance in TRT. When looking at the variables, it shows that CRF (β= 7.865, p = .006) and age (β = .677, p = .001) both have a positive significant effect on TRT. The direct effect of TE on TRT (β= -.142, p = .187) and the interaction effect (CRF x Information, β= -1.380, p = .715) both show to be non-significant. Finally, the analysis shows that both the mediation effect with precise and imprecise information are significant (Precise = BootLLCI = .007 and BootULCI = .182 and imprecise = BootLLCI = .000 and BootULCI = .197). Lastly the complete moderated mediation model shows to be non-significant (BootLLCI = -.100 and BootULCI = .049) 4.3 Evaluation of hypotheses

2

Found analysis.html

on:

http://www.afhayes.com/introduction-to-mediation-moderation-and-conditional-process-

Page | 18

Hypothesis 1 predicted a positive effect of TE on TRT, meaning the higher the TE is, the higher the risk as taken by the team will be. The moderated mediation analysis shows a slightly negative result of β = -.142 on TRT. The probability of this effect is .187. All effects with a significance level below .05 are proved to be significant. With these results hypothesis 1 is rejected. Hypothesis 2 predicted a positive effect of TE on CRF, meaning the higher the TE is, the higher the CRF will be. The moderated mediation analysis showed a slightly positive effect of β= .008 TE on CRF. When looking at the significance level of the effect it shows to be .101 meaning non-significant. Therefore hypothesis 2 is rejected. Hypothesis 3 predicted a positive effect of CRF on TRT, meaning the higher the CRF of the team is, the higher the TRT will be. The moderated mediation analysis shows a positive effect of β=7.86. This effect has a significance of .006, meaning it is statistically significant. Therefore hypothesis 3 was accepted. Hypothesis 4 is a partial mediation effect, predicting a positive effect of TE on TRT, through its positive effect on CRF. This mediation effect is measured twice; once with precise information and once with imprecise information. The effect size for precise information is β= 0.068 with a 95% confidence interval which does not include zero (BootLLCI = .007 and BootULCI = .182) meaning the effect is significant. The effect size for imprecise information is β 0.056 with a 95% confidence interval which does not include zero (BootLLCI = .000 and BootULCI = .197) meaning also this effect is significant. Because both paths show a positive and significant effect hypothesis 4 is accepted. Hypothesis 5 predicts the moderation effect of preciseness of information on the effect of CRF on TRT. The analysis shows an interaction effect (CRF x Information) of β= -1.380 with a p-value of .715. This means the effect is not significant and hypothesis 5 will be rejected. Hypothesis 6 predicts the moderated mediation effect of all the variables included in the model and as explained in the previous hypothesis. It predicted that as teams have higher TE, they would also have higher CRF which would lead to higher TRT. When this effect was exposed to imprecise information, it was expected to negatively influence the effect of CRF on TRT. The moderated mediation analysis shows an index of -0.012 with bootstrap at a confidence interval level of 95% including zero (BootLLCI = -.100 and BootULCI = .049). Therefore hypothesis 6 was rejected. Table number three shows the outcomes of all of the hypothesis for this research. Page | 19

Hypothesis

Coefficient

Significance

Rejected/

(β)

(p)

Accepted

.148

Rejected

.008

.126

Rejected

.493

.002

Accepted

.068

Significant

Accepted

.715

Rejected

6. The positive effect of team-efficacy on team risk -.012

Non-

Rejected

taking is partially mediated by collective regulatory

significant

1. Team-efficacy has a positive effect on team risk -.130 taking 2. Team-efficacy has a positive effect on collective regulatory focus 3. Collective regulatory Focus has a positive effect on team risk taking 4. Team-efficacy has a positive effect on team risk taking, through its positive effect on collective regulatory focus. 5: The positive effect of collective regulatory focus on -1.380 team risk taking will be weaker when imprecise information is applied to the decision-making process

focus and the effect of collective regulatory focus on team risk taking is moderated by preciseness of information. Table number 3

Page | 20

5. Conclusion and recommendations The aim of this research is to try and answer the research question ‘To what extent does teamefficacy influence collective regulatory focus and how do both affect team risk taking? And; how does the precision of information affect the relationship between collective regulatory focus and team risk taking? In order to answer this question a conceptual model, based on a theoretical framework was formed leading to six hypotheses. Below the outcomes of these hypotheses will be discussed. Based on theory the expectation was made that TE would lead to more risk taken by the team. The data shows a slightly negative result of β=-.142 on TRT, meaning that as TE increased with 1, TRT decreases with .142. In other words, the teams that where included into this experiment show less risk taking behavior as their TE goes up. This effect was however not significant so no conclusions can be drawn upon it. Next, TE was also expected to have an influence on the CRF of the team. The data showed a slightly positive effect of β=.008 TE on CRF, meaning as TE increases with 1, CRF increases with .008. For the CRF measure it means that the higher the team scores are, the more promotion focus (instead of a prevention focus) the team is. This effect however was not statistically significant withholding it from a conclusion to be drawn based on this data. The third hypothesis expects a positive effect of CRF on TRT of β=7.86. The data shows that as CRF increases with 1, TRT increases with 7.865. With a p-value of .006 this effect shows to be statistically significant. Therefore based on this specific data it can be said that CRF has a positive and significant effect on TRT. Next, a partial mediation effect of TE on TRT through CRF was hypothesized. This means that teams with high TE take more risky decisions when they also have high CRF. When looking at the output, an effect size of β=0.068 for precise information and β=0.056 for imprecise information with a 95% confidence interval is found. Both effects are found to be statistically significant meaning that there is a slightly positive mediation effect with both precise and imprecise information. The moderation or interaction effect describes the influence of PI on the effect between CRF and TRT. With an effect of β= -1.380 it shows that as information is imprecise there is a negative influence on the effect of CRF on TRT. This would mean that the positive relation between CRF and TRT would be lowered when information is precise. Though, because the effect has a p-value of .715, it is not proven statistically significant and no conclusion can be drawn based on this data. In the last section of the moderated mediation analysis a final result is given for the model as a whole. This Page | 21

shows that the moderated mediation effect as displayed in hypothesis 6 cannot be accepted because of non-significant results (BootLLCI = -.100 and BootULCI = .049). This means that from the results of this research it cannot be statistically proven that the positive effect of TE on TRT is partially mediated by CRF and the effect of CRF on TRT is moderated by preciseness of information. Next to the main variables of this research, three control variables where entered (mean age of team members, team size and team tenure). Age showed to have a significant result on both CRF and on TRT. The effect of age on CRF is β= -.035 meaning that as age increases with 1, CRF decreases with .035. This actually says that as the average age of a team is higher they will be less promotion focused and more prevention focused. When looking at the effect of age on TRT is shows a positive effect of β= .677. This means that as the average age of a team rises, it will become more risk taking. For the control variables team size and team tenure no significant effects where found. When concluding the findings from this research it shows that there is a proven effect of CRF on TRT in such a way that teams with higher CRF, take more risks. Also it shows that the full mediating effect of CRF between TE and TRT is present. This means that the effect of TE on TRT only hold when the CRF is high. In order to see if higher and more significant scores could be reached the research would be recommended to be executed with a larger data set. It is expected that the outcomes of this research can be interesting to organizations is such a way that there are certain personality traits that influence the behavior of teams when taking risks. For future research it might be interesting to further examine the relation between TE and TRT because the theoretical arguments are made for the effect in its individual form. If this effect would not be present on team level it might be good to see what the main influential factor is for this difference. A third recommendation for future research is to see if TE influences CRF or if this effect is the other way around (this is further explained in the chapter 6 of this research). A fourth recommendation would be to perform more research on the mediating effect of CRF between TE and TRT. This effect is not shown in theory before. When CRF is a big influence in the relation between TE and TRT it might be something that is usable in practical situations. It might help managers compose a team that fits into the expectations of the organization. On the non-significant effect of PI as a moderator also the recommendation stands to further research this. Theory is quite explicit that this should influence the risk taking. In chapter 6 a limitation is explained that might have caused this effect not to show. In possible future research this should be taken into account.

Page | 22

6. Discussion and limitations Out of six hypothesis stated in this research only two where accepted. For the expected effect of TE on TRT it is quite surprising that the effect holds for individuals but apparently not for teams. This could be because the team dynamics change the risk taking in such a way that is even becomes slightly lower when TE is high. A theoretical argument might be that the effect is actually an inverted u-shaped one. This was argued by Whyte (1998) saying that TE might lead to low decision making performance once it reaches a level that is too high because overconfidence might lead to unwanted high levels of TRT. Next to this theoretical reason it might also be that the way of measuring TE in this research was not as clear as strived for. One striking thing was that several teams that were able to quickly come to a consensus on a risky decision had quite some problems coming to a consensus on the TE questions. This might indicate that TE is something that is quite difficult to answer for a team as a whole. These same limitations could have influenced the effect of TE on CRF which also showed to be non-significant. When looking at the expected effect of TE on CRF it shows a slightly positive but non-significant effect. When looking at the theory section of this research (paragraph 2.3) it describes an article by Keller (2006) that works exactly the opposite way (RF affects efficacy). The reason this research uses the effect of TE on CRF is because it reasons as such that higher TE gives the team a certain task specific confidence making them have a higher CRF. The reason that this effect did not show could be because of previously named limitations with the measurement of TE or because of the theoretical construct of it. For CRF a possible limitation was the fact that it was, due to practical implications, measured on an individual level instead of on team level. The practical implication was that the proverbs where taking way to long for the team to reach consensus on. Theory on CRF says that CRF is not simply the sum of the RF score of all team members. Though the questionnaire introduced the questions in such a way that the team members had to answer the questions with regards to the team. This is an often used way of measuring a team concept and therefore no real implications are expected. When aggregating the individual scores to a team level the measure of agreement was high enough for all 14 items. Therefore this is not expected to be a true interference with the measurement. When looking at the mediation effect it shows that it was positive for both precise as imprecise information though the difference between these two was non-significant. On this mediating effect no theory was found before.

Page | 23

This is consistent with the outcomes of the moderating effect of information completeness on the effect of CRF on TRT. This showed to be non-significant but did have the direction that was expected (imprecise information leads to a weaker effect of CRF on TRT). A possible reason for this might be that the sample size was too small or that the teams did not see the imprecise information as was intended. During the experiments it was noted that quite some teams said that they thought the numbers in the imprecise cases where just the same as in the precise situation because we would pick the most centered number. This happened even though it was explained to them that this was not the case (the true numbers were defined based on randomness within the given range). Therefore some teams did not change their decision purely on the PI making it possible that the effect was not as hypothesized. Throughout the whole experiment, it was noticed that there was an unintended time pressure on the cases. In order to keep the experiment within a maximum of an hour time (to persuade teams to participate) there was quite a hurry in the answering of the questionnaires and the cases. Sometimes participants commented that if they would have had more time they might have made more extensive calculations for the risk. This could have influenced the risk taken by the team. A next striking thing was that hardly any team tried to, or was able to correctly calculate the risk. This is consistent with the theory in risk taking as described in section 2.1 of this research. A limitation of this research that was almost impossible to evade is that it is not being measured in a real situation. The cases used for the experiment where created in such a way that they would be applicable to teams from various sectors and organizational position within the organizations. As a consequence this could influence the normal team dynamics and roles present in real situations. Also the urge for the team want to perform could have been an influential factor. With the use of the good cause (KWF Kankerbestrijding) it was tried to give the teams a goal to take the cases seriously though this was something that was often quickly forgotten by the participants. From a statistical point of view the sample size might have been more interesting once it was bigger. Also, a limitation is that in order to dichotomize the PI, it was needed to duplicate the results of the CRF scale. Because of the size of the sample it was impossible to use the syntax with the clustering option (maximum of 20 clusters). A different option would have been to use multilevel models. This however is a technique that is not mastered sufficiently to perform.

Page | 24

References Bandura, A. (1997). Self-efficacy: The exercise of control. New York: Freeman. Bandura, A. (2000). Exercise of Human Agency Through Collective Efficacy. Current Directions in Psychological Science, 75-78 doi: 10.1111/1467-8721.00064. Bandura, A. (2006). Guide for constructing self-efficacy scales. In A. Bandura, Self-Efficacy Beliefs of Adolescents, (pp. 307–337). Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences. Hillsdale, NJ: Lawrence Earlbaum Associates. Dahlstrom, M. F., Dudo, A., & Brossard, D. (2012). Precision of Information, Sensational Information, and Self-Efficacy Information as Message-Level Variables Affecting Risk Perceptions. Risk Analysis, 32(1), 155-166. doi: 10.1111/j.15396924.2011.01641.x. Florack, A., & Hartmann, J. (2007). Regulatory focus and investment decisions in small groups. Journal of Experimental Social Psychology, 43(4), 626-632. doi:10.1016/j.jesp.2006.05.005. Guzzo, R. A., Yost, P. R., Campbell, R. J., & Shea, G. P. (1993). Potency in groups: Articulating a construct. British Journal of Social Psychology, 87–106 DOI: 10.1111/j.2044-8309.1993.tb00987.x. Higgins, E. T. (2002). How Self-Regulation Creates Distinct Values: The Case of Promotion and Prevention Decision Making. JOURNAL OF CONSUMER PSYCHOLOGY, 177191. doi:10.1207/S15327663JCP1203_01. Jones, J. E., & Bearley, W. L. (2001). Facilitating Team Development: A View From the Field. Group Facilitation, 3(2), 56-65. Keller, P. a. (2006). Regulatory Focus and Efficacy of Health Messages. JOURNAL OF CONSUMER RESEARCH, 109-114. Retrieved from: http://www.jstor.org/stable/10.1086/504141. Knight, D., Durham, C. C., & Locke, E. A. (2001). The relationship of team goals, incentives, and efficacy to strategic risk, tactical implementation, and performance. Academy of Management Journal, 44(2), 326-338. Retrieved from: http://www.jstor.org/stable/3069459. Krueger, N., & Dickson, P. R. (1994). How believing in ourselves increases risk taking: percieved self-efficacy and opportunity recognition. Decision Sciences;, 25(3), 385400. doi:10.1111/j.1540-5915.1994.tb00810.x.

Page | 25

March, J. G., & Shapira, Z. (1987). Managerial Perspectives on Risk and Risk Taking. Management Science, 33(11), 1404-1418. Retrieved from: http://www.jstor.org/stable/2631920. Pallant, P. (2005). SPSS SURVIVAL MANUAL, A Step by Step Guide to Data Analyzing using SPSS for Windows (version 12). Open University Press. Shupp, R. S., & Williams, A. W. (2008). Risk preference differentials of small groups and individuals. The Economic Journal,, 188(525), 258–283. doi:10.1111/j.14680297.2007.02112.x. Sitkin, S. B., & Pablo, A. L. (1992). Reconceptualizing the Determinants of Risk Behavior. The Academy of Management Review, 17(1), 9-38. Stable URL: http://www.jstor.org/stable/258646. Stekelenburg, J. (2006). PROMOTING OR PREVENTING SOCIAL CHANGE. Instrumentality, identity, ideology and group-based anger as motives of protest participation. PH.D. Dissertation, Faculty of Psychology and Pedagogic, VU University Amsterdam. Amsterdam: Ridderprint. Wallace, M. A., Kogan, N., & Bem, D. J. (1962). GROUP INFLUENCE ON INDIVIDUAL RISK TAKING1. Journal of Abnormal and Social Psychology, 65(2), 75-86. doi:10.1037/h0044376. Whyte, G. (1998). Recasting Janis’s Groupthink Model: The Key Role of Collective Efficacy in Decision Fiascoes. ORGANIZATIONAL BEHAVIOR AND HUMAN DECISION PROCESSES, 185–209. doi:10.1006/obhd.1998.2761.

Page | 26

Appendices Appendix 1: Questionnaire control variables 3. Wat is uw leeftijd:

____

4. Uit hoeveel mensen bestaat uw team origineel?

____

5. Hoe lang bestaat uw team in deze samenstelling?

____

Appendix 2: Questionnaire Team-efficacy De vragen uit deze vragenlijst worden anoniem verwerkt!

In de volgende ronde krijgen jullie een soortgelijke opdracht als jullie zojuist gehad hebben. Voor deze volgende ronde vragen wij jullie hoe jullie verwachten minimaal te scoren op de effectiviteit van jullie besluitvorming. Het betreft hier dus jullie effectiviteit voor deze specifieke opdracht. Vraag één persoon uit het team de tabel in te vullen en probeer zo veel mogelijk overeenstemming te bereiken over de antwoorden. De effectiviteit niveaus zijn als volgt te bepalen: 1=

Erg ineffectief; er is geen wederzijdse overeenstemming en geen toepasbare oplossing

5=

Gemiddeld effectief; er is wederzijdse overeenstemming op sommige punten en er zijn meningsverschillen over andere

10 =

Zeer effectief; er is wederzijdse overeenstemming en een toepasbare oplossing

De door jullie geschatte zekerheid in het behalen van een score moet worden uitgedrukt in een percentage tussen 0% en 100% (zie voorbeeld). Bijvoorbeeld:  0% betekent dat het team dit niveau niet kan behalen  50% betekent dat het team dit niveau net wel of net niet kan behalen  100% betekent dat het team zeker weet dat het minimaal deze score kan behalen.

Score op effectiviteit in besluitvorming

Zekerheid in minimaal te Voorbeeld

1 = erg ineffectief | 5 = gemiddeld effectief | 10 = zeer effectief

behalen score (0% tot 100% zekerheid)

1. Ons team zal in de volgende ronde minimaal een score van 9-10 behalen op effectiviteit

40% zeker

2. Ons team zal in de volgende ronde minimaal een score van 7-8 behalen op effectiviteit

50% zeker

3. Ons team zal in de volgende ronde minimaal een score van 5-6 behalen op effectiviteit

80% zeker

4. Ons team zal in de volgende ronde minimaal een score van 3-4 behalen op effectiviteit

100% zeker

5. Ons team zal in de volgende ronde minimaal een score van 1-2 behalen op effectiviteit

100% zeker

Appendix 3: Questionnaire Collective regulatory focus De vragen uit deze vragenlijst worden anoniem verwerkt!

Geef aan in hoeverre jij vindt dat de onderstaande spreekwoorden op jullie als team van toepassing zijn.

1. Ergens helemaal voor gaan

O

O

O

O

O

O

O

2. Niet geschoten is altijd mis

O

O

O

O

O

O

O

3. Voorkomen is beter dan genezen

O

O

O

O

O

O

O

4. Waar een wil is, is een weg

O

O

O

O

O

O

O

5. De kat uit de boom kijken

O

O

O

O

O

O

O

6. Schoenmaker, blijf bij je leest

O

O

O

O

O

O

O

7. Oost west, thuis best

O

O

O

O

O

O

O

Page | 28

8. Wie niet waagt, wie niet wint

O

O

O

O

O

O

O

9. Men moet niet over een nacht ijs gaan

O

O

O

O

O

O

O

10. Het leven naar je hand zetten

O

O

O

O

O

O

O

11. Verandering van spijs doet eten

O

O

O

O

O

O

O

12. Onder zorgen gebukt gaan

O

O

O

O

O

O

O

13. Je horizon verbreden

O

O

O

O

O

O

O

14. Doe maar gewoon, dan doe je gek genoeg

O

O

O

O

O

O

O

Appendix 4: Experiment protocol PREPERATION Time

Who

What

Check

30 minutes before Entire team (at least Welcome at the organization, wait for all start experiment. three of us) of us to arrive 20 minutes before Entire team (at least - Place all the chairs and tables in the right team shows up three of us) setting - Set up video camera as visualized in logistics - Set up laptop as visualized in logistics - Check if laptop and video camera work (Host) Observers

- Set up observation places as visualized in logistics (chair, observation sheet, pen) Divide who observes who: - Each observer observes a maximum of three participants (see logistics of experimental setup for a visual representation of where each observer will stand during the experimental rounds)

Host

- Decide where to stand during the introduction, and where to sit during the experimental phase (only if different than layout in logistics) Page | 29

- Have the materials for the experiment (script, questionnaire A, B, & C) ordered Host + Observer 1

- Seat the participants at their appointed spots - Ask if the formal leader of the team is present or not (fill in in ‘observation sheet host’)

WELCOME Time

Who

What

5 minutes

Entire team

- Take positions in front of the team

Host

- Welcome speech

Check

Goedendag, welkom iedereen. Allereerst wil ik u allen bedanken dat u vandaag de tijd heeft genomen om hier aanwezig te zijn. Mijn naam is (naam van de gastheer/dame), ik ben vandaag uw host. Aan uw linker- en rechterzijde vindt u mijn medestudenten (naam observer 1, naam observer 2, naam observer 3). Zij zullen dit experiment vandaag ondersteunen. Samen met nog (……) andere studenten studeren wij aan de Universiteit van Tilburg en hopen wij, door middel van dit experiment, dit jaar te slagen voor onze Master Organization Studies. Mocht u als team geïnteresseerd zijn in feedback op uw teamprocessen, dan kunnen wij u deze feedback verschaffen nadat alle experimenten zijn uitgevoerd. U kunt ons dit na afloop van het experiment laten weten. Het experiment waarin u vandaag gaat deelnemen bestaat uit vier korte rondes. In deze rondes wordt u gevraagd om als team een besluit te nemen. Verder willen wij u ook vragen om een vragenlijst in te vullen. Deze vragenlijst is verdeeld in drie delen. Het is de bedoeling dat deze vragenlijsten individueel ingevuld worden mits dit anders is aangegeven. Wij willen jullie dan ook vriendelijk verzoeken niet te overleggen met uw teamgenoten gedurende het invullen van de vragenlijst. Als dank voor jullie bijdrage aan dit experiment, zullen wij €500 doneren aan KWF Kankerbestrijding. Echter, afhankelijk van jullie prestatie wordt dit bedrag mogelijk nog groter, of kleiner. We gaan jullie straks vier keer vragen om een bedrag van €1.000.000 te investeren. Nadat alle vier beslissingen zijn genomen zal de computer bepalen hoeveel geld jullie hebben verdiend of hebben verloren. Je hebt winst en verlies dus in eigen hand. Jullie resultaat delen we door twintigduizend, en dat bepaalt wat er naar KWF gaat. Maken jullie bijvoorbeeld €100.000 winst, dan komt er €5 bij het bedrag dat we zullen doneren. Als jullie echter €100.000 verlies maken, dan doneren we €5 minder aan KWF. Kortom, jullie beslissingen bepalen uiteindelijk de omvang van de donatie voor KWF. Wij vragen u geen gebruik te maken van hulpmiddelen (zoals telefoons, woordenboeken en dergelijke). Wij verzoeken u om uw telefoon uit of op stil te zetten. Je naam is niet nodig, omdat de vragenlijsten anoniem worden verwerkt. Tijdens het experiment Page | 30

zullen opnames gemaakt worden met een video camera. Deze beelden worden niet gepubliceerd. We zullen ze alleen gebruiken als eventueel bewijsmateriaal dat dit experiment heeft plaatsgevonden. (SEINTJE GEVEN UITDELEN VRAGENLIJST 1)

QUESTIONNAIRE A Time

Who

1 minute

Host

What

Check

- Introduction questionnaire A - Tell the participants they can start

Host

- Hand out questionnaire A to the participants

Wanneer alles duidelijk is en er geen vragen meer zijn, mag u starten met het invullen van de eerste vragenlijst. Doe dit alstublieft individueel. Wanneer u vragen heeft mag u me deze altijd stellen. 10 minutes

Participants

- Fill in questionnaire A

Observers

- Write number and appearance of your participants on your observation sheets (e.g. Participant: 1.3, Appearance: bold and glasses)

Host

- When finished, collect questionnaire A

PART 1.1 EXPERIMENT (expansion – precise information) Time

Who

What

2 minutes

Host

- Start introducing the experiment and explain the tasks of the participants.

Check

- Tell participants how long they have for each round (7,5 minutes) Host

- Hand out the vignette expansion (X) to each team member

Observers

- Take position for the observation of team processes

Bedankt voor het invullen van de vragenlijst. We gaan nu beginnen aan het eerste deel van het experiment. Voor dit deel van het experiment vragen wij jullie om als team te bepalen hoeveel geld je in twee keuzes wilt investeren. Denk eraan dat jullie beslissingen bepalen of jullie geld Page | 31

verdienen of verspelen voor het KWF. Jullie hebben de opdracht voor de eerste ronde ontvangen. We zullen starten met een opdracht over een bedrijfsuitbreiding: Stelt u zich voor dat uw team is aangewezen door het bestuur van uw organisatie om advies uit te brengen met betrekking tot een mogelijke bedrijfsuitbreiding naar twee andere landen: Brutopia en Fantasia. Er is een bedrag van €1.000.000 beschikbaar, en de directie vraagt uw team om dit bedrag over twee aangewezen landen te verdelen. Die verdeling is helemaal aan jullie, maar al het beschikbare geld dient geïnvesteerd te worden. Jullie missie is uiteraard om een zo positief mogelijke uitkomst te verkrijgen. Jullie hebben 7,5 minuut de tijd om deze beslissing te maken en die tijd gaat nu in. We verzoeken jullie om de gemaakte keuze op het formulier in te vullen wat midden op de tafel ligt. Veel succes! 7.5 minutes

Participants

- Read the scenario - Discuss insights with team members - Make a decision on how to allocate the money - Write down decision on the answer sheet in middle of table.

Host

- After 5 minutes, tell the participants that they have 2.5 minutes left for this round. - Remind them again after 6.5 minutes

30 seconds

Observers

- Fill in the observation sheets

Host

- Collect material (answer sheet, observation sheet, vignette)

PART 1.2 EXPERIMENT (new project – precise information) 2 minutes

Host

- Introduction new (precise information)

project

- Tell participants how long they have for each round (7,5 minutes) Host

- Hand out vignette New project (X)

Observers

- Take position for the observation of team processes

Jullie ontvangen op dit moment de tweede opdracht van mijn collega’s. Ook in deze ronde zullen jullie geld gaan verdelen over twee keuzealternatieven. Page | 32

Deze ronde zal betrekking hebben op de start van twee nieuwe projecten, project Home en project Swan, die jullie als team moeten gaan uitvoeren. De directie is in gesprek met twee andere organisaties over samenwerking in een project. Deze projecten, project Home en project Swan, worden allebei aan jullie team uitbesteed en aangezien jullie hoog in het vaandel staan bij de directie hebben jullie als team ook inspraak met betrekking tot de investering in beide projecten. Jullie kunnen dan ook als team beslissen of het bedrijf het beschikbare geld in één of beide projecten investeert. Wederom kan er een bedrag van 1.000.000 euro worden verdeeld over beide keuzeopties. Hoe het geld wordt verdeeld over de twee projecten is geheel aan jullie, echter al het geld dient wederom geïnvesteerd te worden. Jullie hebben wederom 7,5 minuut de tijd om deze beslissing te maken en die tijd gaat nu in. Veel succes. 7.5 minutes

Participants

- Read the scenario - Discuss insights with team members - Make a decision on how to allocate the money - Write down decision on the answer sheet.

Host

- After 5 minutes, tell the participants that they have 2.5 minutes left for this round. - Remind them again after 6.5 minutes - Fill in the allocation of money from round 1 into excel.

Observers

- Fill in the observation sheets

Host

- When finished, collect material (answer sheet, observation sheet, vignette)

Time

Who

What

1 minute

Observers

- Hand out questionnaire B

Host

- Introduction questionnaire B

30 seconds

QUESTIONNAIRE B Check

Nu jullie klaar zijn met het eerste deel van het experiment, vragen we jullie weer om een korte vragenlijst in te vullen. Daar mag u nu mee beginnen. 10 minutes

Participants

Fill in questionnaire B

Host

- Fill in the allocation of money from round 2 into excel. Page | 33

Host

- When finished, questionnaire B

collect

PART 2.1 EXPERIMENT (expansion – imprecise information) Time

Who

What

Check

3 minutes

Host

- Start introducing round 3, and explain the tasks of the participants. - Tell participants how long they have for each round (7.5 minutes) - Hand out the vignette Expansion (Y)

Observers

- Take position for the observation of team processes

Zijn jullie benieuwd naar de antwoorden op de kennis vragen uit deze questionnaire? (DEZE ANTWOORDEN PAS GEVEN WANNEER ALLE QUESTIONNAIRES ZIJN OPGEHAALD) Japan heeft een grotere economie dan China

NEE, andersom

Google is de meest bezochte webpagina ter wereld

JA

BNG is de officiële afkorting van Bruto Nationaal JA Geluk De EU vlag heeft 24 sterren NEE, 12 Het rendement is de opbrengst van een investering

JA

Door inflatie ontstaat er koopkrachtdaling van het JA geld Groenland gezien vanaf een satelliet is voornamelijk JA wit Coca Cola was eind 2013 een meer waardevol merk NEE, andersom dan Apple 5+3x2 =16 en 3x(2+4)=18

NEE, eerste antwoord is 11

Economische stabiliteit verwijst naar een JA afwezigheid van buitensporige schommelingen in de macro-economie Page | 34

Een lamprei is een jong van een konijn

JA

Een muntje van 5 euro cent is in oppervlakte groter JA dan een muntje van 10 cent AEX bestaat uit een fondsenaantal van 20

NEE, 25

Een fusie is een samenwerkingsverband tussen NEE, Fusie is organisaties samenvoegen, dit is joint venture We gaan nu beginnen aan het tweede deel van het experiment. Ook voor dit deel van het experiment vragen wij jullie om als team te bepalen hoeveel geld je in de twee keuzes wilt investeren. We komen nog even terug op de beslissing om te internationaliseren naar twee nieuwe landen. De directie heeft gezien dat de concurrentie zich internationaal ook uitbreidt, en zich in het bijzonder richt op de landen Ardenia en Malaguay. De directie wil daarom ook in deze landen aanwezig zijn en heeft wederom €1.000.000 hiervoor uitgetrokken. Het is de bedoeling dat jullie als team weer gaan bepalen hoe veel geld er geïnvesteerd zal worden in elk van de twee landen. Echter de informatie omtrent deze landen is beperkt, vandaar dat alle cijfers zijn aangegeven in ranges. Houdt er rekening mee dat een range van 30-60% zowel erg laag (31%) als hoog (59%) uit kan vallen. Jullie 7,5 minuut om deze ronde te spelen gaat nu in. Succes. 7.5 minutes

Participants

- Read the scenario - Discuss insights with team members - Make a decision on how to allocate the money - Write down decision on the answer sheet

Host

- After 5 minutes, tell the participants that they have 2.5 minutes left for this round. - Remind them again after 6.5 minutes

30 seconds

Observers

- Fill in the observation sheets

Host

- Collect material (observation sheet, answer sheet, vignettes)

PART 2.2 EXPERIMENT (new project – imprecise information) 3 minutes

Host

- Start introducing round 4, and explain the tasks of the participants. Page | 35

- Tell participants how long they have for each round (7.5 minutes) - Hand out the vignette New project (Y) Observers

- Take position for the observation of team processes

Twee grote klanten hebben het bestuur benaderd voor de uitvoering van twee verschillende projecten, genaamd Project Hydra en Project Arrow. Het bestuur heeft jullie team aangewezen voor de uitvoering van deze projecten. Een bedrag van €1.000.000 mag weer verdeeld worden over de projecten. Echter de informatie omtrent deze projecten is beperkt, vandaar dat alle cijfers zijn aangegeven in ranges. Houdt er rekening mee dat een range van 50-80% zowel erg laag (51%) als hoog (79%) uit kan vallen. Jullie 7,5 minuut om deze ronde te spelen gaat nu in. Succes. 7.5 minutes

Participants

- Read the scenario - Discuss insights with team members - Make a decision on how to allocate the money - Write down decision on answer sheet.

Host

- After 5 minutes, tell the participants that they have 2.5 minutes left for this round. - Remind them again after 6.5 minutes - Fill in the allocation of money from round 3 into excel.

Observers

- Fill in the observation sheets

Host

- Collect material

QUESTIONNAIRE C Time

Who

What

1 minute

Host

- Introduction questionnaire C

Host

- Hand out questionnaire C + all the vignettes

Check

Nu dat jullie klaar zijn met het tweede deel van het experiment, vragen we jullie om de laatste, korte vragenlijst in te vullen. En daar mogen jullie nu mee beginnen. Succes. 8 minutes

Participants

- Fill in questionnaire within 8 minutes Page | 36

Host + observers

- Fill in the allocation of money from round 4 into excel. - Calculate final score for the experiment

Observer 3

- When finished, questionnaire C

collect

COMPLETION EXPERIMENT Time

Who

What

5 minutes

Host

- Announce final score (in total you won/lost €…… ; this means a donation of €…… to KWF)

Check

- Word of thanks Nu alle rondes geweest zijn, is de tijd om de score bekend te gaan maken. Gedurende de verschillende rondes hebben jullie het beschikbare budget verdeeld over verschillende opties. Of jullie beslissingen positief of negatief uitgevallen zijn, zal deels bepaald worden door de computer. Via een randomizer beslist de computer of jullie verlies, lage winst of hoge winst zullen hebben. Kortom, de uitkomst van de computer speelt een rol in jullie eindbedrag. Dit betekend dat wanneer jullie ‘verloren’ hebben, jullie het niet automatisch slecht hebben gedaan. Ik ga nu 1x op de knop drukken. De totale winst die jullie met de besluiten hebben gemaakt bedraagt €…… Dit zal resulteren in een donatie ter waarde van €……. aan KWF Kankerfonds. Het start bedrag wat wij doneren aan het KWF is 500 euro, dit beteken dat jullie bedrag hier vanaf/erbij op wordt geteld. Wij willen u hartelijk bedanken voor de interesse en deelname aan ons experiment. Hiermee hebt u ons enorm geholpen met onze dataverzameling, wat ons hopelijk weer een stap dichter bij het afstuderen brengt. [Verder willen we u nog vragen om dit experiment vertrouwd te houden indien er nog meer teams van uw organisatie deelnemen.] Indien gewenst sturen wij u feedback nadat alle experimenten zijn uitgevoerd. Mocht u nog vragen over het experiment of het gebruik van de data hebben, stel ze gerust. Wij wensen u verder nog een prettige dag. 5 minutes

Host + observers

- Answer questions - Note if the team wants to receive feedback after all the experiments are conducted - Give everyone a hand and thank them personally

BEFORE LEAVING THE ROOM Page | 37

Time

Who

What

10 minutes

Entire team

- Put the room back in its original state

Check

- Gather the materials in an envelope (Observation sheets, questionnaires, answer sheets). - Save everything on the laptop

Page | 38

Appendix 5: Experiment logistics Materials We brought along the following materials and printouts for every experiment. Materials

Number

Check

Video Camera and tripod

2

Pens

1 for each participant

Marker

1 for each participant

Laptop

1

Set of cards with participant numbers

1 (1-15)

Envelope to collect all the paperwork

1

Calculator

2

Blanco paper

4 for each participant

Printed materials

Number

Vignettes

1 of each round for each team member, 1 for the host

Forms to write down final decision of the team

1 for each round

Questionnare set A, B & C

1 for each participant

Protocol

1 for the host

Observation sheet

4x number observers (1 for each observer each round)

Time schedule form

1 for the host

Contract in case participants refuse to be filmed

1

Check

Sheet to write down general control variables and special 1 things that happened during the experiment (Like a phone went off or there was some disturbance) Confidentiality As researchers we will need to separate the different teams from each other, both in order to send them the results and for documentation regarding where the data is from originally. However, the data is only shared among the researchers and the research quality committee. Otherwise, the data will be completely anonymous, so in the final documents, the thesis, and the dataset, nothing private about the teams will be revealed. Thus, one document containing sensitive information and one copy of the videotaped experiment will be stored for the quality committee. It is important to emphasize that the research committee keeps all documents confidential for anyone outside of the committee. The video and data is stored with the team members Page | 39

Interruptions and extraordinary events We will ask all non-participating employees to exit the room to minimize distractions and influence. If there is a team leader or supervisor present, they will be asked to either join the experiment with his/her team or to leave the room, as they might disturb or influence the decisions of the team. We will also ask all participants to not use their cell phones or other apparatuses that may distract them. However, if someone is expecting an important call the employee can leave the room if the call is received, and we will then be careful to notice if their absence has any noticeable effect on the decision-making of the team. If someone arrives late, he/she will be able to join the experiment when the next round/vignette begins and will have to fill in the required questionnaires. Of course, a number of unforeseen distractions could occur, from teams not understanding the experiment to a fire alarm going off. However, we hope the 15-minute buffer we have to conduct the experiment will prepare us for all such unexpected events. Room set up

Up front, we will tell the organization we will need a room to fit all team members as well as our equipment, and they should preferably be able to sit around a table to discuss. The cameras will be set up so to capture the faces of all the employees, so that it is possible to analyse who said what at a later stage. The host will be seated so he/she is visible for all participants, while the observers will be seated out of the way of discussion, but so that they have a clear overview to do their observational tasks. Appendix 6: Vignettes Bedrijfsuitbreiding (X) Stelt u zich voor dat uw team is aangewezen door het bestuur van uw organisatie om advies uit te brengen met betrekking tot een mogelijke bedrijfsuitbreiding naar twee andere landen: Brutopia en Fantasia. Er is een bedrag van €1.000.000 beschikbaar voor deze uitbreiding. De directie vraagt jullie team om dit bedrag over de twee aangewezen landen te verdelen. Die verdeling is helemaal aan jullie, maar al het beschikbare geld dient geïnvesteerd te worden. De vraag van de directie is dus om hen te adviseren hoeveel geld er volgens jullie Page | 40

geïnvesteerd dient te worden in beide uitbreidingsinitiatieven, uiteraard met een missie om zoveel mogelijk positieve resultaten te boeken. Brutopia Brutopia is geografisch gezien dichtbij en heeft een vergelijkbare cultuur met het land waarin jullie bedrijf gevestigd is. Ook is de economie stabiel en tonen de economische vooruitzichten aan dat dit de komende jaren niet zal veranderen. Wanneer jullie besluiten om geld te investeren in een bedrijfsuitbreiding naar het land Brutopia is er een kans van 25% om voor iedere geïnvesteerde euro een winst van €0,60 te behalen. De winst kan ook relatief klein uitpakken, namelijk €0,10 bovenop iedere geïnvesteerde euro. De kans dat dit gebeurd is 50%. De totale kans is dus 75% dat deze uitbreiding leidt tot positieve resultaten. De financiële afdeling van jullie bedrijf heeft geschat dat er weinig redenen zijn waarom deze uitbreiding zal falen. Verder heeft het bestuur veel vertrouwen in jullie als team op basis van eerder behaalde resultaten. Op basis van bovengenoemde schattingen worden jullie dan ook geacht om goed te presteren en zullen er geen bonussen toegekend worden indien de uitbreiding positieve resultaten heeft voor de organisatie. Mocht de bedrijfsuitbreiding falen (kans is 25%) dan is dit financieel gezien ongunstig, aangezien de gehele investering verloren zal gaan. Echter, een gefaalde uitbreiding heeft geen verdere negatieve gevolgen die noemenswaardig zijn voor het bedrijf. Fantasia De markt waarin jullie opereren is in het land Fantasia groeiende en er zijn voldoende kansen om een goed aandeel te verkrijgen. Fantasia ligt wel op een ander continent. Door de relatief grote afstand moeten jullie beseffen dat er sprake is van een verschil in cultuur en werkwijze. De economie is echter groeiende en de vooruitzichten zijn gunstig, wat betekent dat jullie snel uit kunnen breiden als de juiste beslissingen gemaakt worden. Hierdoor kan de winst dus hoog zijn! Indien er namelijk succes bereikt wordt, maakt het bedrijf voor iedere geïnvesteerde euro een winst van €3,40. De verwachtingen op basis van de huidige situatie in het land, voorspellen dat de kans op een succesvolle bedrijfsuitbreiding 20% is. Als de business in Fantasia succesvol is, ontvangt elk teamlid een bonus. Ook is de kans zeer groot dat jullie team in de toekomst vergelijkbare prestigieuze taken opgedragen krijgt door het bestuur. De kans is echter relatief groot dat de bedrijfsuitbreiding negatieve gevolgen heeft voor het bedrijf. Zo is er een kans van 30% dat het bedrijf voor iedere geïnvesteerde euro slecht €0,90 terugkrijgt. De uitbreiding kan echter ook flink falen – wat een kans van 50% heeft – waarbij het volledige geïnvesteerde bedrag verloren gaat. Dit betekent ook dat de reputatie van de organisatie er flink onder te lijden zal hebben, waardoor er nog eens 40% bovenop het investeringsbedrag verloren gaat door reputatieschade. Bedrijfsuitbreiding (Y) Stelt u zich voor dat uw team is aangewezen door het bestuur van uw organisatie om advies uit te brengen met betrekking tot een mogelijke bedrijfsuitbreiding naar twee andere landen: Ardenia en Malaguay. Er is een bedrag van €1.000.000 beschikbaar voor deze uitbreiding. De directie vraagt jullie team om dit bedrag over de twee aangewezen landen te verdelen. Echter, de informatie omtrent deze landen is beperkt, vandaar dat alle cijfers zijn aangegeven in ranges. Houd er rekening mee dat een range van 30-60% zowel erg laag (bijv. 31%) als hoog (bijv. 59%) uit kan vallen. De verdeling van het geld is helemaal aan jullie, maar al het beschikbare geld dient geïnvesteerd te worden. De vraag van de directie is dus om hen te adviseren hoeveel geld er volgens jullie geïnvesteerd dient te worden in beide Page | 41

uitbreidingsinitiatieven, uiteraard met een missie om zoveel mogelijk positieve resultaten te boeken. Ardenia De economie in het land Ardenia is stabiel en als bedrijf zijn jullie bekend met haar markt. De competitie is laag, waardoor er een goede kans is op succesvol integreren in de markt van Ardenia. Wanneer het bestuur besluit om een bedrijfsuitbreiding te doen naar dit land en deze uitbreiding blijkt succesvol, dan zal het bedrijf voor iedere geïnvesteerde euro een winst van €0,60 verdienen. De kans dat dit gebeurt, is 10-40%. Er is echter ook een kans van 35-65% dat het succes beperkt blijft tot een winst van €0,10 bovenop iedere euro. Een kleine winst is dus zeker niet ondenkbaar in dit scenario. Het bestuur van jullie bedrijf heeft het vertrouwen dat jullie goed advies geven op basis van jullie kennis. Het land is relatief dichtbij en heeft een vergelijkbare cultuur. Verder is de markt bekend voor jullie bedrijf. De kansen zijn daarom ook relatief klein dat de volledige investering verloren zal gaan. Indien de volledige investering toch verloren gaat zal de organisatie een behoorlijke financiële tegenvaller moeten verwerken. De kans dat dit gebeurt, is 10-40%. Er zullen echter geen verdere negatieve consequenties zijn voor de organisatie die noemenswaardig zijn. Zoals vermeld heeft het bestuur vertrouwen in jullie team en een succesvolle uitbreiding van het bedrijf. Hiervoor zal jullie team geen extra beloningen ontvangen. Malaguay De economie in Malaguay maakt momenteel een groeispurt door. De competitie in de markt is matig, waardoor een uitbreiding naar dit land aantrekkelijk kan zijn voor een snelle groei van de business. Als jullie besluiten te investeren in een bedrijfsuitbreiding naar Malaguay en die uitbreiding een groot succes blijkt te zijn, dan zal jullie bedrijf voor iedere geïnvesteerde euro een winst behalen van €3,40. De schattingen van de financiële afdelingen tonen aan dat de kans op groot succes tussen de 5% en 35% ligt. In geval van succes zal elk teamlid een fraaie bonus ontvangen en bovendien kan het team rekenen op positieve feedback en uitdagende opdrachten in de toekomst. Malaguay ligt op een ander continent waardoor de cultuur behoorlijk anders is dan in het land waar jullie bedrijf momenteel opereert. Hierdoor moet het bedrijf rekening houden met mogelijke problemen tijdens het aanpassen aan deze nieuwe omgeving. De kans dat de uitbreiding negatieve consequenties heeft voor jullie bedrijf is dus relatief groot. Zo is er een kans van 15-45% dat het bedrijf voor iedere geïnvesteerde euro slechts €0,90 terugkrijgt. In dit geval is er dus sprake van een klein verlies voor jullie organisatie. Het is echter ook mogelijk dat de uitbreiding faalt. In dat geval gaat de gehele investering verloren plus nog eens een verlies van 40% van de investering door reputatieschade. De kans op een dergelijk verlies is 35-65%. Nieuw team project (X) Stelt u zich voor dat uw team is aangewezen door het bestuur van uw organisatie om advies uit te brengen met betrekking tot het starten van twee nieuwe projecten. De directie is in gesprek met twee andere organisaties over samenwerking in een project. Deze projecten, project Home en project Swan, worden allebei aan jullie team uitbesteed en aangezien jullie hoog in het vaandel staan bij de directie hebben jullie als team ook inspraak met betrekking tot de investering in beide projecten. Jullie kunnen dan ook als team beslissen of het bedrijf het beschikbare geld in één of beide projecten investeert. Er kan een bedrag van €1.000.000 Page | 42

worden verdeeld over beide keuzeopties. Hoe het geld wordt verdeeld over de twee projecten is geheel aan jullie, echter al het geld dient wederom geïnvesteerd te worden. Project Home Jullie organisatie heeft een langdurige en sterke relatie met de partner die dit project aandraagt. Het team is bekend met de verwachtingen van de partner, evenals de werkwijzen van de partner. Bovendien zijn jullie meer dan vertrouwd met de technologieën die worden gebruikt in dit project. Als dit project een succes blijkt te worden, dan verwacht het bedrijf dat iedere geïnvesteerde euro een winst van €0,70 zal opleveren. De kans op succes is 30%. Succes betekent een verhoogd vertrouwen van het management in jullie team en daarnaast een nog sterkere relatie met de partner. Er is echter ook een lagere winst mogelijk - met een kans van 50% - waarbij voor iedere geïnvesteerde euro een winst van €0,10 wordt geboekt. Er is dan ook een kans van 80% dat dit project met deze partner positieve resultaten oplevert. Omdat het management er min of meer vanuit gaat dat dit project positieve resultaten gaat opleveren, zijn er aan dit project geen bonussen of salarisverhogingen verbonden voor het team. Wanneer jullie team er niet in slaagt dit project tot een goed einde te brengen, dan zal de gehele investering verloren gaan. De kans hierop is 20%. Naast het feit dat dit een behoorlijke financiële tegenvaller is, zullen er echter geen verdere negatieve of noemenswaardige consequenties zijn voor jullie team en de organisatie. Project Swan Project Swan betreft de samenwerking met een bestaande partner, maar er dient wel gewerkt te worden met een technologie waar beide bedrijven onbekend mee zijn. De partner heeft de expertise om deze nieuwe technologie te implementeren in het project. Als team zullen jullie wel zo snel mogelijk bekend moeten worden met de technologie. De kans dat dit project tot een groot succes zal leiden is 25%. Succes betekent in dit geval wel dat het bedrijf voor iedere geïnvesteerde euro een behoorlijke winst behaald van €2,40. Bij succes kan ieder teamlid een bonus verwachten in de nabije toekomst. Eveneens zal, bij succes, jullie team in de toekomst vaker worden gevraagd om projecten uit te voeren voor het bedrijf. Het project kan echter ook negatieve gevolgen hebben voor de organisatie. Zo is er een kans van 45% dat er voor iedere geïnvesteerde euro slechts €0,80 zal worden terugverdiend. Indien dit project volledig faalt, zal dit resulteren in het verlies van de totale investering. Hier komt nog bij dat reputatieschade er voor zorgt dat er extra kosten zullen zijn. In dit geval is er een totaal verlies van de investering plus een extra €0,50 per geïnvesteerde euro. De kans op een dergelijk verlies is naar schatting 30%. Nieuw team project (Y) Stelt u zich voor dat uw team is aangewezen door het bestuur van uw organisatie om advies uit te brengen met betrekking tot het starten van twee nieuwe projecten. De directie is in gesprek met twee andere organisaties over samenwerking in een project. Deze projecten, project Hydra en project Arrow, worden allebei aan jullie team uitbesteed en aangezien jullie hoog in het vaandel staan bij de directie hebben jullie als team ook inspraak met betrekking tot de investering in beide projecten. Echter, de informatie omtrent deze projecten is beperkt, vandaar dat alle cijfers zijn aangegeven in ranges. Houd er rekening mee dat een range van 50-80% zowel erg laag (51%) als hoog (79%) uit kan vallen. Jullie kunnen dan ook als team beslissen of het bedrijf het beschikbare geld in één of beide projecten investeert. Er kan een bedrag van €1.000.000 worden verdeeld over beide keuzeopties. Hoe het geld wordt verdeeld Page | 43

over de twee projecten is geheel aan jullie, echter al het geld dient wederom geïnvesteerd te worden. Project Hydra Project Hydra is vergelijkbaar met projecten die jullie team eerder heeft uitgevoerd. Jullie zijn dan ook bekend met de werkwijzen van dit nieuwe project. Het bedrijf is ook al in bezit van de software die nodig is om dit project uit te voeren. Verder is jullie team ook enigszins bekend met de verwachtingen en werkwijzen van jullie partner. Door deze factoren is de kans dat dit project succesvol wordt dan ook relatief hoog, namelijk 15-45%. Indien er sprake is van succes dan wordt er voor iedere geïnvesteerde euro een winst van €0,70 behaald. Het is echter ook mogelijk dat het project een gematigde winst oplevert. Hiervoor is er een kans van 35-65% en wordt er voor iedere geïnvesteerde euro €0,10 extra verdiend. Het management verwacht dat jullie team dit project tot een succes zal maken en daarom dient uw team niet te rekenen op al te veel erkenning wanneer het project daadwerkelijk positieve resultaten oplevert. Positieve resultaten zullen er wel voor zorgen dat het vertrouwen van het management in jullie team stijgt, evenals de relatie met de partner. Mocht het project falen dan zijn de vooruitzichten financieel gezien een stuk minder aantrekkelijk. In dat geval zal namelijk de gehele investering verloren gaan. De kans dat het project faalt, is geschat op 535% Project Arrow Project Arrow heeft betrekking op een contract met een nieuwe partner. Voor het project is het gebruik van nieuwe software nodig. De organisatie waar jullie mee samenwerken heeft het systeem recentelijk geïmplementeerd om de efficiëntie te verbeteren. Jullie team zal dan ook zo snel mogelijk vertrouwd moeten raken met deze software om het project tot een succes te maken. Als jullie team erin slaagt om dit project tot een succes te maken, wordt er verwacht dat iedere geïnvesteerde euro een winst van €2,40 gaat opleveren. De kans op succes is 1040%. Het management is bereid een financiële beloning uit te betalen indien het project succesvol afgerond wordt. Omdat jullie team vertrouwd is geraakt met de software van het project zal het team in de toekomst meerdere interessante projecten kunnen verwachten. Het is echter ook goed mogelijk dat de uitkomsten financieel gezien negatief zullen zijn. Er is bijvoorbeeld een kans van 30-60% dat er voor iedere geïnvesteerde euro slechts €0,80 wordt terugverdiend. Verder vertellen de schattingen van het financiële departement dat als het project faalt - de kans hierop is 15-45% - de gehele investering verloren gaat. Bovendien zal er dan nog eens een extra €0,50 bovenop iedere geïnvesteerde euro verloren gaan als gevolg van reputatieschade voor zowel jullie bedrijf als het team zelf. Appendix 7: Calculations of experiment Calculations Expected values (EV) of the scenarios are calculated to be exactly the same within one scenario. In this way there is no rational argument for choosing the risky option in favor of the safe option (or vice versa). Since the vignettes were developed to be similar except for their riskiness, choices will be made purely on the differences in risk between the choice alternatives within a scenario. Between the two scenarios (expansion and new project) the expected values were varied, because otherwise the numbers in the vignettes would be too similar to the other scenario. In the excel sheet the following calculation was used for the expected values: Page | 44

EV(X) = (p1*e1*X) + (p2*e2*X) + (p3*e3*X) – X EV(X) : Expected value for investment X

: Investment in €’s

p

: Probability of outcome, p1=success, p2=moderate success/failure, p3=failure

e

: Amount of win/loss, e1=success, e2=moderate success/failure, e3=failure

Using this calculation in the excel sheet allowed us to adjust the probability of win/loss and the associated payouts in order to reach an optimal distribution that would represent reality, and also clearly differentiate between safe and risky. In the tables below an overview of the definitive probabilities of win/loss and the associated payouts is presented. In the table below we can see the probability of winning, the associated payouts, and the expected value for each scenario. The probability of win/loss and associated payouts are incorporated in the vignettes. A screenshot of this excel tab is shown below.

Expansion Scenario Negative Expected Value (EV) Expansion safe alternative

Expansion risky alternative

75% chance of positive return

20% chance of positive return

Success: investment * 1.6

Success: investment * 4.4

Moderate success: investment * 1.1

Moderate failure: investment * 0.9

Failure: investment *0

Failure: investment * -0.4 (additional loss of 40%)

*In the imprecise information condition a range of 30% is applied to the probability of the outcome; the probability of the precise information condition being the mean of this range Possible allocations Safe option

Risky option

Risk score

Expected value

1.000.000

0

0

- € 50.000

900.000

100.000

10

- € 50.000

800.000

200.000

20

- € 50.000

700.000

300.000

30

- € 50.000

600.000

400.000

40

- € 50.000

500.000

500.000

50

- € 50.000 Page | 45

400.000

600.000

60

- € 50.000

300.000

700.000

70

- € 50.000

200.000

800.000

80

- € 50.000

100.000

900.000

90

- € 50.000

0

1.000.000

100

- € 50.000

Project Scenario Positive Expected Value (EV) Project safe alternative

Project risky alternative

80% chance of positive return

25% chance of positive return

Success: investment * 1.7

Success: investment * 3.4

Moderate success: investment * 1.1

Moderate failure: investment * 0.8

Failure: investment *0

Failure: investment * -0.5 (additional loss of 50%)

*In the imprecise information condition a range of 30% is applied to the probability of the outcome; the probability of the precise information condition being the mean of this range

Possible allocations Safe option

Risky option

Risk score

Expected value

1.000.000

0

0

+ 60.000

€

900.000

100.000

10

+ 60.000

€

800.000

200.000

20

+ 60.000

€

700.000

300.000

30

+ 60.000

€

600.000

400.000

40

+ 60.000

€

500.000

500.000

50

+

€ Page | 46

60.000 400.000

600.000

60

+ 60.000

€

300.000

700.000

70

+ 60.000

€

200.000

800.000

80

+ 60.000

€

100.000

900.000

90

+ 60.000

€

0

1.000.000

100

+ 60.000

€

Excel sheet tab; expected value Play the game A second tab in the excel sheet is used as a game tool during the experiments. At the end of an experiment the chosen allocations by the team are inserted in this excel tab. This excel tab calculates, using randomizers, the actual winnings/losses of the team. Most importantly, this excel tab calculates the charity donation in either green (win) or red (loss). A screenshot of this excel tab is shown below. If interested, the excel sheet can be provided on request.

Page | 47

Excel sheet tab; play the game

Appendix 8: Reliability analysis TE

Page | 48

Appendix 9: Principal component analysis CRF

Page | 49

Page | 50

Appendix 10: Reliability analysis CRF When testing the aggregated scales for normality the results show that the KolmogorovSmirnov it non-significant meaning the distribution of the scores is normal. Also the Trimmed-Means for both scales hardly deviate from the mean score, indicating there are no outliers influencing the data.

Appendix 11: Within group agreement CRF Group number 3 4 5 6 7 8 9 10

rwg 0,925 0,956 0,984 0,949 0,975 0,966 0,958 0,969 Page | 51

11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51

0,913 0,935 0,975 0,946 0,965 0,953 0,963 0,941 0,938 0,983 0,975 0,950 0,948 0,975 0,965 0,975 0,957 0,966 0,968 0,979 0,933 0,963 0,979 0,881 0,971 0,976 0,974 0,953 0,969 0,984 0,881 0,970 0,966 0,863 0,969 0,981 0,976 0,978 0,983 0,982 0,956

Appendix 12: Reliability analysis TRT Page | 52

Page | 53

When assessing the normality of the risk score with precise information it is found that the Mean (38.849) and the Trimmed Mean (38.155) hardly differ meaning no extreme scores are strongly influencing the mean. When checking the boxplot it shows that there are some outliers that might cause this difference. These cases where checked on possible errors when Page | 54

filling in the data. No errors where found, therefore it was decided to leave the variables in the data set. The score on Skewness (.611) is positive, showing the peak is more left than within the normal distribution. The Kurtosis score is .772. The Kolmogorov-Smirnov score (.018) shows that it is significant suggesting a violation of the assumption of normality. This violation is often seen with small samples sizes which in this case could also be the reason. In order to further examine the normality of the data the normal Q-Q plot is checked for a straight line which is visible. The Detrended Normal Q-Q Plot should show hardly no clustering and most points around the line of 0,0 which is the case. Because of the sample size and all the plots pointing to normality of the data it is assumed that the assumption of normality is not violated.

Appendix 13: Output correlations

Page | 55

Appendix 14: Moderated mediation analysis

Page | 56

Page | 57

Appendix 15: Syntax preliminary analysis TE DATASET ACTIVATE DataSet1. RELIABILITY /VARIABLES=TE_1 TE_2 TE_3 TE_4 TE_5 /SCALE('TE-per-team') ALL /MODEL=ALPHA /STATISTICS=DESCRIPTIVE SCALE CORR /SUMMARY=TOTAL CORR. Appendix 16: Syntax preliminary analysis CRF DATASET ACTIVATE DataSet2. FACTOR /VARIABLES C1_RF_Promo C5_RF_Prevent C6_RF_Prevent

C2_RF_Promo

C3_RF_Prevent

C4_RF_Promo

C7_RF_Prevent C8_RF_Promo C12_RF_Prevent C13_RF_Promo

C9_RF_Prevent

C10_RF_Promo

C11_RF_Promo

C3_RF_Prevent

C4_RF_Promo

C14_RF_Prevent /MISSING PAIRWISE /ANALYSIS C1_RF_Promo C5_RF_Prevent C6_RF_Prevent

C2_RF_Promo

Page | 58

C7_RF_Prevent C8_RF_Promo C12_RF_Prevent C13_RF_Promo

C9_RF_Prevent

C10_RF_Promo

C11_RF_Promo

C14_RF_Prevent /PRINT INITIAL CORRELATION KMO EXTRACTION ROTATION /FORMAT SORT BLANK(.10) /PLOT EIGEN /CRITERIA MINEIGEN(1) ITERATE(25) /EXTRACTION PC /CRITERIA ITERATE(25) DELTA(0) /ROTATION OBLIMIN /METHOD=CORRELATION.

FACTOR /VARIABLES C1_RF_Promo C5_RF_Prevent C6_RF_Prevent

C2_RF_Promo

C3_RF_Prevent

C4_RF_Promo

C7_RF_Prevent C8_RF_Promo C12_RF_Prevent C13_RF_Promo

C9_RF_Prevent

C10_RF_Promo

C11_RF_Promo

C3_RF_Prevent

C4_RF_Promo

C10_RF_Promo

C11_RF_Promo

C14_RF_Prevent /MISSING PAIRWISE /ANALYSIS C1_RF_Promo C5_RF_Prevent C6_RF_Prevent C7_RF_Prevent C8_RF_Promo C12_RF_Prevent C13_RF_Promo

C2_RF_Promo C9_RF_Prevent

C14_RF_Prevent /PRINT INITIAL CORRELATION KMO EXTRACTION ROTATION /FORMAT SORT BLANK(.10) /PLOT EIGEN /CRITERIA FACTORS(2) ITERATE(25) /EXTRACTION PC /CRITERIA ITERATE(25) DELTA(0) /ROTATION OBLIMIN Page | 59

/METHOD=CORRELATION.

RELIABILITY /VARIABLES=C1_RF_Promo C2_RF_Promo C10_RF_Promo C11_RF_Promo C13_RF_Promo

C4_RF_Promo

C8_RF_Promo

C6_RF_Prevent

C7_RF_Prevent

/SCALE('Promotion') ALL /MODEL=ALPHA /STATISTICS=DESCRIPTIVE SCALE CORR /SUMMARY=TOTAL CORR.

RELIABILITY /VARIABLES=C3_RF_Prevent C9_RF_Prevent C12_RF_Prevent

C5_RF_Prevent

C14_RF_Prevent /SCALE('Prevention') ALL /MODEL=ALPHA /STATISTICS=DESCRIPTIVE SCALE CORR /SUMMARY=TOTAL CORR.

EXAMINE VARIABLES=CRF_Promotion_Agg CRF_Prevention_Agg /ID=TeamNr /PLOT BOXPLOT STEMLEAF HISTOGRAM NPPLOT /COMPARE GROUPS /STATISTICS DESCRIPTIVES EXTREME /CINTERVAL 95 /MISSING PAIRWISE /NOTOTAL. Appendix 17: Syntax preliminary analysis TRT DATASET ACTIVATE DataSet1. EXAMINE VARIABLES=Risk Page | 60

/PLOT BOXPLOT STEMLEAF HISTOGRAM NPPLOT /COMPARE GROUPS /STATISTICS DESCRIPTIVES /CINTERVAL 95 /MISSING PAIRWISE /NOTOTAL. Appendix 18: Syntax correlations CORRELATIONS /VARIABLES=Age team_size Tenure TE RF Risk /PRINT=TWOTAIL NOSIG /STATISTICS DESCRIPTIVES /MISSING=PAIRWISE.

Appendix 19: Syntax multicollinearity analysis DATASET ACTIVATE DataSet1. REGRESSION /MISSING LISTWISE /STATISTICS COLLIN TOL /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT Regulatory_Focus /METHOD=ENTER Team_Efficacy.

Page | 61