Competitive Versus Non-Competitive Goals in Group Decision-Making

Method

Seventy-nine university students (50 female and 29 male) from different study programs took part in the study. Their age varied between 18 and 45 years with a mean age of 22.4 years among men and 26.3 years among women.

Procedure

One week before the experiment, the participants were (a) informed of the main aim of the study and asked to participate, and (b) asked to complete a questionnaire with background data such as age and sex, and other variables not used in this study. One week later when they arrived at the experiment they were randomly divided into groups of three to five persons, most often four. The total number of groups was 20; 5 with only men, 10 with only women, and 5 groups with both men and women. These groups were randomly assigned to one of two conditions: (a) to assemble a team for a relay running competition with the goal to maximize the chance to win the competition, or (b) to assemble a team for a relay running competition with the goal to have fun and to maximize the sense of community within the team.

Participants in each group were placed around a table on which two sheets of paper, lying upside down, and a pencil were placed. The session started with the experimenter reading the instructions, and then the participants were asked to pick up the sheets. The first contained the instruction about the task and how to proceed with it, the time limit of 15 min, and the goal specification to which the group was assigned (see Table 1). It also contained pictures of 20 target persons (10 men and 10 women) with information about their first name, age, occupation, and favorite hobby. Occupations and hobbies assigned to the targets were categorized as traditionally feminine, masculine, or gender neutral based on ratings made by another group of students (N = 11), who rated a larger set of occupations and hobbies on five-step scales ranging from typically feminine to typically masculine. These three sets of occupations and hobbies were as evenly distributed as possible across the 10 female and 10 male target persons in the stimulus material.

OPEN IN VIEWER

Table 1.

Written InstructionsGiven to Participants.

General information to all groups
You are members of the IMO association. You are assigned the task to put together a relay team for a 2-day long running competition. Unfortunately, you are injured and cannot run yourselves.
Information to groups with the competition task:	Information to groups with non-competition task:
IMO has achieved well in earlier years and your task is to put together a team to maximize the chance of winning.	IMO has had many laughs and pleasant memories in earlier years and your task is to put together a team to maximize a sense of community.

The other sheet of paper was a protocol, containing the different relay sections, and on which the group should write their final decision. There were 10 different relay sections of various lengths, and the groups were told that they could use as few or as many runners as they wished and that they were also allowed to split sections. The session was filmed by the use of a surveillance system with three cameras set at different angles.

After 15 min, the experiment was stopped and the experimenter collected the group’s protocol. Immediately afterward, they completed a questionnaire again, among other things asking about their reflections of their experiences of the decision task. In the present study, only the data from the group discussion were used and not the questionnaire data. Finally, they were debriefed about the specific aim of the study and about how the data were to be handled. Participation in the experiment was a course requirement (i.e., demonstration of experimental research methods) but being part of research was voluntary. After debriefing, the participants signed an agreement consenting to the use of the data they provided for research purposes.

Data Analysis

The filmed group sessions were first transcribed and then transferred to NVivo, a qualitative computer-based analysis program. The transcriptions were coded blindly into the IPA (Bales, 1999) categories with no a priori knowledge about the experimental conditions. IPA is comprised of four major categories, each with three sub-categories (see Table 2). Coding was carried out by one of the authors and a research assistant independently of each other for four groups. The inter-rater reliability was 86% across the four categories. The remaining 16 groups were coded by one judge.

OPEN IN VIEWER

Table 2.

Bales’ Interaction Process Analysis (IPA) Categories and Examples.

Functional area	Category	Quotes
Socio-emotional area: Positive reactions	Shows solidarity, seems friendly Shows tension release Shows agreement	“Thanks for the help” “Laughs, that’s funny” “Yea, I agree”
Task area: Information sharing	Gives suggestions Gives opinion Gives orientation	“We should do like that” “I think we should pick him” “We should have both men and women”
Task area: Questions	Asks for orientation Asks for opinions Asks for suggestions	“How shall we start?” “Which one do you prefer?” “Any ideas?”
Socio-emotional area: Negative reactions	Shows disagreement Shows tension Shows antagonism	“I think you’re wrong” “We should follow the rules I told you that” “You’re talking too much!”

The next step was to analyze the content of the interactions. One theme emerging from the discussions centered on whom to pick for the team and who should be included or excluded. Thus, one category labeled inclusion and another category labeled exclusion were established. Another theme emerged regarding the argument behind possible choices. There was no explicit information given about the qualifications or personality of the available persons. However, information about these qualifications triggered arguments referring to implicitly associated beliefs (i.e., stereotypes). One such theme was related to their qualifications, or lack of qualifications, to run. Another was related to attributes relevant to form a good group climate. As such, they could be said to correspond to the stereotype typology of competence and warmth (Fiske et al., 2002). Coding of statements was carried out by the authors independently of each other for a sub-set of four groups, while the rest was categorized by one person. The inter-rater reliability for the category inclusion was 78%, for exclusion 91%, for competence 88%, and for warmth 64%. These four categories, together with examples, are presented in Table 3.

OPEN IN VIEWER

Table 3.

Examples of Statements for the Categories of Inclusion, Exclusion, Competence, and Warmth for the Competition and Non-Competition Conditions, Respectively.

Direction of information	Competition	Non-competition
Inclusion	“Pilots are great athletes, he should be included”	“Perhaps he isn’t a good runner, but he is a psychologist and may help the team to come together”
Exclusion	“He likes wine tasting, probably an alcoholic, he can’t run and shouldn’t take part in the team”	“It will be boring for him, who’s interested in sports, to be in a group with people only interested in food”
Competence	“He’s a fireman and plays ice hockey; firemen usually have a good physique so he should be a good runner”	“The pilot and the ice hockey player should be together, they’re both masculine”
Warmth	Not mentioned	“We should pick Juha since he’s a priest and can keep the group together”

Results and Discussion

The IPA analysis revealed a total of 4,433 statements possible to categorize into the four main categories (positive reactions, information sharing, questions, and negative reactions). The most frequent type of statements referred to information sharing, constituting 66% of the total number of statements. The second most frequent category was questions, with 17% of the total number of statements. This means that there was a lot of information sharing which was not initiated by asking for information, but was given freely. The category of positive reactions was of about the same magnitude as questions (14%; 1). Finally, there were very few negative reactions expressed in the group sessions (3%).

Most of these categories were equally common in the groups with different conditions (competitive vs. non-competitive goal). The only exception was the category of positive reactions. There were more positive reactions in the groups with a non-competitive goal (M = 41.00, SD = 18.91) than in the groups with a competitive goal (M = 25.66, SD = 13.74), F(1, 18) = 4.44, p < .05, η² = .20. However, this difference was generated by the sub-category labeled shows tension release, which was the only sub-category with a significant difference between the groups, F(1, 18) = 8.67; p < .01, η² = .33. Thus, the main finding was that the interaction process did not differ particularly between the groups with different goals set for the task.

The remaining analysis focused on the content of the discussions within the groups (see Table 3). Discussions about how the target persons could perform were prominent in every group discussion, regardless of the goal set for the task. Almost immediately after conversations began, statements relating to competitiveness (e.g., “How should we pick the best team?”; “I think this one is very good as he likes to play soccer”) were made. What is important here is that arguments to find the most competent runners were related to the stereotypic attributes associated with hobbies, age, sex, occupation, or a combination of these. For example, the male pilot and fireman were often assumed to be good performers: “He is a fireman and plays ice hockey, firemen usually have a good physique so he should be a good runner.” Others were considered to be poor performers: “He works at kindergarten and likes knitting, he is probably not the man for our team.”

For groups with a competitive goal, statements about the target persons’ competence were positioned in the beginning of the discussion. Persons presumed to have a high competence (i.e., strong, athletic, fit) were identified to be selected before those who were presumed to have a low competence (i.e., weak, unathletic, and unfit). Groups with a non-competitive goal sometimes stressed the need to have competent runners in the team, but the reason was that they assumed that those people liked running and would probably be happier if they were allowed to run, which should lead to a better group climate. The quantitative data confirm the differences between groups with a competitive and a non-competitive goal, F(1, 18) = 13.67; p < .01, η² = .42. The target persons’ presumed competence was elaborated more in the discussions in groups with the competitive goal (M = 13.67, SD = 4.25) than in groups with a non-competitive goal (M = 7.63, SD = 2.62).

The first few minutes of the discussions in the groups with the non-competitive goal were devoted to explaining what was meant by a good group climate. Statements such as “We are supposed to select a team to reach a good group climate, not a winning team,” or “It is most important that they have fun together” permeated the discussion. Groups with a non-competitive goal also discussed the need for warmth among the team; the central catchphrase was, “To get a good group climate.” They argued that “Everyone should be involved,” implying that this would lead to a better group climate; “We should pick Juha since he is a priest and can keep the group together,” “Everyone should take part in the team,” and “Who can contribute with what to get a great group climate?” and so on. The quantitative data confirm that groups with a non-competitive goal discussed the need for a good group climate more (M = 7.63, SD = 5.4) than groups with a competitive goal did, where this need was not discussed at all (M = 0), F(1, 18) = 28.24; p < .001, η² = .61.

As the task was to pick a team for a relay running competition, arguments about inclusion and exclusion would be central to the groups’ discussions. This was immediately evident in the groups with a competitive goal. A strategy discussed in these groups was to include many athletes to yield a better performance: “If we use more athletes, everyone may run a shorter distance and this will make a better result.” However, another strategy was to use fewer runners for the same purpose: “Juha is good, he can run two or more sections.” Inclusion strategies were often motivated by attributes assigned to the target persons based on preconceptions associated with the information given: “Pilots are great athletes, he should be included”; “Men are better runners than women.”

Groups with a non-competitive goal also discussed inclusion issues, but from another starting point, as it was more or less implicitly understood that all members should be included in the team. Inclusion statements thus focused on who should be placed in which section. The dialogue would contain arguments like, for example, that those who presumably liked running should be included as this would probably lead to a better group climate. Otherwise they talked of a different set of competences they assumed would contribute to the atmosphere in the team: “Perhaps he is not a good runner, but he is a psychologist and may help the team come together.” Their discussion also focused on matching people with the same interests, arguing that this would lead to a good interaction and a better group climate: “Here we have two teachers, they’ll match each other.” Some of the groups with a non-competitive goal argued in a different direction that those with contrasting interests could run together as opposites may attract. All in all, the attributes they used to argue for the inclusion of people in the relay team were of the kind that would optimize the sense of community within the teams.

Exclusion of members from the team was a central issue for groups with a competitive goal. It was important to exclude target persons they assumed to perform poorly based on preconception of categories such as age, sex, hobby, occupation, or a combination of these: “Old people run slower than young people,” “She likes downhill skiing, I don’t think she will be a good runner,” and “He likes wine tasting, probably an alcoholic, he can’t run and shouldn’t take part in the team.” The habit of using preconceived stereotypes was taken a step further in the discussions with even more generalizing statements based on vague group characteristics: “He doesn’t look like a good runner and he’s a teacher, probably in computing, and they can’t run” and “I have a friend and she plays soccer and can only run short distances, I don’t think that female soccer players are good long distance runners.” In the groups with a non-competitive goal, however, discussions about exclusion of people from the relay team were rare.

Thus, it seems that arguments regarding both inclusion and exclusion strategies were more important for groups with a competitive goal set for the task. This is confirmed by the quantitative results. The groups with the competitive goal had more inclusion statements (M = 18.75, SD = 10.49) than the groups with the non-competitive goal (M = 9.00, SD = 5.04), F(1, 18) = 5.92, p < .05, η² = .25. The result was similar regarding exclusion statements: competitive goal groups (M = 10.33, SD = 7.87), non-competitive goal (M = 2.13, SD = 1.55), F(1, 18) = 8.35, p < .01, η² = .32. Groups with a competitive goal used both statements of inclusion and exclusion more often than groups with a non-competitive goal. It seems that inclusion and exclusion mechanisms are both vivid concepts in the competitive situation to a greater extent than in the non-competitive setting where inclusion of all seems to be an implicit precondition.

Conclusion

The aim of this study was to study interaction processes and interaction content in groups that were given different goals for a decision task. The IPAs revealed that the most frequent form of interaction was that of giving information to each other. Given that information sharing was four times more frequent than asking for information suggests that information was freely and spontaneously given in all groups. The IPA revealed only minor differences between the groups, indicating that the interaction process was not dependent on the goal set for the interaction.

Also clear from the results is that the arguments in all groups went well beyond the information given about the target persons. Often, attributes associated with group stereotypes were used to characterize the persons who were to be included or excluded in the relay teams. Many of those attributes were related to the basic stereotype dimensions of competence and warmth (Fiske et al., 2002). Attributes related to competence were used in all groups regardless of the goal for the task, but in different ways. In groups with a competitive goal, positive competence attributes were used to include people on the team, whereas negative competence attributes were used to exclude people from the team. In groups with a non-competitive goal, positive competence attributes were sometimes used to include people but for reasons that had to do with having fun and for optimizing a good climate within the team. Arguments relating to warmth were nearly exclusively used in the groups with a non-competitive goal.

Inclusion and exclusion were both important concepts in the groups with a competitive goal. Attributes relating to both competence and lack of competence were communicated in the groups to argue for inclusion or exclusion of people. In groups with a non-competitive goal, exclusion was seldom discussed. There seemed to be an unspoken agreement that all the target persons should be included in the team. As such, when there was a discussion about competence in a group, it seldom took the form of lack of competence. Similarly, to reach a good group climate, discussions about inclusion were generally about positive attributes and not characteristics that might have negative effects on the group climate.

It is important to remember that the groups in this study were randomly formed and not real groups. Participants were students, and groups had 15 min to solve a relatively simple task with no right or wrong answers. It is difficult to tell whether or not the results are applicable to other situations with other tasks, groups, and populations. More research is needed with a variation of tasks and participants. The question of what would happen in a group across time with the same kind of task is also still open. Would the patterns identified become stronger or would some other processes appear? The inter-rater reliability analysis revealed that statements relating to the warmth attribute and a good group climate were more difficult to classify. This indicates that such reasoning is less ordinary in these circumstances, despite the fact that sports associations have both kinds of imperatives as conditions for their task (Vogel et al., 2003). This suggests that the competitive goal seems to be dominant regardless of conditions. Thus, there is a need for applying research to situations where a non-competitive goal is dominant and a competitive one is sub-ordinate. As mentioned in the introduction, research on decision-making has so far most often been carried out within an implicit competitive paradigm (Ahn et al., 2003), and this one-sidedness has seldom been questioned (Kray & Thompson, 2005). This study reports an attempt to address this problem and our results clearly stress the need for research that explores the consequences of different kinds of goals.

A conclusion suggested from these results is that a competitive goal leads to both inclusion and exclusion mechanisms where both positive and negative stereotypes are used, whereas a non-competitive goal leads to inclusion mechanisms and the use of positive stereotypes only. It is interesting to note that in the groups with a competitive goal, no references were made to attributes that related to warmth and creating a good group climate, whereas in groups with a non-competitive goal, there were arguments relating to both warmth and competence. This suggests that groups in a non-competitive condition could not free themselves of the competitive nature of the task and sometimes had to remind themselves that they were supposed to put together a team to maximize the sense of community. This indicates that these groups experienced the dilemma between the two imperatives of performance and the common good, while the groups with the competitive goal did not. This is ironic, as the importance of creating a good team spirit to perform well is so often stressed within sports (and also in many other organizational situations). Our results pinpoint how easily competitive goals gain hegemony over other forms of goals. There is a strong desire to perform well but at the same time this might mean exclusion of participants, which is motivated by the use of negative stereotypes. Our results clearly suggest that stereotypes are easily applied to legitimize decisions about inclusion and exclusion. This is especially problematic from a democratic point of view (Vogel et al., 2003). It has to be acknowledged that both types of goals coexist in many circumstances, but it seems that non-competitive goals need to be recaptured, over and over again, to be taken seriously.