The following is an excerpt of the intervention methodology. For more information, please see the full text of the article on the publisher's website or through your institution's library.

STUDY 1: CROSSED VS. CONVERGENT BOUNDARIES

Methods

Subjects. Subjects were 458 graduate students in management participating in a school orientation activity intended to introduce them to group decision making. [...]

Design. The independent variable consisted of two conditions of boundary configurations. In one condition category and group boundaries crossed within teams, while in the other condition these boundaries were convergent. Dependent variables included measures of group and team identification, intergroup bias, intergroup conflict, and team performance.

Procedure.

At the beginning of the experiment, the subjects in each session were divided into two categories, with each category having approximately the same number of subjects. [...]

After categories had been formed and the experimenter introduced the topic, the exercise materials were handed out. The materials were copied on paper the same color as the ribbons, so that each category had materials of a different color. The exercise was a version of the Sub-Arctic Survival Task [...]. [...] In the background information, subjects were told to imagine that they were passengers on a plane that had crashed in the Canadian wilderness during the winter. [...] They were told that they salvaged 15 items from the crash (e.g., magnetic compass, 1 hand axe, bottle of water purification tablets, etc.) and their task was to rank these items based on their importance for survival (1=most important). [...] These rankings would be scored by comparing them to experts' rankings. [...]

After reading the materials, each subject was paired with one other person to form a two-person task group. In the convergent condition, the two people were from the same color category, while in the crossed condition, the two people were from different color categories. [...]

After the two-person groups completed their ranking, subjects individually answered questions about category identification, group identification, and intragroup conflict. [...] Subjects were asked "how strongly do you identify with your (green or red) color category?" on a scale of 1 (not at all) to 7 (very strongly). They were also asked to rate the statement "I identify with this group" on a scale from 1 (strongly disagree) to 7 (strongly agree). Intragroup conflict was measured by asking "how much conflict was there in your group?" on a scale of 1 (none at all) to 7 (a great deal).

[...] The team task was to rank the same 15 items. Each group was then paired with another group to form a four-person team. In forming the teams, each crossed group joined another crossed group so that the category and group boundaries were crossed within each of these teams. Each convergent group joined a convergent group from the other category so that the category and group boundaries were convergent within each of these teams. [...]

After subjects completed the team task, they individually completed a postexperimental questionnaire containing several items. [...] To measure intergroup bias, subjects were asked to separately rate each other person on the team on how competent, cooperative, valuable to the discussion, competitive, friendly, trustworthy, and similar to you they were. Each attribute was rated for each other team member on a scale from 1 (not at all characteristic) to 7 (extremely characteristic). Performance evaluations were collected by asking subjects to rate team members' individual contributions to the team product by assigning each other member from 0 (worst contribution) to 100 (best contribution) points. The ratings were independent (e.g., every member could receive 100, or every person could receive 0). Self-reports were gathered for the amount of conflict that occurred during the team interaction. Subjects were asked "how much conflict was there in your team?" and "to what extent were there differences of opinion in your team?" (1 = none at all; 7 = a great deal). Measures of time taken to complete the task and team performance were also collected. [...]

STUDY 2: NESTED SOCIAL DILEMMAS

Methods

Design. A 4 x 2 between-subjects factorial design was employed to test the hypotheses. Independent variables were common fate (4 conditions) and organizational status (2 conditions). Subjects were randomly assigned to their experimental condition. Dependent variables were the allocation of $5.00 among subjects' private, subgroup, and collective accounts, and estimations of allocations for a typical member of the subject's own and other subgroup.

Procedure. Each subject belonged to a natural subgroup nested within a natural collective. Specifically, the subjects were students in a business school (the collective) and each student belonged to a "section" (the subgroup) defined by, and nested in, the business school. [...]

Students from six sections participated in the study. Each section was anonymously paired with one other section to form the two subgroups making up the collective. Thus, each pair of sections comprised a nested social dilemma that was completely separate from the other pairs. Subjects did not know with which other section their section was paired. The business school group boundary encompassing both sections served as the collective.

Each subject was endowed with five dollars and given the task of allocating this money among three accounts. Each subject had his or her own private account; dollars allocated to the private account were multiplied by 1, with the money accruing to the individual holder of the account. Subjects could also allocate to their subgroup account, which included all of the participants from their section. Dollars allocated to the subgroup account were multiplied by 2.5, with the resulting pool of money divided equally among all participants in the section. The collective account included all participants from both sections. Each dollar allocated to the collective account was multiplied by 4, with the resulting pool of money divided equally among all participants from both sections. Thus, each subject had a personal total payoff that was the sum of 1) the amount in their private account, 2) their equal share of the amount in their subgroup account, and 3) their equal share of the amount in the collective account.

These three allocation options presented subjects with a dilemma. On one hand, regardless of the accounts to which other participants allocated their money, each individual maximized his or her outcome by allocating all five dollars to his or her private account. [...] On the other hand, the collective profit (i.e., the total amount of money made by everyone participating in the exercise) was highest if every participant allocated to the collective account, lower if every participant allocated to their subgroup account, and lowest if everyone allocated to their private account. [...]

Common fate manipulation. After all decisions had been made, a personal total payoff was calculated for each subject. [...] Subjects received payment if they won a coin flip. In the individual fate condition, subjects were told that a coin would be flipped separately for each subject. In the collective common fate condition, subjects were told that a single coin would be flipped to determine whether or not all participants in the collective would receive payment. In the subgroup common fate condition, subjects were told that a separate coin would be flipped for each subgroup. A fourth condition created a common fate at the subgroup level, but added a psychological analogue of intergroup competition by creating an "us against them" mentality. In this condition, subjects were told that one coin would be flipped to determine whether the participants in their subgroup qt the other subgroup would be paid. [...]

Status manipulation. The collective was a natural group with high status stemming from several sources, most notably its top position in rankings of business schools. This naturally-occurring high status was evoked in one condition and not evoked in the other condition.

STUDY 3: NESTED SOCIAL DILEMMAS

Method

Subjects. Subjects were 99 graduate management students from Northwestern University and the University of Chicago [...]. [...]

Design. A 2 x 4 between-subjects factorial design was used to test the hypotheses. The independent variables were common fate (2 conditions) and subgroup composition (4 conditions). Dependent variables included subjects' individual allocation decision and their predictions of their own and the other subgroup members' allocations.

Procedure.

The exercise described in study two was used in this study with a few modifications. [...] In the current study, each subject had to decide how to allocate five dollars among their private account, subgroup account, and collective account. The multipliers for dollars allocated to each account were the same as in study two (1 for the private account, 2.5 for the subgroup account, and 4 for the collective account), and each subject got a personal total payoff that was the sum of their share of each account.

Operationally, the current study differed from study two in two important ways. First, the size of the subgroups and collective were held constant; each subgroup consisted of three people and each collective of six. The members of each subgroup and collective were completely anonymous to each other, with the exception of knowing some information about their real group memberships. Second, rather than recruiting subjects through their student mailboxes, the exercise was run in four different classes (two classes at each business school). [...]

Common fate manipulation. [...] In the intergroup competition condition, a coin was flipped to determine which subgroup's members would receive their payoffs. [...] In the condition with no intergroup competition, a coin was flipped separately for each subgroup to determine whether the members of each subgroup would receive their payoffs. [...]

Subgroup composition manipulation. [...] In all conditions, the subject's own subgroup consisted of the subject and two other anonymous people from the subject's own class. The composition of the other subgroup was manipulated. In the four conditions, the three members of the other subgroup were specified as anonymous members of A) the subject's own class, B) another class in the same business school, C) a group outside of [the subject's business school], or D) a class in [the other business school]. These four conditions were intended to manipulate the psychological distance between the members of the subject's own subgroup and those of the other subgroup, with psychological distance increasing from condition A to condition D. [...]