Hypothesis 1: Spontaneous cooperation Hypothesis 1a – Participants in the experimental condition will report higher levels of cooperation compared to participants in the control group. Hypothesis 1b – Groups in the experimental conditional will cooperate to a greater extent compared to groups in the control condition.

—

Hypothesis 2: Shared emotions Hypothesis 2a – Participants in the experimental condition will report stronger individual emotions and perceive stronger shared emotions compared to participants in the control condition. Hypothesis 2b – Across both conditions, the similarity between participants’ self-reported individual emotions and their perceived shared emotions will be positively associated with reported levels of cooperation. Hypothesis 2c – The relationship between emotional similarity and cooperation will be stronger in the experimental condition than in the control condition.

—

Hypothesis 3: Common fate Hypothesis 3a – Participants in the experimental condition will perceive higher common fate compared to participants in the control condition. Hypothesis 3b – Across both conditions, common fate will be positively associated with reported levels of cooperation. Hypothesis 3c – The relationship between common fate and cooperation will be stronger in the experimental condition than in the control condition.

—

Hypothesis 4: Solidarity Hypothesis 4a – Participants in the experimental condition will perceive higher solidarity (belonging, identification and entitativity) compared to participants in the control condition. Hypothesis 4b – Across both conditions, solidarity will be positively associated with reported levels of cooperation. Hypothesis 4c – The relationship between solidarity and cooperation will be stronger in the experimental condition than in the control condition.

—

Hypothesis 5: Joint intentions Hypothesis 5a – Participants in the experimental condition will report greater joint intention compared to participants in the control condition. Hypothesis 5b – Across both conditions, the primary goal (I-goal versus We-goal) will moderate the positive association between joint intention and reported levels of cooperation. Hypothesis 5c – The relationship between joint intention and cooperation will be stronger in the experimental condition than in the control condition.

—

Independent variable

Experimental manipulation

The experimental manipulation consisted of two conditions (control=0, experimental=1) with different auditory signals. In the experimental condition, participants heard a fire alarm and the verbal instructions: “Attention! This is an emergency. Please follow the emergency exit signs and immediately leave the building.” The purpose of the experimental condition was to simulate a collective threat event (i.e., a building evacuation). As participants were leaving, they encountered an obstacle consisting of two tables, four chairs and five metal recycling containers which hindered their ability to exit the building. The purpose of the obstacle was to strengthen the saliency of the threat. The control condition followed the same procedure, including the obstacle, but the auditory signal was an announcement bell and the verbal instructions: Attention. This task has now ended. Please leave your cubicles and follow the exit signs to go out.” The auditory signal was played over a speaker (Ultimate Ears Wonderboom 3) situated near the exit.

Independent variable

Perceived shared emotions

We chose 15 emotions and measured the extent to which participants felt each on an individual (self-emotions) and group level (other-emotions). Participants rated their agreement on a 7-point scale (-3-Strongly disagree, 3-Strongly agree) with the following statements: “As I heard the auditory signal and got up to leave, I felt [alarmed]” and “In my view, the other participants felt [alarmed]” (𝛼self-emotions = 0.66, 𝛼other-emotions = 0.72). Participants were also asked to list other emotions that were not included but that they felt.

Independent variable

Common fate

We measured common fate using a 3-item scale from Drury (2016) and included two additional items (5-items), example item: “We were all in this together” (𝛼 = 0.80). Participants rated their agreement on a 7-point scale (-3-Strongly disagree, 3-Strongly agree).

Independent variable

Solidarity

We measured three components of solidarity (belonging, identification, and entitativity) (Van Mourik Broekman et al., 2018). We measured belongingness to the groups with five items, example item: “I felt as one with the others” (𝛼 = 0.82, Need Threat Scale, Van Beest & Williams, 2006). We sampled five items from Leach et al., (2008) to assess identification with the group, example item: “I identify with the others in the group” (𝛼 = 0.81). Lastly, we measured entitativity with other participants using four items, example item: “I feel that the others and I are a unit” (𝛼 = 0.89, Jans et al., 2011). For all solidarity components, participants rated their agreement on a 7-point scale (-3-Strongly disagree, 3-Strongly agree). We then aggregated the mean scores for each component into a single aggregated score for solidarity for each participant (𝛼 = 0.84).

Independent variable

Joint intention

We measured joint intention in two ways. The first consisted of a single question aimed at understanding participants’ primary goal when encountering the obstacle. Participants responded on a bipolar slider from “I wanted to get myself out” (-3, I-goal) to “I wanted to get everyone out” (3, We-goal). We then designed six items to test joint intentions. Participants rated their agreement on a 7-point scale (-3-Strongly disagree, 3-Strongly agree), example item: “We jointly saw to it that the obstacle was cleared” (𝛼 = 0.81).

Dependent variable

Cooperation

Cooperation was measured in two ways: 1) self-report; and 2) video/ audio data. For the self-report measure, we designed six items to measure cooperation, example item: ‘The other participants and I worked well together’ (𝛼 = 0.76). Participants rated their agreement on a 7-point scale (-3-Strongly disagree, 3-Strongly agree). Video and audio data were collected using four Lamax X3.1 Atlas video cameras. The videos had a wide frame of view with a video resolution of 1080 pixels and a frame rate of 30 frames per second. The cameras were positioned just below the ceiling and angled to capture as much of each room as possible. There were two cameras situated in the room outside the cubicles and two cameras situated in the room with the obstacle (see supplementary material for layout). The recordings took place in a laboratory setting. The video and audio recordings were transcribed verbatim. We then developed an ethogram (see Sieben & Postmes, unpublished) to analyze observed behaviors. Inter-rater reliability was assessed, and discrepancies were resolved through discussion.