Participants in the alarm condition will report higher levels of cooperation and group processes (perceived emotions, common fate, solidarity, and joint intentions) than in the announcement condition (H1).

—

Regardless of experimental condition, cooperation will be positively associated with group process variables (perceived emotions, common fate, solidarity, and joint intentions) (H2).

—

The relationship between cooperation and group process variables (perceived emotions, common fate, solidarity, and joint intentions) will be stronger in the alarm condition than in the announcement condition (H3).

—

Independent variable

Experimental condition

The experimental condition consisted of two conditions (announcement=0, alarm=1). The auditory signals for each condition were identical to Willemsen et al. (2025). 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 and seven chairs, which hindered their ability to exit the building. The purpose of the obstacle was to strengthen the saliency of the threat. The auditory signal was played over a speaker (JBL Charge 5) situated in the rafters above the participants.

Dependent variable

Cooperation

We measured cooperation using a six-item scale (Willemsen et al., 2025), example item: ‘The other participants and I worked well together’ (𝛼 = 0.84). We removed two items (planned cooperation and leadership) post-hoc. The remaining four items were aggregated into a single cooperation score.

Dependent variable

Perceived emotions

We used 15 emotions to measure the extent to which participants felt each on an individual (self-emotions) and group level (other emotions) (Willemsen et al., 2025). Emotions were clustered into six factors. Three for self-emotions and three for other-emotions. For each factor, we calculated a single emotion score: (a) Anxious emotions (𝛼self-emotions = 0.85 and 𝛼other-emotions = 0.83); (b) Enthusiastic emotions (𝛼self-emotions = 0.84 and 𝛼other-emotions = 0.84); and (c) Confused emotions (𝛼self-emotions = 0.67 and 𝛼other-emotions = 0.62). One item (control) was removed from both self- and other-emotions.

Dependent variable

Common fate

We measured common fate by adapting a three-item scale from Drury et al. (2016) and including one additional item, example item: “It was everyone for themselves” (𝛼 = 0.81). We aggregated items to create a single common fate score.

Dependent 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 (𝛼 = 0.77, Need Threat Scale, van Beest & Williams, 2006). We sampled five items from Leach et al. (2008) to assess identification with the group (𝛼 = 0.89). Lastly, we measured entitativity with other participants using four items (𝛼 = 0.89, Jans et al., 2012). We aggregated the mean scores for each component into a single aggregated score for solidarity for each participant (𝛼 = 0.88).

Dependent variable

Joint intentions

We used the same six items to test joint intention (Willemsen et al., 2025), example item: “We jointly saw to it that the obstacle was cleared” (𝛼 = 0.79). Items were aggregated into a single joint intention score.