When accidents happen, immediate responders or civil response persons are often first to respond, before a professional emergency response (Bakke et al., 2015; Pilemalm et al., 2013; Ramsell et al., 2017). Immediate responders are simply people who happen to be nearby when the accident occurs and decide to help. A civil response person, on the other hand, is a volunteer that has received rudimentary training in emergency response and who is dispatched to provide initial basic response before the arrival of professional emergency response (Pilemalm et al., 2013; Ramsell et al., 2017). Both types of responders face difficult circumstances as they try to provide aid, particularly when there is more than one responder providing aid. As a group they have no or little previous shared history, may lack sufficient resources, and have likely not trained to perform together (Whittaker et al., 2015; Majchrzak et al., 2007). What little research has been done on first aid groups shows that they sometimes, but not always, provide better aid when working together (Takei et al., 2014; Pelinka et al., 2004).

Swift trust has been proposed to be an important enabler for cooperation in emergent ad-hoc groups in dangerous situations (Olsen, 2018) and has been linked to increased first aid performance in an experimental setting (Brodin et al., 2025). Olsen et al. (2020) investigated factors associated with swift trust in first aid using video vignettes and found that an autocratic leadership style together with emotional stability elicited the highest trust. Affiliation with institutional organizations, i.e., an expertise marker, has also been proposed and found to positively influence swift trust (Meyerson et al., 1996; Blomqvist and Cook, 2018; Barrett, 2025). However, this has not been tested in the context of first aid.

Olsen et al. (2020) used video vignettes to investigate swift trust formation. An alternative method for this type of study would be text vignettes. Text vignettes are more resource effective, as they do not require recording the different scenarios. However, as the majority of the people have never experienced an actual first aid event, they may not be able to visualize a first aid scenario with sufficient accuracy based on text only. To explore if text vignettes are a viable alternative for research in a first aid context, the current study will attempt to replicate the results of Olsen et al. (2020) using text vignettes instead of video vignettes.

The current study aimed to replicate the findings of Olsen et al. (2020) and also to investigate the effect of an expertise marker on swift trust in first aid. Two research questions were investigated in relation to the aim: (1) can the result from Olsen et al. (2020) be replicated with text vignettes? and (2) what effect does a marker of expertise have on trust in first aid?

Previous research on swift trust and first aid can be constructed into a theoretical framework (see Figure 1). Swift trust is defined by Meyerson et al. (1996) as trust in temporary systems where the decision to trust was made in an instant due to time constraints and lack of previous shared familiarity. Previous research has identified assigning and clearly communicating roles (Kroeger et al., 2021) and brief verbal interaction (Schilke and Huang, 2018) as fundamental to swift trust development and accuracy. Swift trust formation in professional emergency responder teams has been found to be category-based and to rely on third-party recommendations (Xu and Zhao, 2011). In early emergency response by immediate responders (i.e., people giving first aid in emergencies before the arrival of professional responders), swift trust has instead been posited to develop due to the recognition of actions being done expertly (Majchrzak et al., 2007). In other words, swift trust in emergency response by immediate responders may depend on individuals’ active engagement with the shared task (Meyerson et al., 1996; Barrett, 2025) and the apparent membership of organizations (Majchrzak et al., 2007), i.e., institutional categories (Meyerson et al., 1996; Blomqvist and Cook, 2018; Barrett, 2025). The positive effect of organizational membership and institutional categories on swift trust is shown in the framework through a dashed arrow from expert marker to swift trust separately from leadership and emotional stability, as there was little overlap in the first aid literature (see Figure 1). The line is dashed, as the effect is hypothetical in the context of first aid.

Previous research on swift trust in the first aid context has also found autocratic leadership and emotional stability to increase swift trust in a first aid context (Olsen et al., 2020). Olsen et al. conducted a vignette study where participants rated their trust after viewing a short video with an actor displaying autocratic or democratic leadership, and emotional stability or instability. The results showed that the combination of autocratic leadership and emotional stability elicited the largest trust ratings by participants, while democratic leadership and emotional instability elicited the lowest trust (Olsen et al., 2020). Emotional stability is defined in Olsen et al.’s (2020) as an individual’s tendency to slow arousal and fast inhibition (Eysenck and Eysenck, 1985). Low emotional stability, or emotional instability, would then be quick to arousal and have slow inhibition. Autocratic leadership involves taking control of the decision-making and responsibility for the performance of the group or team while democratic leadership asks for input and shares information and decision-making with the group (Bass and Stogdill, 2008; Hannah et al., 2009; Hyllengren et al., 2011). In the framework, the findings of Olsen et al. (2020) that autocratic leadership and emotional stability elicit more swift trust than democratic leadership and emotional instability are visualized by combining the arrows from leadership and emotional stability before reaching swift trust (see Figure 1).

Olsen et al. (2020) used video vignettes to investigate the role of leadership behaviors on swift trust formation in a first aid context. Video vignettes can convey detailed behavior but can be resource-intensive to develop to a sufficient level of quality (van Zelderen et al., 2024). Comparatively, vignettes using text require fewer resources to develop but may lack sufficient detail to immerse the participants in the scenario. The first aid domain is unfamiliar to the majority of people, and the use of text vignettes puts a lot of responsibility on the participant to visualize the described scenario correctly enough for the study’s purpose. It is therefore of interest to test the viability of text vignettes to investigate psychological aspects of first aid research. The hypothesis for RQ1 on using text vignettes instead of video vignettes to investigate swift trust in first aid is thus (see H1 in Figure 1):

Whereas Olsen et al. (2020) found that leadership and emotional stability influenced swift trust formation, it is unknown how this effect is affected by the presence of an organizational membership or expertise marker (H2 in Figure 1). One expertise organization in the first aid space currently under development is the civilian response person (Swedish: civil insatsperson, or CIP) initiative (Pilemalm et al., 2013; Ramsell et al., 2017). The CIP initiative aims to identify areas of Sweden that are underserved concerning the emergency response capacity in the region. In these areas, local residents are recruited, educated in basic emergency response, and then included as an early local response capacity to provide basic emergency response before the arrival of professional first responders. These CIP responders are provided with some materials, e.g., a high-visibility reflective vest with “civil insatsperson” written on the back. This vest functions as a marker of both the person’s expertise in emergency response and affiliation with the professional emergency response organization. Thus, the vest functions as an expertise marker since it explicitly communicates the membership of the CIP initiative and emergency response organization (Meyerson et al., 1996; Blomqvist and Cook, 2018; Barrett, 2025). Therefore, the hypothesis for RQ2 regarding the effect of an expertise marker on swift trust in first aid is as follows (see H2 in Figure 1).

A text-vignette study with a 2 (leadership and emotional stability) by 2 (expertise marker) design was conducted. The two leadership and emotional stability conditions were based on the study by Olsen et al. (2020). Olsen et al. used the separate dimensions of leadership (autocratic versus democratic) and emotional stability (stable versus unstable) for their four conditions. The current study is a partial replication, as only the condition of autocratic leadership combined with emotionally stable behavior and the condition of democratic leadership with emotionally unstable behavior are included. These conditions were selected as they showed the largest differences in reported trust in Olsen et al. (2020). The second dimension added in the current study was the inclusion of a visible expertise marker, see Figure 2 for conditions.

The text vignettes in this study are based on the video vignettes used by Olsen et al. (2020), where each vignette is a textual description of the same scenario and behavior as the corresponding video vignette. An expertise marker was added to this description. The text vignettes were validated by eight emergency response experts to ensure that each condition entailed plausible behavioral descriptions. A convenience sample of 93 participants (43 females, 50 males, and 1 preferred not to answer) with an average age of 28.7 years (SD = 9.9) was then recruited from a university in southern Sweden. Participants received no monetary compensation. The study entailed no risk, danger, collection of sensitive personal data, or physical or psychological manipulation, and thus no prior ethical approval was required according to Swedish law.

The Mann–Whitney U-tests for leadership and emotional stability ratings were significant. Leadership was rated significantly different in conditions with autocratic leadership and emotional stability (condition 1 & 2; N = 41, Mdn = 2) compared to conditions with democratic leadership and emotional instability (condition 3 & 4; N = 37, Mdn = 4) U = 120, p < 0.01, r = −0.23. Emotional stability was rated differently in conditions with autocratic leadership and emotional stability (conditions 1 & 2; N = 44, Mdn = 5) compared to conditions with democratic leadership and emotional instability (conditions 3 & 4; N = 46, Mdn = 2) U = 155, p < 0.01, r = 0.2. This indicates that vignettes 1 and 2 were rated as more autocratic and emotionally stable than vignettes 3 and 4, whereas vignettes 3 and 4 were rated as more democratic and emotionally unstable.

In the 2×2 ANCOVA on the effect of leadership, emotional stability, and expertise marker presence on rated trust, the covariate propensity of trust was found to be significantly related to trust, F(1, 88) = 8.4, p = 0.005, r = 0.23. There was also a significant main effect of leadership behavior and emotional stability on trust F(2, 88) = 56.1, p < 0.001, partial η² = 0.39, and a significant interaction effect between leadership behavior and emotional stability and the presence of an expertise marker, F(2,88) = 5.29, p = 0.024, partial η² = 0.06. A post-hoc analysis comparing all four conditions with each other found four significant differences between conditions (see Table 2; Figure 3).