“That’s how the trust began”: Forming trusting friendships from adolescence to adulthood

Adolescence as a sensitive period for the development of trusting behavior

As adolescents begin forming close relationships with peers, the nature of their friendships transforms. While childhood friendships center around companionship and shared hobbies, mutual trust becomes more salient from adolescence to adulthood as friendships shift to fulfill the needs for security, intimacy, and support (Blieszner & Roberto, 2004). Adolescents also increasingly differentiate friends from neutral peers (Güroğlu et al., 2014) or strangers (Güroğlu et al., 2014; Sweijen et al., 2023) when choosing towards whom to behave in a trusting manner. If successfully navigated, these changes foster the development of trust-relevant skills, like mentalizing (Desatnik et al., 2023; Güroğlu, 2022), differentiating aspects of the social world (e.g., distinguishing between generalized trust and trust in specific others; Flanagan & Stout, 2010), or managing negative emotions (Riediger & Klipker, 2014).

The ability to flexibly adapt one’s trusting behavior based on the course of the interaction also markedly improves from early adolescence to adulthood, as evidenced by findings from trust game research. In this experimental paradigm, the trustor can keep the entire initial endowment or transfer any amount to the trustee. The trustee receives the amount multiplied by the experimenter, choosing to keep or share the increased endowment with the trustor (Lewicki & Brinsfield, 2012). In the study by van den Bos et al. (2012), early adolescents, middle adolescents, and young adults played the trust game with interaction partners who differed in how cooperative they were in sharing money. All age groups noticed that their counterparts differed in trustworthiness. However, only middle adolescents and young adults adapted their behavior to the behavior of their partners; that is, they increasingly shared more with trustworthy partners and less with untrustworthy partners. Similarly, participants in Sijtsma et al.’s (2023) longitudinal study decreased their cooperation with untrustworthy trust game partners across three consecutive years spanning early to middle adolescence.

In line with the developmental task perspective, which considers the formation of close friendships a developmental task of adolescence, sufficiently developing the abovementioned skills in adolescence continues to influence the quality of relationships into adulthood (Masten & Coatsworth, 1998). For example, Allen et al. (2020) found that social competence demonstrated by forming and sustaining close friendships in adolescence uniquely predicted romantic life satisfaction in adulthood. Adolescents who can manage the socio-emotional demands of these new friendships can rapidly develop abilities crucial for establishing close, trust-based relationships across the lifespan.

Understanding the trust process

Notwithstanding these developmental changes, the current theorizing postulates that the trust process leading to trusting behavior follows a universal trust sequence model (Dietz, 2011; Dietz & Den Hartog, 2006; Six & Latusek, 2023). People first assess other persons’ trustworthiness to form expectations about their behavior. Positive expectations culminate in the decision to trust (i.e., to render oneself vulnerable). For trust to materialize in trusting behavior, such as sharing sensitive information, one must take the leap of faith to trust, that is, accept the uncertainty and vulnerability that the positive expectations may have been erroneous (cf. Möllering, 2006; Rempel et al., 1985). The outcomes of trusting, such as whether the other person keeps or shares sensitive information, affirm or modify trustworthiness beliefs, and, in turn, the tendency to trust the same person again. One will be more inclined to share another secret with someone who has kept previous secrets to themselves than someone who disclosed them. On the contrary, if the other person seems untrustworthy and people hold negative expectations about their behavior, they withhold the decision and the leap of faith to trust. As a result, they refrain from acting in a trusting manner or commence protective actions (e.g., limiting contact with the untrustworthy person; Six & Latusek, 2023).

Nonetheless, the trust process may not be universal. Prior experiences with trusting others can carry over into new relationships, altering how people progress through the trust process, not only with a particular person but also in general. The reinforcement learning perspective suggests that previous experiences with trusting other people influence individuals’ willingness to be trusting in new relationships. For some, initially trusting and taking weightier risks might be more rewarding as it does not prematurely preclude possibly high-quality connections. Others may find their comfort level in remaining more reserved, for example, after being hurt in previous relationships (Schilke et al., 2021). Likewise, according to the social learning theory, experiences gained in particular situations affect one’s expectancies and trusting behavior in comparable situations (Landrum et al., 2015; Thielmann & Hilbig, 2015). For example, positive experiences of confiding in a peer can strengthen one’s willingness to trust peers with vulnerable information and assume a more open and less self-protective stance when confiding in peers in the future. These effects of reinforcement and social learning can be particularly pronounced in adolescence, which is characterized by a relatively high sensitivity to social rewards (Güroğlu, 2022). Whether there are individual variations in the trust process and how they may unfold thus warrants empirical exploration.

The common methods in current trust research allow limited insights into the trust process, as they typically focus on other aspects of trust (e.g., trustworthiness) or are not best suited for studying processual phenomena. The existing questionnaire-based research has extensively concentrated on trustworthiness, especially on evaluating the other person’s characteristics (McAllister, 1995; Rempel et al., 1985), but only a few questionnaire studies tapped into intentions to act in a trusting manner (e.g., Gillespie, 2012), and even fewer included items assessing trusting behavior (e.g., Breuer et al., 2020). In addition, questionnaires offer only a static snapshot of trust based on pre-defined categories. As such, they are of limited use in exploratory research focusing on the process of trust formation because they do not capture the full complexity and temporal dynamics of the trust process (i.e., how the course of the trust process may vary and change over time). The other line of empirical research employing trust games is limited to a single trust-relevant situation – cooperation on dividing resources – and is highly structured, offering participants little room for variability in handling the situation. Furthermore, while people distinguish between financial risk and the no less costly social betrayal in trust games (Alós-Ferrer & Farolfi, 2019), the presumed personal risk and vulnerability when engaging in trusting behavior in close relationships are much higher. Thus, the trust game does not allow the exploration of the broader patterns of trust in real-life situations and the establishment of trust in others in the longer term beyond a limited series of interactions.

The present study

The trust sequence model posits that people decide to trust others if they appraise them as trustworthy while accepting the risk that their assessment may be incorrect. It, therefore, presumes that people undergo the trust process uniformly (Dietz, 2011; Six & Latusek, 2023). On the other hand, the social learning theory (Landrum et al., 2015; Thielmann & Hilbig, 2015) and reinforcement learning perspective (Schilke et al., 2021) jointly suggest that people differ in how they navigate the trust process based on prior trust experiences and their outcomes. However, empirical understanding of how the course of the trust process may vary remains limited, especially during adolescence, which marks a critical period for initiating trust-based friendships (Blieszner & Roberto, 2004) and trust-relevant abilities (Allen et al., 2020; Desatnik et al., 2023). The present study sought to fill these gaps by qualitatively exploring variations in the trust process in the context of friendships. We analyzed the data collected from focus group interviews with early, middle, and late adolescents and a comparative sample of adults using reflexive thematic analysis to address the following research question: How do people experience the process of forming trust in friends from early adolescence to adulthood?

Study design

The present study was a secondary analysis of qualitative data from a large-scale focus-group investigation of meanings and sources of trust and distrust in institutional and interpersonal contexts. The first part of focus group interviews explored trust in authorities and institutions during the COVID-19 pandemic, with the second part concentrating on trust in close others. The large-scale project aimed to collect rich data for use in a descriptive project report focused on sources of trust and distrust and for further analyses, with reflexive thematic analysis as the intended analytical approach.

Participants

There were 39 Czech participants, divided into four age groups: early (11-12 years, n = 11), middle (14-15 years, n = 9), and late (18-19 years, n = 8) adolescents and adults (30–50 years, n = 11). All participants were White, reflecting Czechia’s population, which is largely racially and ethnically homogenous. See Table 1 for details about participants’ age, gender, educational status, the educational status of adolescents’ parents, and community type.

OPEN IN VIEWER

Table 1.

Participant characteristics.

Characteristics	Age group
	11–12 (M = 11.55, SD = 0.50)	14–15 (M = 14.33, SD = 0.47)	18–19 (M = 18.25, SD = 0.43)	30–50 (M = 37.27, SD = 5.41)
Gender
Women	5	5	4	5
Men	6	4	4	6
Highest education level
Elementary school	9	5	–	–
Grammar school (lower) a	2	4	–	–
Grammar school (upper) a	–	–	4	–
High school	–	–	4	4
University or college	–	–	–	7
Mother’s highest education level
High school	3	3	4	–
University or college	8	6	4	–
Father’s highest education level
High school	4	1	4	–
University or college	7	8	4	–
Community type
Rural	4	3	2	3
Suburban	1	3	3	1
Urban	6	3	3	7

Procedure

We held eight online semi-structured focus group interviews in total, two per age group, from June to September 2022. Participants were recruited through social media and personal contacts using non-probability sampling. We balanced the focus groups by gender, education level, and community type. We used a dual moderator approach, with each focus group featuring one main moderator and an auxiliary moderator (Krueger & Casey, 2000). The average duration of a focus group was 68 minutes (Min = 63 minutes, Max = 74 minutes). In this study, we analyzed the second part of the focus group interviews, which lasted on average 25 minutes (Min = 19 minutes, Max = 30 minutes), focusing on the participants’ answers to the following questions:

1. “Up to now, we have talked a lot about COVID and the government. Now, we would like to know how you feel about people you know. Try to imagine a person that you trust. You do not have to say who it is; just picture this person. Could you tell us why you trust this person?”

We coded participants’ responses to identify people they recalled. All participants spoke primarily about friends, with a few participants also mentioning family members. In this analysis, we focused on the excerpts about friends.

Ethics

All participants included in the study signed an informed consent. For the participants younger than eighteen, we also obtained informed consent from their parents. We removed identifying information from the transcripts. We identified participants by the number of their respective focus group (e.g., FG1) and pseudonyms (e.g., Carl). Importantly, the participants agreed that their anonymized data could be used in multiple research subprojects for secondary data analysis. The Research Ethics Committee of Masaryk University approved the research (approval No. EKV-2019-072).

Data analysis

The flexibility of reflexive thematic analysis allowed us to study patterns of the variations in the trust process, adopting an inductive, semantic, and experiential approach epistemologically underlined by tempered realism (Braun & Clarke, 2022), which we found suitable for studying an underexplored phenomenon. Our coding was, therefore, driven by the data rather than existing theoretical constructs, focusing mainly on exploring the meaning at the explicit level, even though we shifted toward generating more latent codes as we progressed with the analysis. We aimed to capture and explore people’s lived experiences under the tempered realist assumption that while the reality of participants’ experiences and meanings are reflected in their language in a largely uncomplicated manner, researchers and participants influence each other through their values and assumptions.

We followed Braun and Clarke’s (2022) six-phase guidelines: (1) familiarization with the data, (2) initial coding, (3) generating initial themes, (4) developing and reviewing initial themes, (5) defining, refining, and naming themes, and (6) producing the final report. We used MAXQDA (VERBI Software, 2024) to analyze data and store analytical memos. All authors were pre-familiarized with the data as they took part in the data collection and processing, with the second and fourth contributing authors also interviewing the participants. The first and second authors re-read the transcripts multiple times before the analysis to refamiliarize themselves with the raw data. The first and second authors worked in pairs to generate and refine codes and to generate, develop, refine, and name the themes. The first author generated the initial codes and candidate themes, with the second author reviewing the codes and themes and suggesting changes. The first two authors then met regularly to exchange the MAXQDA files, coding schema, and theme delineations; discuss and refine the codes and themes; resolve discrepancies; and check the themes against the data. The first two authors also selected participant quotes and produced write-ups. The third and fourth authors assessed the themes and write-ups for clarity, non-redundancy, and appropriateness. For an illustrative example demonstrating the theme generation process, see Table S1 in Supplementary Materials.

We followed the procedures outlined by Nowell et al. (2017) for conducting trustworthy thematic analysis. We employed prolonged engagement with the data, researcher triangulation, an audit trail of code and theme generation, and a reflexive journal. We used memos in the transcripts to track coders’ emotions, attitudes, experiences, or prior knowledge of the topic. Diagramming was used to explore connections within and among themes. Finally, we described the coding and analytical process in detail. Regarding the secondary nature of the conducted analysis, we adhered to the recommendations by Ruggiano and Perry (2019). All authors of this study collaborated on the large-scale data collection, the analytical approach in this study aligned with the methodology of the large-scale data collection, and participants explicitly agreed to the data being used for secondary analyses.

Experiential Trust Building

The central feature of Experiential Trust Building was that trust increased gradually based on ongoing positive interactions. Participants acquired the positive experiences either in a slower and more spontaneous manner through (a) Accumulating Positive Experiences or faster and more deliberately through (b) Trust Testing. Both subthemes were present across all age groups.

Absence of Reasons to not trust

The core of this trust process variation lies in trusting by default unless the other person gives one sound reason not to, “usually trying to trust people from the beginning until… well until those people lose my trust one way or another” (Max, 31yo, man, FG5). The initial trust typical of this variation can be viewed as a general tendency to trust others, as portrayed in the quote by Sofie (18yo, woman, FG8):

I sort of have this approach that I’m probably too trusting, but I trust even strangers until they prove to me that I shouldn’t trust them. (….) Because I simply hope that people are kind and mean well. (…) I just feel like, why not trust people? I mean, if they give you the reason [not to], then of course, don’t [trust them].

Rather than building trust through acquiring positive experiences, trust was deepened when negative experiences did not come up as the relationship progressed: “I have known them for a long time, but the main thing is, they have never given me a reason not to trust them” (Cyril, 14yo, man, FG4). The participants most mentioned the lack of disappointments and focused on the fact that their trusted friends had not done anything to taint the trust they had placed in them, discussing the overarching “experience with that person [the friend] that they have never failed my trust” (Karin, 36yo, woman, FG2). Increasing trust was, therefore, not a matter of what the friend had done but rather of what the friend had not done.

The adult group emphasized not only needing reasons to not trust but also needing multiple disappointments to diminish trust or warrant distrust. Benny (38yo, man, FG2) described: “It’s more about the experiences. I trusted him [the friend] twice, it didn’t work out twice, so no more [trust], right.” Adults also considered intentionality and expressed greater tolerance for mistakes, which was not present in the adolescent age groups, while deeming it acceptable to cease contact after several letdowns:

Agnes (46yo, woman, FG2): Many times, a person makes mistakes, makes a misstep, that can happen to anyone. But if I feel it was intentional or it happens repeatedly. In that moment, it [the trust] goes down and down.

Relying on Intuitive Connection

Several adolescents used an affect-based strategy when forming trust in friends, relying primarily on their intuition. Immediacy was inherent in this variation of the trust process, with trust emerging after only “knowing the person [the friend] briefly… very briefly” (Johana, 14yo, woman, FG6). Participants commonly used the phrase “we clicked” to portray the intuitive sense that allowed them to quickly establish trust based on a feeling of connection: “We somehow immediately clicked, and I just felt like I could trust them” (Beata, 14yo, woman, FG4). How the feeling came about was “hard to explain” (Magda, 12yo, woman, FG3) for some participants, further demonstrating that trusting the other person, in this case, was not something the participants had deliberately thought about but something that came with easiness and they just felt, as Beata said.

At the same time, trust based on an immediate feeling of connection sparked discussion, with a couple of adolescents expressing that quickly forming trust based on intuition may be possible, although difficult to imagine, as Dita (14yo, woman, FG4) conveyed:

How much you click with the person... (…) I mean, I think it’s possible, but I haven’t experienced it myself yet, trusting a person after a short period. I think, at the minimum, always after we get to know each other properly. So maybe a year, sometimes even two [years] before I would be able to tell [them] anything, a bigger secret or something.

Similarly, Cyril (14yo, man, FG4) added: “I don’t get to know the person all at once. Like that, I would immediately click with them. I must get to know them first to confide [in them] or something [like that].”

Even though two other adolescents did not rely on the immediate connection when forming friendships, they used negative gut feelings as a warning signal: “It also protects us from something, and it’s good to listen to it. It’s just that when we feel inside that we don’t trust someone, it’s wrong to trust them” (Kurt, 12yo, man, FG3). Ondra (18yo, man, FG8) described protective actions, such as “not talking about more confidential stuff with that person,” should follow.

Regarding the developmental aspect, this theme depicts a variation of the trust process that was conceivable only for some adolescent participants who considered it a way to either swiftly start or refrain from trusting. Adults, on the other hand, did not mention affect-based trust at all.

Trusting the Right/Wrong Person

For several adolescents who centered on their friend’s character, the course of the trust process hinged on choosing or less deliberately encountering the “right” or the “wrong” person. Many adolescents also drew the dichotomy between trusting “good” versus “bad” people. Corroborating the essentialist stance that individuals possess personality traits that make them innately right and good or wrong and bad, there was an air of certainty that the other person would eventually reveal their true nature:

We thought that this person was either good or maybe even [had] some other positive… another positive quality. (…) We believed they were a certain way, but when it turned to what they were really like, it disappointed us and crushed us. (Magda, 12yo, woman, FG3)

Trusting the wrong people was viewed as having detrimental consequences. For example, “when we don’t confide in a good person, it can further add to our burden and make things even worse” (Beata, 14yo, woman, FG4). For some participants, trusting the wrong person shed a negative light on the trustor as well, as Dušan (14yo, man, FG6) advocated: „I think it’s wrong to trust bad people or people who shouldn’t be trusted. (…) One mustn’t be naïve. One must not trust everyone and immediately.” Trusting the wrong person was, therefore, not only linked to being harmed but also considered an indication of one’s poor judgment.

Participants differed in the amount of control they perceived to have over trusting the right person. The youngest adolescents assumed a more passive role, expressing the belief that trusting the right or wrong person remains largely a matter of chance. Artur (12yo, man, FG1) aptly described this perspective: “Trust is like a lottery – you either bet on a good person whom you can trust, and you know they won’t disappoint you, or you bet on someone who… whom you trust, and then they betray you.” In contrast, older adolescents offered a rationalized view and were more proactive, underscoring the necessity of “finding the right person” (Laura, 15yo, woman, FG4) or “choosing the right people” (Johana, 14yo, woman, FG6). As such, they believed they could foresee and prevent the negative consequences of trusting a bad person: “When we know that the person is not entirely good and could do us dirty or give us bad advice, I would rather not trust them, and I wouldn’t take their advice” (Anežka, 18yo, woman, FG8). Older adolescents thus placed a high value on their agency, positing themselves as the choosing subjects. Ascribing greater importance to chance also arguably entailed more uncertainty and risk than feeling relatively assured in one’s ability to rationally decide whether to trust based on the assessment of the other person’s personality.

While the participants who elaborated on this variation of the trust process differed in their emphasis on rationality and perceived control, their view was overall quite reductionist in its narrow focus on personality characteristics and on dichotomizing others into two categories – as either good or bad based on their true nature. Moreover, the previous two patterns of the trust process were relational, i.e., trust was construed as a dynamic and reciprocal process formed within the relationship between two people. On the other hand, the relational aspect was largely absent in this final theme. Labeling the friend as an inherently good person worth trusting was also exclusive to adolescents.

Limitations and future directions

Even though the data were intended for secondary analyses and the data collection methodology aligned with our analytical approach, the secondary nature of the analysis entails disadvantages. First, the moderators did not concentrate on inquiring about the trust process. Second, the first part of the focus group interviews dealt with trust in authorities during the COVID-19 pandemic, with the second part exploring interpersonal trust. The early questions may have prompted abstract reflections on trust and increased awareness of trust-related risks as participants discussed trust violations committed by the authorities during the pandemic. However, no participants referred to the pandemic in the second half of the interviews. We purposefully employed the imagination task instruction (detailed in the Method section) to facilitate the topic transition. These design aspects should be considered when planning similar research. From a practical standpoint, the second half of the interviews was also affected by participant fatigue, which was most apparent among the youngest adolescents. All participants nonetheless remained engaged throughout their respective focus groups. While the focus group format sparked fruitful discussion, it may have been more difficult for some participants to talk about vulnerable experiences or deviate from the prevailing stance when in a group as opposed to an individual interview.

While our findings suggest that adolescents and adults might differ in relying on feelings and moral essentialism when forming trust in friends, more studies are needed, as our design did not allow us to establish the magnitude of identified differences. Future research should also explore the process of refraining from the leap of faith into trusting others. A better understanding of how people avoid entering trust-relevant situations is of great importance because the process of not trusting or distrusting may also vary and, in some cases, prematurely preclude the trust diagnostic opportunities for learning whether the other person can be trusted or not (Rempel et al., 1985; Six & Latusek, 2023). Our study provided detailed insights into the experiences of Czech adolescents and adults. Czechia is characterized by moderate generalized interpersonal trust (European Social Survey European Research Infrastructure, 2023) and a highly homogenous racial and ethnic composition. However, variations of the trust process or their commonality may be influenced by cultural factors. We encourage further research to explore the trust process in low- and high-trust societies and to include participants from diverse ethnic and racial backgrounds. We did not collect data on participants’ sexual orientation and disabilities; thus, we encourage future studies to consider these characteristics. Finally, investigating the impact of individual differences on the trust process would also be valuable.

Conclusion

This study advances existing research by empirically exploring the variations of the trust process within friendships, focusing on adolescence as a crucial period for embarking on first trust-based relationships with peers (Blieszner & Roberto, 2004) and acquiring skills necessary for forming and maintaining trusting relationships (Allen et al., 2020). The qualitative design allowed us to gain unique insights into adolescents’ and adults’ experiences with forming trust in friends. Unlike extant trust research, which often focuses on experimental situations and intentions, we explored a real-life context and past actions. Our findings extend the most recent version of the trust sequence (Six & Latusek, 2023) by describing several variations in the course of the trust process. Additionally, this research contributes to the growing evidence regarding decreases in moral essentialism from childhood to adolescence (Heiphetz, 2020) and underscores the role of affect in adolescents’ close peer relationships (cf. Riediger & Klipker, 2014; Somerville, 2013; Veenstra & Laninga-Wijnen, 2023).