Abstract
Research-practice partnerships (RPPs) rely on trust, yet prior work rarely specifies how trust becomes observable in routine interactions. This single-case study explores how trust manifests and is facilitated in a teacher-researcher co-design partnership within a five-year NSF-funded bioinformatics project. Participants included three high school teachers and three university researchers. Using a qualitative approach, we analyzed recordings from a three‑day co-design workshop to examine trust enactment; implementation observation notes and debrief transcripts to trace trust-facilitating practices; and six semi-structured interviews to capture partners’ interpretations across contexts. Guided by a cyclical, multidimensional trust framework, we found that co-design trust was enacted through authority sharing, commitment to shared goals, and transparent, constructive dialogue. Across the broader partnership, trust was sustained through routines, norms, and strategies spanning five trust dimensions, including classroom visits, nonjudgmental communication, and tailored support. Findings offer actionable indicators for designing trust-supporting infrastructures in co-design RPPs.
Keywords
Introduction and Background
Collaborative design (co-design), a form of research–practice partnership (RPP), involves teachers and researchers in joint decision-making to develop and refine educational innovations (Penuel et al., 2007). Sustained partnerships can strengthen teachers’ agency and pedagogical practice while supporting the development of research-based, practice-relevant learning experiences (Penuel et al., 2007, 2011). Central to the success and longevity of these partnerships is the establishment and sustainment of trust between partners (Brown & Allen, 2021; Denner et al., 2019; Henrick et al., 2017).
An expanding body of research has examined trust in RPPs, particularly in relation to power asymmetries, differing priorities, and organizational cultures (Coburn et al., 2013; Denner et al., 2019; Henrick et al., 2017). In response, scholars have identified practice-level mechanisms for cultivating and sustaining trust, including communication and relational norms (e.g., clear dialogue and shared understandings of roles and goals), joint-work infrastructures (e.g., routines, shared language, and decision processes that support predictability and shared ownership), and equity- and power-attending strategies that foreground historical, cultural, and community contexts (Brown & Allen, 2021; Denner et al., 2019; Henrick et al., 2017; Lezotte et al., 2022; Ryoo et al., 2015; Vetter et al., 2022). This body of literature together reflects an emerging view of trust in RPPs as a multifaceted, dynamic, and relational construct shaped through ongoing interaction, communication, and negotiation of joint work.
Despite this progress, trust in RPPs is still often treated as a diffuse condition of effective partnerships, leaving its multidimensional nature largely implicit. While prior studies identify trust-supportive practices, there are fewer empirical accounts showing how trust becomes observable in routine interaction, how partners interpret specific actions as trust-facilitating or trust-eroding across activity settings, or how these interactional indicators connect to infrastructures that can be intentionally designed, monitored, and revised over time (Denner et al., 2019; Henrick et al., 2017; Lezotte et al., 2022). Without clearer conceptual and empirical accounts of how trust is enacted and sustained, it remains difficult to evaluate whether partnership infrastructures support trust in practice or to design activities that sustain collaborative decision-making over time.
Addressing these gaps, we define trust in co-design RPPs as partners’ mutual willingness to accept vulnerability, grounded in expectations of one another’s trustworthiness and cultivated through repeated interaction over time (Bryk & Schneider, 2002; Costa et al., 2018; Edwards-Groves et al., 2016). Because trust is socially constructed through interaction, it becomes visible in interdependent collaborative activities where partners coordinate contributions, negotiate decisions, and take professional risks (Dietz & Den Hartog, 2006; Lewicki et al., 2006). Accordingly, this study examines trust enactment during a later-stage co-design workshop as a concentrated site for observing interactional indicators of trust (RQ1), and traces across earlier partnership activities to identify the routines, norms, and strategies that participants enacted and experienced as stabilizing trust over time (RQ2). In doing so, we move beyond abstract accounts of trust to document how trust becomes observable and designable through interactional mechanisms across partnership activities (Henrick et al., 2017). Specifically, we ask:
RQ1. In what ways was relational trust enacted and experienced during a co-design workshop, as evidenced by collaborative, risk-taking, and monitoring behaviors and partners’ interpretations of these interactions?
RQ2. What routines, norms, and strategies did teacher- and researcher-partners enact and experience across the broader partnership that facilitated the development and sustainment of multidimensional relational trust?
To address these questions, this paper draws on a sustained teacher-researcher partnership within a five-year NSF-funded project to develop and refine a high-school bioinformatics curriculum (Yoon et al., 2022). Our goal here is not to evaluate partnership effectiveness but to identify mechanisms through which trust is enacted and sustained over time in a co-design RPP.
Conceptual Framework: Relational Trust in RPPs
Trust is broadly defined as a willingness to accept vulnerability based on positive expectations of another’s trustworthiness in their behaviors, intentions, and characteristics (Rousseau et al., 1998). In RPPs, where researchers and practitioners depend on one another to pursue shared goals (Penuel et al., 2007), this willingness is not merely an individual trait but a relational state that is developed, reinforced, or weakened through repeated interactions and shared experiences (Brown & Allen, 2021; Denner et al., 2019; Henrick et al., 2017). Following Edwards‑Groves et al. (2016), we use relational trust to emphasize this interactional, socially constructed nature; throughout, we use trust and relational trust interchangeably.
In our RPP, we conceptualize relational trust as having two interconnected components (Figure 1): a cyclical, socially constructed process through which trust is formed, enacted, and negotiated over time, and a multidimensional set of relational qualities reflected in partners’ actions and interpretations. The first component specifies how trust develops and is revised through interaction; the second specifies what relational qualities are at stake in those interactions. Figure 1 depicts the temporal process through which trust is formed, enacted, interpreted, and recalibrated over time, whereas Table 1 summarizes the five cross-cutting dimensions through which that process becomes visible in practice.
Conceptualization of Relational Trust in Our RPP
Original and Adapted Relational Trust Dimensions (Edwards‑Groves et al., 2016, 2021)
Note. Adapted definitions emphasize mutual enactment and experience across teacher and researcher roles; dimensions may co-occur within a single interaction or decision.
Relational Trust as A Cyclical, Socially Constructed Process
Grounded in the integrative model of trust (Mayer et al., 1995; Schoorman et al., 2007) and social exchange theory (Blau, 1964), we conceptualize relational trust in RPPs as an iterative, socially constructed cycle that unfolds through partners’ interactions. As illustrated in Figure 1, this cycle moves from partners’ “initial and evolving perceptions of one another and their relational and interactional dynamics” to “trust as belief”—expectations of trustworthiness, “trust as decision”—willingness to accept vulnerability, and “trust as action”—vulnerability expressed through observable behaviors, followed by “relational interpretation and response” that confirms or updates subsequent perceptions over time.
The cycle begins with partners’ initial perceptions of one another and of the partnership relationship. Partners often enter RPPs with preexisting assumptions about others’ intentions, competence, and reliability, shaped by prior interactions, professional roles, status, and shared histories (Mayer et al., 1995; Schoorman et al., 2007). In our case, prior mentor-student relationships between the principal investigator and two teacher-partners likely informed early perceptions of expertise, supportiveness, and credibility, which in turn shaped initial expectations of trustworthiness and willingness to enter the partnership. These perceptions shape “trust as belief”—partners’ expectations of one another’s trustworthiness—and “trust as decision”—judgments about whether to accept vulnerability in the relationship and in joint work (Dietz & Den Hartog, 2006; Mayer et al., 1995). Those judgments are then enacted as “trust as action,” expressed through observable behaviors such as joining the partnership, delegating authority, asking for help, or taking initiative in shared decision-making (Dietz & Den Hartog, 2006).
During these interactions, partners interpret one another’s actions in light of prior history, existing expectations, and the evolving relationship—what we term “relational interpretation and response” (Brown, 2009; Kroeger, 2012; Möllering, 2005). For instance, when researchers adjust a co-design agenda based on teacher feedback, teachers may interpret this as respect and willingness to share control; when teachers implement and iterate on proposed changes, researchers may interpret this as commitment and reliability. Partners then respond in ways that may be reciprocal, consistent, and aligned with expectations of trustworthiness, or not.
Then, these actions, interpretations, and responses feed back into partners’ evolving perceptions, confirming or revising trust beliefs after each interaction (Dietz & Den Hartog, 2006; Mayer et al., 1995). This cyclical and socially constructed view acknowledges the importance of initial perceptions in trust building yet emphasizes that the quality of ongoing interactions and relationships ultimately shapes and reshapes relational trust (Bryk & Schneider, 2002; Lewicki et al., 2006). Over time, relational trust strengthens when cycles of enactment, interpretation, and response are reciprocal and consistent, and it weakens when these patterns become inconsistent or nonreciprocal (Edwards‑Groves et al., 2016; Henrick et al., 2017; Lezotte et al., 2022). This highlights mutuality as a central organizing principle: Trust is reinforced when partnership goals, responsibilities, and benefits are co-constructed and reciprocally negotiated over time (Lezotte et al., 2022). In our conceptualization, mutuality refers to the reciprocal alignment of partners’ actions and responses over time, such that contributions and risks are recognized and reciprocated rather than one-sided.
Relational Trust as a Multidimensional Construct
While the trust cycle in Figure 1 specifies how relational trust is formed, enacted, and revised over time, we also attend to what relational qualities are at stake as partners decide to accept vulnerability and interpret one another’s actions. To describe these qualities, we adapt Edwards‑Groves and colleagues’ (2016, 2021) five dimensions of relational trust: pragmatic, interactional, intersubjective, interpersonal, and intellectual. Although Edwards‑Groves et al. developed these dimensions to characterize the trustworthiness of middle leaders in schools—instructional leaders who occupy roles between classroom teachers and senior administrators and support professional learning and change initiatives—we reinterpret them for RPPs as a shared set of relational qualities that can be mutually enacted and experienced by teacher- and researcher-partners.
Table 1 summarizes the original dimensions alongside our adapted definitions for RPP contexts. We retain the labels to preserve continuity with the original framework, while refining definitions to reflect mutuality and reciprocal expectations between partners. For example, we define the interpersonal dimension in terms of mutual care, personal connection, emotional support, and adaptability to one another’s needs and constraints, and the pragmatic dimension in terms of practical, goal-oriented actions that align partners’ needs with the aims of the partnership.
Importantly, at any point in the trust cycle (Figure 1), a single interaction or decision may express multiple dimensions simultaneously. For instance, when a teacher invites a researcher back into the classroom after a lesson that did not go as planned, that invitation may reflect interpersonal trust (expecting support), intellectual trust (valuing feedback), and pragmatic trust (expecting the visit to help address an instructional need).
Translating the Conceptual Framework into Analytic Indicators and Mechanisms of Relational Trust
For this study, we translated our conceptualization of relational trust (Figure 1) into an analytic approach in two complementary ways: by identifying trust-supporting mechanisms, routines, norms, and strategies, that participants described and, where available, enacted as recurring across the partnership (RQ2), and by specifying behavioral indicators of trust within a concentrated co-design activity (RQ1). Routines are repeated, structured practices that organize collaboration (e.g., regular post-lesson debriefs during implementation); norms are shared expectations for how partners communicate, collaborate, and treat and relate to one another (e.g., nonjudgmental uptake of teacher struggles); and strategies are intentional moves partners use to support one another, solve problems, or sustain relationship quality (e.g., communicating through personal cell messaging) (Henrick et al., 2017; Lezotte et al., 2022; Ryoo et al., 2015).
We treat partners’ enactment of these mechanisms and their interpretations and responses to that enactment as observable indicators of relational trust. In terms of Figure 1, these indicators are most visible in the “trust as action” and “relational interpretation and response” phases, where partners take interpersonal and professional risks, make sense of and respond to these actions throughout their interactions, updating expectations of trustworthiness over time (Brown, 2009; Hadar et al., 2024; Lewicki et al., 2006). Because practices can enact multiple relational qualities at once, indicators often reflect more than one trust dimension. For example, a standing debrief routine may simultaneously express interactional trust (creating space for candid talk), intellectual trust (treating teachers’ judgments as consequential), and pragmatic trust (jointly addressing classroom challenges).
This framing guided our primary analytic focus for RQ2: If routines, norms, and strategies stabilize trust cycles, they should be discernible in data capturing partnership episodes and partners’ accounts of those episodes, such as implementation support, reflection, and co-design. Accordingly, we examined how teacher- and researcher-partners enacted and experienced the five trust dimensions across sustained RPP activities in order to identify the concrete routines, norms, and strategies that participants associated with developing and sustaining relational trust.
The same conceptualization also informed RQ1. If relational trust has been developed and sustained through earlier cycles, it should be observable in later, highly interdependent joint work that requires partners to coordinate contributions and share authority. In our case, the three-day curriculum co-design workshop at the end of year three offered such a late-stage snapshot. Co-design requires partners to negotiate goals, distribute control, and collaborate toward jointly owned artifacts (Penuel et al., 2007), making it a sensitive context for observing how trust is enacted and interpreted in real time (Costa et al., 2018; Henrick et al., 2017).
To examine trust enactment in co-design (RQ1), we operationalized the “trust as action” and “relational interpretation and response” phases of our trust cycle (Figure 1) using three interactional indicator categories—collaborative, risk-taking, and monitoring behaviors—that cut across the five trust dimensions (Table 3). Prior work links higher trust to collaborative engagement such as information sharing, accepting influence, proactive contributions, and commitment to shared goals, and links lower trust to defensiveness and withholding (e.g., Parker et al., 2006; Simons & Peterson, 2000; Zand, 1972). Trust also supports interpersonal and professional risk-taking, such as seeking help and exchanging feedback (e.g., Edmondson, 2003; Mayer et al., 1995). Because monitoring can also signal distrust, we treated monitoring as trust-informed only when oversight was calibrated to keep work aligned with shared goals while preserving partners’ autonomy; this distinction is especially consequential in co-design RPPs where researchers often hold greater structural authority (Coburn et al., 2008; Ferrin et al., 2007; McAllister, 1995). Consistent with Figure 1, we interpreted these behavioral indicators alongside partners’ situated interpretations and responses (e.g., describing actions as equitable, nonjudgmental, controlling, or confidence-building) as evidence that trust expectations were confirmed or recalibrated.
Methods
Research Design
We present a single case study examining the manifestation and facilitation of relational trust in a sustained RPP (Yin, 2018). We employed a sequential qual → QUAL approach (Morse, 2010), first exploring how relational trust emerged during a co-design workshop by analyzing participants’ perceptions and behaviors. This analysis (RQ1 - qual) offered exploratory and illustrative empirical snapshots of how trust was enacted and experienced at a later stage of the partnership. We then identified the norms, routines and strategies that contributed to developing and maintaining this trust throughout the broader RPP activities (RQ2 - QUAL). Figure 2 summarizes our research design.
Summary of Research Design
The co-design workshop included three teacher-partners with one to three years of involvement in the RPP and provided a cross-sectional view of the interactional and relational dynamics between teacher- and researcher-partners. This case helped us identify routines, norms, and strategies that were repeatedly experienced or expected across RPP activities and that facilitated relational trust across stages and boundaries (Henrick et al., 2017). It also allowed us to observe how that trust manifested in a highly collaborative context where reliance on others to achieve shared goals was essential (Costa et al., 2018).
Co-Design Research-Practice Partnership
This study is part of an NSF-funded project developing a 20-lesson, problem-based, STEM-integrated bioinformatics curriculum for high school students that engages them in analyzing local air quality data related to asthma and air pollution (Shim & Yoon, 2024; Yoon et al., 2022). Initially designed by university researchers, the curriculum was implemented over three years with sustained professional development (PD), including classroom visits, just-in-time assistance, and reflective discussions to support teachers’ enactment.
We used targeted recruitment to engage teachers positioned to provide in-depth feedback for refining the curriculum (IES/NSF, 2013). Recruitment focused on environmental science and biology teachers recognized for instructional effectiveness or recommended by school science directors. By Cohort-3, recruitment expanded to lower-resourced schools to test the curriculum across diverse contexts, including one of our teacher-partners who taught at a school serving a low socioeconomic population.
Over three years, implementation and PD evolved in response to COVID-19. In Year 1, five Cohort-1 teachers participated in in-person PD and implemented the curriculum; in Years 2 and 3, PD moved online. Four Cohort-1 teachers supported this transition by revising the PD, with two serving as teacher-facilitators for Cohort-2. In Year 3, ten teachers implemented the curriculum (one from Cohort-1, three from Cohort-2, and six from Cohort-3), with online PD facilitated by two Cohort-1 teachers. Table 2 summarizes the three-year timeline of project activities, cohorts, and partner roles.
Timeline of RPP Activities, Teacher Cohorts, and Partner Roles
Note. AY = academic year; PD = professional development; co-design = curriculum co-design workshop. Teacher codes: P = participated in PD; I = implemented the curriculum unit; RF = facilitated research activities during implementation; DC = PD design collaborator; TF = teacher facilitator in PD; CD = co-designer. Researcher codes: D = curriculum/PD designer; F = PD facilitator; S = implementation support (e.g., school visits, just-in-time assistance); CD (P) = co-design design; CD (D/F) = co-design design and facilitation. An en-dash (–) indicates that the partner was not involved in that phase. Parenthetical counts (e.g., n = 3) indicate the number of teacher-partners in the cohort holding the role.
Co-Design Workshop
Over three years, teacher and student feedback indicated a need for substantial curriculum revisions, particularly to reduce the unit’s length. To address this, a co-design team of one teacher from each cohort and one researcher participated in a three-day workshop to redesign the unit. Alan (Cohort-1), Sarah (Cohort-2), and Rick (Cohort-3) were selected to represent different school contexts and levels of implementation experience: Alan had implemented the curriculum for three years at a selective magnet school; Sarah had implemented it for one year at a charter school outside the district; and Rick had implemented it for one year at a low socioeconomic school within the district.
Because they taught in different schools and districts, the teacher-partners did not know one another before joining the RPP. Their interactions before the workshop were largely mediated through PD structures: Alan interacted with Sarah and later Rick through his teacher-facilitator role in online PD, whereas Sarah and Rick had limited direct interaction before the co-design workshop (Table 2).
The co-design workshop consisted of three in-person sessions held at the research team’s university, each lasting 3.5 to 6.5 hours. Designed by the three researcher-partners (Kaitlin, Madison, and Sooah), the sessions were coordinated and facilitated by Madison. As the sessions unfolded, teachers’ input prompted adjustments to the format and activities, such as replacing planned video reflections with open discussion. Teachers and Madison collaboratively reframed the unit by defining its scope, establishing lesson objectives, and developing materials for the first lesson. Teachers then independently designed lessons aligned with their expertise and interests, which were shared for collective feedback. The workshop resulted in five redesigned lessons; the remaining seven lessons were adapted by the researcher and refined through virtual feedback from teachers.
Research Participants
The participants in this study were three teacher-partners who participated in the co-design workshop and three researcher-partners (Kaitlin, Madison, and Sooah) who supported PD and curriculum implementation. The teachers taught biology or environmental science in Northeastern U.S. high schools and had 17 (Alan), 7 (Sarah), and 21 (Rick) years of teaching experience at the start of the workshop. All three held master’s degrees in education, two from the research team’s university. Their cohort histories, implementation roles, and prior involvement in the RPP are summarized in Table 2.
The researcher-partners were Kaitlin, a learning sciences professor and principal investigator of the project, and her two doctoral advisees, Madison and Sooah, who together developed the curriculum and professional development and supported teachers throughout the RPP activities.
Researcher Positionality
The three researcher-partners served dual roles as both study participants and members of the author team. As a former K–12 teacher, Kaitlin’s shared professional identity with teachers (Tanis & Postmes, 2005) and expertise in teacher PD likely supported early trust by signaling familiarity with classroom constraints and pedagogical competence. She also had preexisting relationships with Alan and Sarah as former master’s students. These prior mentor-student interactions likely provided firsthand evidence of trustworthiness (e.g., responsiveness, nonjudgmental feedback), which shaped their initial perceptions of Kaitlin at partnership entry. In Figure 1 terms, these perceptions informed “trust as belief” and “trust as decision.” Reciprocal familiarity also mattered: Kaitlin viewed Alan and Sarah as capable and reliable, enabling her to delegate meaningful responsibilities. Such relational histories can seed initial trust (Coppola et al., 2004) that is then confirmed or updated through ongoing partnership interactions (“trust as action” and “relational interpretation and response” in Figure 1). Thus, by the co-design workshop, this initial trust was likely calibrated through their sustained interactions. Kaitlin’s dual role as mentor and researcher nonetheless required reflexivity regarding potential interpretive bias.
Beyond the PI’s positionality, the broader research team’s relational context also shaped the partnership. By the time of the co-design workshop, Kaitlin, Madison, and Sooah had collaborated for five to six years. Sooah and Madison—Kaitlin’s doctoral advisees—had worked with her and with each other across multiple projects, developing shared routines for planning, facilitation, and decision-making that likely supported consistent, predictable partner-facing work (e.g., coordinated support and follow-through).
Within this team structure, Madison and Sooah played key intermediary roles in researcher-teacher relationships. They co-designed and facilitated the curriculum and PD and worked closely with teacher-partners. Madison primarily supported Rick, and Sooah supported Alan and Sarah through classroom visits, implementation assistance, and ongoing communication, creating repeated opportunities to confirm or recalibrate trust within each researcher-teacher duo. Madison’s K–12 teaching background may also have supported early perceived trustworthiness, as shared professional identity often shapes positive perceptions in RPPs (Tanis & Postmes, 2005), particularly for Rick, who did not have prior ties to the team. Differences in the form, frequency, and context of these interactions likely contributed to variation in how trust developed.
Because the three researcher-partners held dual roles as RPP facilitators and study participants, the first author—a learning sciences postdoctoral researcher who joined the team after the co-design workshop—led the research design, interviews, analysis, and manuscript preparation. With no prior relationship with the teacher-partners, her outsider position supported candid reflection during semi-structured interviews. She also emphasized that the interviews were not intended to evaluate teachers or researchers, but to identify trust-building norms, routines, and strategies that could strengthen RPPs; responses would be shared with the research team without identifying teachers by name.
Kaitlin, Madison, and Sooah are co-authors who contributed contextual clarification, reflexive analytic debriefs, and supported theoretical framing, interpretation, and revision. Although they generated observational and debrief data in their RPP roles, they did not conduct interviews or participate in initial coding or primary analysis. Accordingly, the author team comprises the first author and the three researcher-partners; throughout the manuscript, “we” refers to this author team.
Data Sources and Collection
We drew on four data sources to address our research questions: (a) audio recordings from the co-design workshop (partial), (b) semi-structured interviews with three teacher-partners and three researcher-partners, (c) researchers’ classroom-implementation observation notes, and (d) audio-recorded post-lesson debriefs.
For RQ1, we analyzed the audio recordings of the first day and part of the second day of the co-design workshop and the interview segments focusing on the workshop to document trust-relevant behaviors and participants’ interpretations of those interactions. Teachers requested the remaining co-design sessions not be recorded to maintain a relaxed atmosphere; therefore, interviews served as the primary data source for later workshop dynamics, with recordings used to contextualize and triangulate participants’ reflections. For RQ2, we analyzed the interview segments focusing on partnership experiences across the broader RPP, along with implementation-era observation notes and debriefs, to identify routines, norms, and strategies that participants enacted and experienced as trust-facilitating.
Interviews (n=6; ~45 minutes each) were conducted via Zoom four months after the workshop, using a single protocol that included prompts about both the co-design workshop and the broader partnership. Informing both RQs, the prompts were organized around four themes: (1) motivations and expectations for joining and continuing the partnership, (2) factors influencing participation in the co-design workshop, (3) reflections on the co-design planning, interactions and activities, and (4) relationships and interactions between partners across the broader partnership (see Supplemental Appendix A).
The interviews were critical for capturing participants’ expectations and situated interpretations of partnership dynamics that were not directly observable in recorded interactions (“relational interpretation and response” in Figure 1; Bryk & Schneider, 2002). These interpretations shape whether trust beliefs are confirmed or revised over time, facilitating how trust is enacted and experienced in an RPP (Dietz & Den Hartog, 2006). Thus, we used interviews as a primary data source to examine how trust manifests in co-design interactions, and the routines, norms, and strategies partners associated with sustaining trust across the broader RPP.
For RQ2, we supplemented these interviews with observation notes and audio-recorded post-lesson debriefs collected during the three teachers’ curriculum implementations by each teacher’s assigned researcher-facilitator. Observation notes followed a structured protocol documenting student and teacher engagement, enactment of PD, implementation challenges, curriculum adaptations, and student learning. Debriefs (~5-15 minutes; audio-recorded) were brief semi-structured conversations after lessons that focused on teachers’ rationales for instructional decisions, plans for subsequent lessons, challenges encountered, and support needs. They were usually researcher-led but occasionally co-led by teachers during collaborative problem-solving. Informal social talk often accompanied these debriefs, contributing to variation in content and length across teacher-researcher pairs.
The number of observations and debriefs varied across teachers due to differences in implementation years, facilitator classroom-visit frequency during implementation, and geographic constraints. We collected 16 observation notes from Alan (three-year implementer), 15 from Rick (one-year implementer), and none from Sarah (one-year implementer) due to her remote teaching location. Debrief counts were 15 for Alan, 8 for Rick, and 2 for Sarah. All audio recordings (co-design sessions, interviews, debriefs) were transcribed using a professional AI service and thoroughly reviewed by the first author for accuracy.
Data Coding and Analysis
We employed qualitative content analysis (QCA) to systematically categorize our data and identify patterns, themes, and meanings relevant to our RQs (Schreier, 2012). Because QCA attends to both manifest content (what partners said or did) and latent content (the meanings and interpretations partners attached to those actions), it allowed us to examine both observable/described trust-relevant behaviors and partners’ context-specific interpretations through which trust was constructed and negotiated (“trust as action” and ‘”relational interpretation and response” in Figure 1) (Elo & Kyngäs, 2018).
Across both RQs, the unit of analysis was a meaning unit (a clause, sentence, short turn of talk, or discrete observation-note entry) in which a partner enacted, described, or evaluated a trust-relevant action, interaction, expectation, or relational experience (Schreier, 2012). In the co-design workshop and post-lesson debrief transcripts, meaning units consisted of a turn of talk or a short exchange of adjacent turns expressing a trust-relevant action or interpretation; in interviews, meaning units were clauses/sentences in which participants recounted, interpreted, or evaluated co-design and broader RPP interactions and relational dynamics; and in classroom observation notes, meaning units were discrete entries (e.g., a sentence or bullet) documenting an observed trust-relevant action or interaction and/or the observer’s evaluative interpretation (e.g., evidence of teacher competence, uptake of feedback). Meaning units could receive multiple codes when they included multiple trust-relevant actions or interpretations.
Our analysis was abductive: We began with an a priori coding frame for each RQ and iteratively refined it through cycles of deductive and inductive coding as the data extended, nuanced, or challenged initial categories or behaviors (Elo & Kyngäs, 2018; Schreier, 2012).
RQ1
We first applied a deductive coding frame from trust-in-collaboration literature that organized trust-related behaviors into three functional categories—collaborative, risk-taking, and monitoring—operationalized through observable/described actions and interpretive reflections (Table 3).
Coding Frame for Trust-Informed Behaviors in Co-Design (RQ1)
Note. Unilateral investment was treated as trust-informed when described/experienced as voluntary commitment; it can also signal imbalance if framed as compensatory or resentful. Accepting influence from others may function as collaborative (integrating ideas) and/or risk-taking (ceding control/face) coded based on the local function in context. Monitoring behaviors can signal distrust when excessive; they were coded as trust-informed only when participants framed them as supportive coordination that preserved autonomy.
Using this frame, we segmented the workshop transcript and interview data into meaning units and deductively coded each unit to a primary category (collaborative, risk-taking, or monitoring) and behavior(s) as a first pass, flagging ambiguous units for refinement (Elo & Kyngäs, 2018). Two independent researchers double coded the data and resolved discrepancies through debriefing sessions (Patton, 2015).
Because the deductive frame did not fully capture relational and interpretive dynamics, we followed up with inductive coding, systematically re-examining meaning units to identify emergent processes, stances, and meanings beyond the initial frame (Schreier, 2012; Saldaña, 2016). This inductive process surfaced new process-oriented (e.g., disagreeing with others) and value- and emotion-related codes (e.g., frustration with the process). We iteratively moved between the deductive framework and emergent codes, using constant comparison to examine how behaviors functioned across contexts (Elo & Kyngäs, 2018; Graneheim et al., 2017; Schreier, 2012). This process revealed instances where the same behavior functioned differently depending on context. For example, accepting influence from others included both integrating input toward a shared goal (collaborative) and allowing a partner to lead (risk-taking). We then compared and clustered related codes into higher-order themes that draw on the collaborative, risk-taking, and monitoring behaviors in Table 3. We treated these behaviors as trust-informed when they enacted voluntary professional/interpersonal vulnerability and/or reliance on others’ expertise and good intent in interdependent work, and when partners’ interpretations and responses treated these actions as evidence that trust expectations were being affirmed or recalibrated. Next, we revisited relevant literature and held analytic debriefs with Madison (workshop facilitator) to check contextual accuracy and theoretical coherence. We also solicited feedback from teacher-partners on our interpretations of data (member checking) to ensure they are aligned with their understandings and perspectives (Graneheim et al., 2017; Lincoln & Guba, 1985).
RQ2
Using meaning units (defined above), we applied Edwards-Groves et al.’s (2016, 2021) five dimensions of relational trust—interpersonal, intellectual, pragmatic, intersubjective, and interactional—as an initial deductive coding frame (Table 1). We deductively coded interview, debrief, and observation-note meaning units in which partners described, enacted, or evaluated trust-relevant actions and interpretations (“trust as action” and “relational interpretation and response” in Figure 1) for the trust dimension(s) they reflected (Schreier, 2012).
To move from dimensions to trust-supporting mechanisms, we conducted inductive open coding to identify specific behaviors, and the meanings participants attached to them (Bieber & Viehoff, 2023). Open-coding generated both process-related codes (e.g., responding quickly, communicating issues) and emotion- or value-related codes (e.g., value of responsiveness, nonjudgmental environment). We then used axial coding to compare and cluster codes into broader behavioral categories, attending to how similar codes could function differently depending on the relational quality at stake (Elo & Kyngäs, 2018; Graneheim et al., 2017; Saldaña, 2016). For example, responding quickly and timely communication constituted reciprocal reliability when reflecting partners’ dependability and responsiveness in communication, and mutual personal regard when reflecting care and attentiveness to one another’s well-being beyond professional responsibilities. Throughout, we iteratively checked emerging categories against relevant theoretical and empirical work to support conceptual coherence (Braun & Clarke, 2006).
Given variation in observational coverage across participants and activity settings, we treated behaviors as recurring mechanisms when participants explicitly characterized them as typical/expected across partnership episodes and/or when they appeared in multiple segments and, where available, across data sources. We also memoed contextual conditions that participants described as stabilizing or destabilizing trust (e.g., researchers’ prior teaching experience), attending to how trust was also shaped by roles and histories (Denner et al., 2019; Farrell et al., 2019; Penuel & Gallagher, 2017).
We further coded behaviors for directionality—that is, whether they were expected of or enacted by teachers, researchers, or both (e.g., mutual personal regard as mutually expected; genuine care as unilaterally expected of researcher-partners). To enhance credibility and validity, the first author and an external researcher independently coded the data and reconciled discrepancies through analytic debriefing; we also debriefed with the research team to check interpretive accuracy and contextual fit (Lincoln & Guba, 1985; Patton, 2015; Yin, 2018). After consensus, we produced a matrix mapping behaviors to trust dimensions and directionality, alongside associated routines/norms/strategies and contextual conditions, which is summarized in Table 4.
Relational Trust-Supporting Mechanism Categories by Trust Dimension
Note. Directionality indicates whether the mechanism category was typically expected of/enacted by researcher-partners (R), teacher-partners (T), or both (Mutual) in this dataset; mechanism categories may be instantiated through routines, norms, and/or strategies.
Finally, we solicited teacher-partner feedback on our interpretations of the findings (member checking) to confirm resonance and refine our analytic claims (Lincoln & Guba, 1985). Consistent with RQ2’s purpose, our analysis was design-oriented rather than evaluative: We synthesized participants’ accounts into trust-supporting routines, norms, and strategies that we or other co-design RPPs can adapt, generating actionable and transferable design considerations grounded in partners’ expectations and motivations (Bieber & Viehoff, 2023; Kunnel & Quandt, 2016; Li, 2007).
Results
RQ1. Trust-Informed Behaviors in Co-Design Workshop
We identified three themes in how relational trust was enacted and interpreted in the co-design workshop: shared decision authority through moderated facilitation, commitment to shared goals through diverse participation, and transparent and constructive interaction. Because the co-design workshop was only partially recorded, interviews serve as the primary evidence for partners’ accounts of trust enactment and their retrospective interpretations (“trust as action” and “relational interpretation and response”; Figure 1). We use the available recordings as illustrative interactional vignettes; Supplemental Appendix B provides extended excerpts and analytic annotations.
Shared Decision Authority Through Moderated Facilitation
Participants’ interactions and reflections suggest that decisions about what the team would design (e.g., unit scope and priorities) and how the team would work (e.g., agenda, activity formats, pacing) were responsive to teacher input rather than researcher-controlled, an especially consequential dynamic in co-design RPPs where researchers often hold greater structural authority (Coburn et al., 2008).
Teacher-partners described shaping what the workshop prioritized and, in turn, what the redesign focused on. For example, Rick recalled proposing a major structural reframe, “I just said, there is so much in here . . . and it should be divided up into two different programs: the bio component and environmental component . . . ” Sarah similarly argued for adapting the unit “specifically for environmental science . . .” In participants’ accounts, these inputs shaped subsequent scoping discussion, positioning teachers’ classroom expertise as central to product decisions; Madison treated these proposals as consequential, accepting teacher influence and relying on teachers’ judgments (Table 3).
Process decisions were likewise negotiated rather than imposed. In the recorded episode, Madison opened the question of how to proceed with the design process, and Alan proposed turning the project’s five guiding ideas (culturally relevant learning; local inquiry; real‑world problem solving; data literacy; 21st‑century skills) into lesson objectives to inform sequencing and pacing (see Supplemental Appendix B for the extended sequence):
Before we start . . . design[ing] these lessons, what’s the next step? we want to map out . . . the objectives for each lesson . . .
Do we want to turn this [project’s five big ideas] into objectives? Because that’s also going to give us an idea [of] how many lessons it should be . . .
Yeah . . . maybe we start breaking that down . . .
This opening exchange illustrates authority-granting and reliance on others (Table 3): Madison invites teacher input on structuring the process and treats Alan’s suggestion as consequential. Madison’s immediate uptake reflects accepting influence, a trust‑informed behavior that supports shared decision authority while her facilitative framing (“map out objectives”) maintains moderated control oriented toward the shared deliverable.
Madison further described leaving key process parameters open while keeping the group oriented toward an outcome: “I don’t have an agenda here other than that we need a curriculum when we leave . . . Let’s build this together . . . they took ownership of the process.” This framing signaled authority-granting and reliance on others by delegating discretion over how the work would proceed while treating teachers’ judgments as central. At the same time, Madison described actively coordinating time and focus, “I was the person who kept us on tasks . . . ‘Okay, we’re getting off-topic like we need to come back,’” which we coded as moderated control because it maintained alignment with the shared deliverable without reclaiming substantive decision authority. She reinforced that stance by positioning teachers as the instructional content experts: “I am not a teacher . . . I value all of your expertise . . . tell me how to fix [issues in the existing unit].” Together, these moves indicate calibrated facilitation that protects shared goals while preserving teacher autonomy—monitoring enacted in a way participants experienced as trust-supportive rather than controlling (Table 3).
Teachers experienced these patterns as leveling. Alan emphasized the equalizing effect of Madison’s approach: “I really got a deeper appreciation of Madison and how she runs stuff! I felt like there were four people in the room that were equals just kind of [went] through our ideas.” Alan’s characterization here functions as a “relational interpretation and response” that suggests that these authority-sharing moves were experienced as trust-affirming. These insights together suggest shared decision authority enacted through teachers’ consequential input on both product and process decisions, and through the researcher’s willingness to incorporate that input while using calibrated coordination to keep the work aligned with shared goals.
Commitment to Shared Goals through Diverse Modes of Participation
The second theme captured partners’ shared commitment to producing a workable redesign under uncertainty, reflected in complementary participation modes: putting forward and refining ideas in group discussion and taking responsibility for drafting lesson materials during distributed design work (Table 3). Focusing on the former pattern, Alan described this collaborative sensemaking as the group “not know[ing] exactly where we were going,” yet moving forward by proposing ideas and extending them—“one person says, ‘Oh, that’s a good idea. Let’s add to it.’” His account points to trust enacted through interdependence: partners stayed oriented to the shared goal while keeping ideas open to revision and allowing others’ input to shape next steps.
This pattern was visible as the team developed a day‑one “hook” for the final project. Madison’s prompt elicited successive contributions and refinements (see Supplemental Appendix B for the extended sequence):
. . . if you’re introducing this project on day one, what is the hook? . . . We have to make it personal.
. . . survey . . . to get information that I can say, “Here’s your class of students . . . that . . . has a breathing condition . . . in their immediate and extended family . . . then brainstorm ways that we’re going to fix this.”
You might have to follow up more on the air pollution . . . they might not know what PM is, but I agree. That’s a good bridge . . . What is PM2.5?
Right. And how do we get that reading? Here’s your sensor.
Here, partners openly shared concrete design ideas and resources and treated suggestions as revisable by refining and extending one another’s proposals (e.g., Alan’s PM2.5 bridge and Rick’s immediate uptake). Alan raised a potential feasibility issue without defensiveness, and it was taken up and elaborated in the flow of talk. These moves reflect trust-informed behaviors because partners made their reasoning visible and revisable and allowed others’ expertise to shape the evolving design in service of the shared deliverable.
Commitment was also expressed through the second participation mode—distributed lesson drafting after the workshop discussions. In this shortened episode, Sarah contributed less verbally than Alan and Rick, but her commitment was evident in subsequent distributed design work: She drafted two lessons while Alan and Rick each drafted one, representing voluntary additional effort toward the shared deliverable (Table 3). In interviews, Sarah attributed her early quietness to peer dynamics and self-positioning relative to the other teachers rather than to distrust of the researcher; we return to how participants made sense of these interactional dynamics in the next theme.
Transparent and Constructive Interaction
Whereas Theme 2 emphasized commitment to the shared deliverable through what partners contributed and how (discussion vs. distributed drafting), this theme focuses on interaction quality: how partners voiced critique, uncertainty, and alternative ideas and how those contributions were taken up in ways that kept the work constructive. Across accounts and recorded talk, partners engaged in open communication: They shared ideas and concerns, exchanged constructive feedback, and took up one another’s input in ways that moved the design forward (Table 3).
Sarah’s interview suggests that this openness emerged over time as she became more comfortable in the group. She initially held back, explaining: I was a little reserved because I thought, what is my experience compared to these people? . . . who’ve taught longer . . . was a little intimidating. So, day one was maybe a little awkward . . . by the second day, we started to get along better . . . maybe it just has to do with spending time together and hearing each other’s ideas.
Her account suggests her early hesitation reflected peer dynamics and perceived differences in teaching experience more than distrust of the researcher, and that increased familiarity supported more open exchange. Because the final workshop days were not recorded, we rely on participants’ reflections to characterize this shift.
In contrast, Alan and Rick challenged and refined ideas from the outset. Madison recalled an episode where Alan directly reshaped the group’s organizing approach: I had started writing . . . on sticky notes . . . Alan got up and . . . started moving them around, adding sticky notes of his own as we were talking . . . “No, no, I think they should go in this order” . . . which is exactly what we wanted to have happen.
This episode reflects trust-informed behavior because Alan treated the process as open to revision, offering feedback through action and taking initiative to improve the shared work, while Madison interpreted and accepted this intervention as productive influence rather than a challenge to her approach (Table 3).
The same openness to critique surfaced when the group debated how to support students’ scientific action in the culminating project (see Supplemental Appendix B for the extended sequence). When Madison proposed having students “do a little bit of research, [and] put together a list of things to volunteer,” Alan suggested using quantification to connect action to evidence, “what would be the effect of CO2? . . . If you planted a tree? or 20 trees?”—and Rick refined the idea toward more immediate, observable actions: “trees take lifetimes to grow . . . raise garden beds on the sidewalk and see how that changes the air quality.” This exchange illustrates open communication and constructive feedback: Rick openly challenged and improved Alan’s proposal by grounding it in feasibility and instructional payoff rather than withholding concerns. Madison similarly framed such critique as productive influence rather than threatening: Although she had written the original curriculum, she reported feeling no “resentment” about teachers’ criticisms, and noted: “I felt so lucky to have them and their expertise . . . they were so thoughtful . . .” These exchanges suggest trust-informed behaviors because partners spoke candidly and critiqued ideas with the expectation that others would respect their expertise and good intent, treating disagreement as a shared effort to strengthen the design in service of the partnership’s goals (Table 3).
RQ2. Relational Trust-Supporting Mechanisms across Five Trust Dimensions
Building on the co-design snapshot in RQ1, RQ2 traced earlier partnership episodes and post-co-design interviews to identify recurring trust-supporting mechanisms (routines, norms, strategies) that were described and, where possible, enacted, and experienced as typical across the broader RPP. These mechanisms were organized using the five relational trust dimensions as a presentation frame, while recognizing that a single routine, norm, or strategy can support multiple dimensions simultaneously. We also coded each mechanism’s typical directionality, whether it was typically enacted and/or expected of researcher-partners, teacher-partners, or both, to capture where trust work was reciprocal versus role-specific. Table 4 summarizes the resulting mechanism categories by dimension and directionality with illustrative excerpts.
Interpersonal Dimension
Interpersonal trust-supporting mechanisms reflected relational actions and interpretations that conveyed care, personal connection, emotional support, and adaptability to one another’s needs and constraints. Across partnership episodes and partners’ reflections, three interpersonal mechanism categories were salient: demonstrating genuine care, mutual personal regard, and rapport building.
Demonstrating Genuine Care
This category of mechanism was typically expected of and enacted by researcher-partners. We defined it as researcher-partners’ receptiveness and responsiveness to teacher-partners’ personal and professional struggles, needs, and constraints, in ways that treat teachers’ well-being as integral to the partnership. It was reflected in norms of safe disclosure and nonjudgmental uptake of teachers’ struggles and constraints and enacted through strategies such as validation and accommodation, affective check-ins, and encouraging trial-error during implementation. For example, Rick described Madison’s responsiveness when he felt overwhelmed: “She [Madison] understood when I said . . . ‘I need to take a couple of days off from doing this’ . . . [She said] no problem.” Here, Madison validates Rick’s struggle and adjusts expectations without questioning.
These strategies often surfaced in post-lesson debriefs. In one debrief, Madison checked in on Rick’s experience and highlighted progress across repeated attempts:
The past couple of times you’ve been feeling lost and frustrated.
I’ve been screwing up and not doing it correctly because it’s taking so long.
Do you still feel lost and frustrated?
. . . Last week, I was like, oh my god, this is going to be a nightmare . . . I feel . . . more confident seeing the students’ progress . . .
The first time you were figuring it out . . . the second time, it was natural, and you made bigger connections.
When we’re trying new things in our class . . . the first time does not come out right at all . . .
In this episode, the routine of post-lesson debriefing provided a recurring space for affective check-in and validation, while framing implementation as improving through trial and error (“the first time . . . the second time . . .”). In his later interview, Rick described this stance as confidence-building, noting it reduced his anxiety: “I was nervous . . . but getting that feedback when things didn’t go well . . . that’s why you guys are awesome.” Rick’s reflection suggests that the debrief routine, paired with nonjudgmental norms and strategies of validation/accommodation and trial-and-error encouragement, was interpreted as evidence of researcher-partners’ genuine care for teachers’ well-being, affirming confidence in researchers’ trustworthiness (Figure 1).
Mutual Personal Regard
We defined this mechanism category as partners’ reciprocal, voluntary accommodations, often logistical (time, availability, effort), that went beyond formal role expectations in ways that signal personal valuation beyond instrumental contribution. In participants’ accounts, it was enacted through strategies like discretionary schedule flexibility and rapid informal coordination (e.g., exchanging numbers/texting), and reflected norms that informal channels were acceptable for coordination/support and that partners would mutually accommodate shifting priorities. For example, Alan emphasized Sooah’s flexibility, noting, “She has a schedule . . . But at the same time, she always kind of drops everything . . . she was willing to come in when she couldn’t come . . . help out.” He linked this flexibility to informal communication: “. . . we shared numbers, and we text each other like, hey? Are you coming in today for this?” Alan also characterized this as a broader pattern of researchers making themselves available through informal channels when needed: “A lot of teachers are really appreciative . . . [researcher] group have gone out of their way to say, ‘Hey, if you need something, let me contact you informally . . . ’”
Teachers framed these accommodations as reciprocal, responding to researchers’ flexibility by adjusting their own time and effort as partnership needs shifted. Alan described aiming to “be reciprocal in the relationship . . . It’s about reading what’s important to you and me.” Taken together, these accounts suggest that mutual accommodation and informal communication were interpreted as evidence of partners’ valuation of one another and of sustained relational balance.
Rapport Building
We defined rapport building as partners’ efforts to build personal connection by sharing personal stories, learning about one another’s lives, and engaging in informal interaction. In participants’ accounts, rapport building reflected norms that personal talk could appropriately accompany partnership work and was enacted through strategies such as personal check-ins and reciprocal self-disclosure. Rick, for example, described Madison regularly asking about his family (“How’s your wife? How are the kids?”) and following up after a family event (“I had to take my son to urgent care . . . Two weeks later, she asked about him”). Kaitlin similarly described sharing about her own life and work to invite reciprocity (“I always share what I am working on . . . understanding that they will share back”).
Rapport building was also supported by mechanisms that created recurring opportunities for informal conversation, including the strategy of assigning a consistent researcher‑facilitator and the routine of classroom visits and post‑lesson debriefs, which often included casual talk alongside implementation support. Madison described how informal post‑class interaction (e.g., helping reset classroom chairs) created space to “talk to catch up,” which she noted “really strengthens those relationships.” Because rapport building was often embedded in these facilitator‑presence routines, participants described how facilitator turnover and geographic distance could disrupt or reduce them. For example, when Madison was no longer able to continue as Rick’s facilitator, he emphasized the loss of her regular classroom presence: “Madison was perfect. I’m gonna miss her . . . she was in my room every day.” Similarly, Sarah’s out‑of‑district location meant fewer in‑person touchpoints during implementation, and she noted that she did not develop rapport with Madison face‑to‑face until much later: “I really didn’t get to know her very well until the co-design year this summer . . . I’d never met her in person.” These accounts suggest that rapport-building mechanisms (e.g., personal check-ins, reciprocal sharing), when supported by facilitator continuity and regular visit/debrief routines, could strengthen personal connection between partners.
Pragmatic Dimension
Pragmatic trust-supporting mechanisms reflected partners’ practical, context-responsive, and goal-oriented actions that participants described as helping align partnership work with teachers’ classroom needs and researchers’ goals. Two pragmatic mechanism categories were salient: targeted activities and professional support and investment in design, PD, and research efforts.
Targeted Activities and Professional Support
We defined targeted activities and professional support as researcher-partners’ tailoring of RPP activities and implementation support to teachers’ needs, goals, and classroom constraints. In participants’ accounts, it reflected a norm that teacher feedback about feasibility would be taken up as consequential and was enacted through strategies such as active listening with follow-through, tailoring PD to teachers’ interests and desired takeaways, and side-by-side problem-solving during implementation. For example, Rick emphasized Kaitlin’s responsiveness to teacher input: “Kaitlin’s really good . . . listens to us and doesn’t just pacify us if it can’t be done . . ., ” and Kaitlin described designing PD so teachers would “walk away with a set of experiences they’re going to love.” Teachers also described just-in-time support during implementation, such as working “side by side to figure out how I could get my students to work with the sensors that weren’t working” (Rick). These findings suggest that teachers interpreted tailored PD and just-in-time support as making participation feasible and instructionally relevant in their local contexts, reinforcing pragmatic expectations that partnership work would remain aligned with classroom realities as needs emerged.
On the other hand, two contextual conditions were noted as shaping how targeted support was enacted across researcher-teacher pairs: differences in facilitator expertise (e.g., pedagogical coaching vs. technical troubleshooting) and recurring sensor malfunctions that made troubleshooting a frequent focus of classroom support.
Teachers’ Investment in Design, PD, and Research Efforts
This pragmatic mechanism category captured teachers’ active contributions to the partnership’s design, professional learning, and research work beyond curriculum implementation, through strategies such as providing substantive feedback on curriculum/PD materials, supporting PD facilitation/coaching for newer implementers, and helping coordinate research tasks (e.g., data collection logistics and contextual interpretation). Across interviews, researchers described the depth of teachers’ feedback as a key indicator of this investment. Sooah emphasized, “We asked teachers to provide feedback. Some teachers would just give a line or two, but some really put a lot of thought into it, giving in-depth feedback . . . an indicator of their commitment.”
Teachers’ investment in the research side of the partnership was also visible within the routine of post-lesson debriefs, which sometimes expanded from implementation reflection to coordinating data-collection logistics. In the excerpt below, Alan offers options and follow-through to support Sooah’s data collection:
Tomorrow is planning the investigation . . . if you want to peek in on some of the breakout rooms tomorrow . . . you’re still gonna get some of that on Friday . . .
I’ll join Friday . . . spend . . . time with [a group] and . . . rotate through.
. . . rotation-four . . . is a little ahead . . . you haven’t spent much time observing.
. . . rotation-three is good enough . . . I’ll reach out to rotation-one . . . for the focus group interview.
. . . I need to follow up on . . . individual surveys . . . I’ll send you that list . . .
This exchange illustrates teacher investment through coordination support and contextual input: Alan proposes feasible observation opportunities aligned with the instructional timeline, shares classroom-grounded information that could inform observation choices, and takes responsibility for follow-up tasks enabling data collection. While Sooah maintains her planned approach, Alan’s moves align to support that plan (e.g., focus group access and survey follow-up), illustrating teacher-partners’ commitment to making research activity workable within classroom constraints.
Intersubjective Dimension
Intersubjective trust-supporting mechanisms reflected partners’ mutual efforts to build shared purpose and a sense of community within RPP. One intersubjective mechanism category was salient: community building and mutual identification.
Community Building and Mutual Identification
We defined community building and mutual identification as partners’ intentional efforts to create a collective identity and sense of belonging by embracing shared values, goals, and experiences and cultivating inclusive spaces where partners felt connected and integral to the partnership’s work. One strategy the researchers consistently noted was positioning teachers and researchers as co-inquirers. Sooah noted, “As a whole team, we put a lot of effort into not just considering our teacher participants as mere research subjects or examples,” and described a norm that frames the work as collective exploration: “. . . this is exploratory . . . process we are all in together . . . we are, as a team, trying to figure out the best way to teach this to our high school students.” Relatedly, Sooah and Kaitlin described using intentional inclusive language as a strategy to reinforce shared ownership: “. . . using ‘We,’ as all of us working together.”
However, teacher accounts suggested that experiences of mutual identification were not uniform across contexts. Sarah, the only teacher from outside the [anonymized city] school district, described feeling “a little bit . . . disconnected,” and she linked this to the unit’s focus on [anonymized city]-specific cases that differed from her teaching context (“We have . . . different air pollutants . . .”). She proposed broadening cases to include diverse contexts— “. . . using people who are different as a contrast can be helpful . . .”—to better connect partnership work to variation in teachers’ settings. Across these accounts, shared positioning and inclusive “we” language were described as intersubjective trust-supporting mechanisms, while Sarah’s reflections underscored how inclusive RPP design can shape teachers’ sense of belonging within the partnership.
Intellectual Dimension
Intellectual trust-supporting mechanisms reflected partners’ mutual recognition and demonstration of professional expertise. We identified two intellectual mechanism categories: recognition and respect for expertise and demonstrating professional knowledge and competence.
Recognition and Respect for Expertise
This category captures how partners made one another’s expertise visible and consequential in joint work through how they interacted and through how instructional authority and discretion were distributed. Across accounts, recognition was treated as something to be shown rather than assumed, expressed through both affirming language and concrete role-positioning moves. Madison described enacting respect for teachers’ practitioner knowledge through a strategy of respectful time negotiation and reciprocity: “The way that I interact with them, the way that I ask for their time, and the way that I give them my time is all influenced by my respect for their knowledge.” Kaitlin similarly emphasized making recognition visible: “He knows that I think he’s terrific, and I show it through my actions.”
Partners described strategies that instantiated this norm in both directions. Researchers recognized teacher expertise by granting discretion, signaling confidence in teachers’ professional judgment during implementation. Alan described experiencing this as freedom to adapt instruction without seeking permission: I might have just modified entirely what that plan was . . . I didn’t really ask for permission. I felt I had a lot of freedom . . . it was definitely implied . . . “We think you’re great. You can keep doing what you want . . .”
Teachers, in turn, recognized researcher expertise through instructional authority sharing during classroom implementation visits, treating researchers’ in-the-moment contributions as legitimate. Rick described Madison as “almost like a second teacher,” indicating that he granted her pedagogical standing to contribute in class: “. . . She noticed I was struggling . . . she didn’t mind helping out . . . It didn’t make me feel judged.” Rick’s framing of Madison as “almost like a second teacher” shows him granting her pedagogical standing to contribute in his classroom. Here, the key mechanism is not the help itself, but Rick’s role-positioning move that makes her contribution legitimate.
Demonstrating Professional Knowledge and Competence
This mechanism category reflects partners’ demonstrations of pedagogical and practical expertise that made competence observable in the partnership’s work. These were most visible within the routine context of researcher class visits and observations during implementation, where partners could see expertise enacted. Partners described two key strategies. First, researchers demonstrated pedagogical competence through in-the-moment instructional support during implementation; Rick recalled that when he was “struggling,” Madison stepped in to “help out.” Second, teachers demonstrated instructional competence through effective enactment that researchers could point to concretely in observation notes; for example, Sooah wrote, “All teachers should model how Alan instructs how to use sensors and prepare for data collection,” and Madison noted, “Rick is doing a great job linking science to current events.”
Participants also referenced Madison’s prior K–12 teaching as an early credibility cue (Madison: “Being able to say I was a teacher . . . immediately bumps up your respect level”), and her just‑in‑time classroom support during implementation offered a concrete instance, reinforcing that perceived pedagogical competence.
Interactional Dimension
Interactional trust-supporting mechanisms reflected how partners sustained predictable, candid, and dependable communication and collaboration across RPP activities. We identified three interactional mechanism categories: regular communication and collaboration, open and safe communication, and reciprocal reliability.
Regular Communication and Collaboration
This mechanism category reflects partners’ recurring coordination touchpoints for reflecting on implementation and aligning next steps. It was most visible in the routine of classroom visits followed by brief post-lesson debriefs, which created a predictable opportunity to revisit how a lesson went and what to adjust. Alan described this pattern as consistent: “Each time Sooah came, she would say, ‘Can you talk about how that just went?’ and then I would share my thoughts, and we would discuss it.” In this account, the regularity of these conversations and the use of prompts to open reflection functioned as a predictable forum for shared sensemaking about the work.
Open and Safe Communication
We defined open and safe communication as partners’ ability to engage in candid, constructive exchanges, raising concerns, offering suggestions, and giving/receiving feedback with the expectation that input would be listened to and taken up rather than dismissed. It was reflected in a norm of constructive responsiveness to feedback and enacted through reciprocal strategies of active listening with follow-through (researchers) and uptake of feedback in practice (teachers). Rick, for example, highlighted that “You listened to my suggestions . . . you listen to us, and you’re open to suggestions.” This openness was reciprocated through teacher uptake of feedback in subsequent instruction; Madison noted in observation notes, “Rick took to heart what we talked about,” documenting how prior discussion carried forward into practice. Together, these mechanisms supported communication quality by keeping suggestions discussable and improvable within ongoing work.
Reciprocal Reliability
We defined reciprocal reliability as partners’ mutual dependability and timely follow‑through on communication and coordination commitments across RPP activities. It was reflected in the norm of timely responsiveness and enacted through strategies such as replying promptly, communicating issues as they arose, and completing time‑sensitive coordination tasks. Sooah emphasized teachers’ timely responsiveness as essential for collective preparation: “Teachers’ commitment was critical . . . whether they are timely and responsive to us . . . Because we’re working as a whole group, and there are a lot of time‑pressing preparations.” She also described parallel expectations for researchers’ responsiveness: “I would reply to teachers probably in three hours, or at least on the same day but not later than two days.” Sarah similarly recalled rapid follow‑through when implementation problems emerged: “Anytime I couldn’t find something, or there was an issue with the devices, I’d email or text her, and she’d get back to me really quickly.” Together, these accounts describe reciprocal reliability as a repeated pattern of timely response and follow‑through that supported ongoing coordination across partnership activities.
Discussion
Although trust is widely recognized as essential to sustaining effective RPPs, it is often treated as a diffuse condition, with limited empirical attention to how trust is enacted in routine interaction or stabilized through partnership infrastructures over time (Denner et al., 2019; Henrick et al., 2017). Using a cyclical, multidimensional relational trust framework, we analyze a late-stage teacher-researcher co-design workshop to identify interactional indicators of trust enactment (RQ1) and the infrastructures partners experienced as stabilizing trust across broader partnership activity (RQ2).
Trust-Informed Behaviors in Co-Design
We identified three interconnected themes of trust-informed behavior across the co-design data: shared decision authority through moderated facilitation, commitment to shared goals through diverse participation, and transparent, constructive interaction.
First, shared decision authority through moderated facilitation illustrates how trust becomes visible under conditions of structural asymmetry. RPP scholarship has long noted that researchers’ institutional status can render practitioner input consultative rather than consequential (Coburn et al., 2008; Denner et al., 2019; Farrell et al., 2019). In this case, teachers experienced the workshop as equitable because facilitation enacted a calibrated balance: Teachers’ input shaped both product and process decisions, while the researcher-facilitator coordinated time and focus in ways participants interpreted as autonomy-preserving rather than controlling. This distinction matters because monitoring can easily be read as distrust when experienced as excessive oversight (McAllister, 1995). Here, monitoring functioned as supportive coordination, aligning with work showing that authority sharing and monitoring can reinforce cooperation when experienced as goal-aligned and competence-affirming (Ferrin et al., 2007; McAllister, 1995). In trust-cycle terms, authority-granting moves constituted vulnerability on the part of the facilitator, while teachers’ interpretation of those moves as “leveling” reinforced positive judgments of researchers’ trustworthiness (Costa & Anderson, 2011; Dietz & Den Hartog, 2006).
The themes of commitment to shared goals and transparent, constructive interaction further demonstrate how trust became observable in action. Across episodes, teacher and researcher partners advanced ideas, allowed them to be shaped by others, voiced critique with expectations of good-faith uptake, and sustained momentum toward shared deliverables. These patterns align with research linking trust to constructive disagreement and sustained task-focused engagement (Breuer et al., 2016; Costa & Anderson, 2011; Simons & Peterson, 2000). Importantly, the uptake of teacher critiques shows that teacher expertise was consequential in the moment, not merely acknowledged in principle, a critical distinction in co-design RPPs where power asymmetries can otherwise blunt practitioner influence (Coburn et al., 2008; Denner et al., 2019; Farrell et al., 2019).
However, these trust-informed interactional norms were not taken up uniformly across teachers. In the recorded segments, Alan and Rick were more verbally active, whereas Sarah was less visible in group talk early on. Sarah’s retrospective account framed this pattern in relation to peer-facing positioning, especially experience/seniority comparisons and limited prior ties within the teacher group. Consistent with multilevel views of trust in teams, these layered “being different” cues—i.e., seniority/prestige, cohort/geographic belonging, and potentially gendered participation norms given that she was the only woman teacher-partner—can shape within-group relational dynamics and interactional patterns (Costa et al., 2018; Jassawalla & Sashittal, 1999; Tsui et al., 1992). This foregrounds teacher-teacher trust dynamics as consequential in co-design partnerships, extending RPP accounts that more often emphasize teacher-researcher role and power asymmetries (Coburn et al., 2008; Kyza & Agesilaou, 2022).
Moreover, Sarah described becoming more comfortable through spending time together and engaging others’ ideas, suggesting that relational trust expectations can recalibrate as interactional history accumulates (Bryk & Schneider, 2002; Lewicki et al., 2006). Her substantive contributions through less public modes (e.g., taking on a larger share of lesson drafting than other teachers) further reinforce that co-design participation is realized across multiple moments and forms of contribution (Bratteteig & Wagner, 2016). Read together, Sarah’s participation patterns complicate any simple mapping of talk time onto trust enactment. Consistent with our relational trust-cycle framing, this underscores that trust is evidenced not only in participation patterns but in how partners interpret those patterns and respond in ways that confirm or recalibrate trust expectations over time (Dietz & Den Hartog, 2006; Mayer et al., 1995). Yet, further research is needed to examine how intersecting positionalities (e.g., gender, professional seniority) within role groups shape these participation trajectories and relational interpretations and uptakes that reinforce or weaken trust in RPP interactions.
Trust-Supporting Mechanisms in RPP
RQ2 extends the co-design snapshot by identifying infrastructures that stabilized relational trust across activity settings. Building on RPP scholarship on joint-work infrastructures and predictable, revisitable routines (Henrick et al., 2017; Ryoo et al., 2015), we add empirical specificity by showing how routines, norms, and strategies stabilize trust cycles by linking trust-enacting actions to relational interpretation and response, rendering trust less diffuse and more designable as a set of interactional affordances. Taken together, the mechanism categories summarized in Table 4 point to three broader takeaways: Trust was stabilized through recurring work-cycle routines, through patterned combinations of nonjudgmental responsiveness and visible follow-through, and through practices that made expertise and shared ownership consequential in joint work.
A central finding is that the most trust-supportive mechanisms were not isolated relational moves but recurring work-cycle routines that created low-risk opportunities for vulnerability, sensemaking, and follow-through. The classroom-visit and post-lesson debrief routine illustrates this dynamic. While prior RPP work emphasizes the value of regular check-ins (Henrick et al., 2017), our findings specify what makes such routines trust-supportive: They repeatedly transformed implementation uncertainty into shared problem solving while making responsiveness, competence, and care observable. In trust-cycle terms, these micro-interactions invited trust as action (e.g., surfacing struggles, offering critique) and generated relational interpretation and response through nonjudgmental uptake, timely support, and visible follow-through (Bryk & Schneider, 2002; Lewicki et al., 2006). This interactional history helps explain why monitoring and coordination during the co-design workshop were experienced as autonomy-preserving rather than controlling, as oversight-like actions were consistently paired with assistance and a nonjudgmental stance, shaping how they were interpreted (Costa & Anderson, 2011; Ferrin et al., 2007).
Limited researcher presence in Sarah’s classes highlights an important boundary condition: when routines and touchpoints are constrained, opportunities for rapport and “withness” may be delayed or reduced, shaping intersubjective trust and contributing to felt disconnection, an equity-relevant concern often noted abstractly in RPP scholarship but less frequently tied to specific infrastructures (Denner et al., 2019; Vetter et al., 2022). More broadly, this illustrates that trust-supportive mechanisms do not map neatly onto single trust domains because partners interpret actions holistically within unfolding activity settings (Edwards-Groves et al., 2016, 2021).
A second cluster of mechanisms centered on how nonjudgmental responsiveness, consequential follow-through, and relationship building operated together as patterned trust work. Vulnerability in RPPs requires teachers to surface constraints and researchers to accept influence or revise plans, actions that carry professional risk under role and status asymmetries (Coburn et al., 2008; Denner et al., 2019; Farrell et al., 2019). Across interactions, nonjudgmental uptake paired with visible follow-through, and supported by rapport, flexibility, and availability, lowered the perceived cost of candor and shaped how feedback, oversight, and disagreement were interpreted as constructive engagement rather than evaluation or dismissal (Bryk & Schneider, 2002; Edmondson, 2003; McAllister, 1995). Treating teachers’ judgments as decision-relevant expertise, rather than acknowledged but disregarded, was critical for sustaining trust, consistent with prior RPP findings (Denner et al., 2019; Farrell et al., 2019).
Relatedly, partners made expertise visible and consequential through recognition, respect, and authority-sharing moves rather than affirmation alone. While competence-based trust is a core basis for reliance under uncertainty (Mayer et al., 1995; Rousseau et al., 1998), in RPPs it is inseparable from role and authority: Researchers and teachers hold different forms of expertise, and trust depends on how these are recognized in practice (Coburn et al., 2008; Farrell et al., 2019). Granting discretion, negotiating time, and legitimizing in-the-moment classroom contributions shaped whether instructional involvement and adaptation were interpreted as collaboration or intrusion, thereby keeping joint-work workable rather than rigidly controlled (Ferrin et al., 2007; Edwards-Groves & Grootenboer, 2021).
The final mechanism theme concerns how shared ownership was enacted through both discourse and labor, including community-building practices (e.g., co-inquirer positioning, inclusive “we” language) and teachers’ substantive investment in design, professional learning, and research work. Together, these practices fostered a sense of collective purpose and made reciprocal commitment visible in ways that sustained partnership work in context (Edwards-Groves et al., 2016, 2021; Henrick et al., 2017; Lezotte et al., 2022). At the same time, Sarah’s account highlights a boundary condition: As the only out-of-district teacher, she experienced disconnection when curriculum work centered city-specific cases misaligned with her context, underscoring that belonging depends not only on inclusive discourse but on whether partnership artifacts substantively recognize contextual variation. This aligns with equity-oriented RPP scholarship emphasizing that trust and inclusion are shaped by power, culture, and whose experiences are centered in partnership activity (Denner et al., 2019; Vakil et al., 2016; Vetter et al., 2022).
Across themes, trust was reciprocal but occasionally role differentiated. While many mechanisms were mutual, others were non-mutual in analytically distinct ways: Some reflected complementary, role-specific contributions essential to joint work, whereas others involved asymmetrically distributed relational labor. Making this directionality visible clarifies how mutuality may be enacted through complementary roles rather than identical behaviors, extending scholarship that treats mutuality as a core partnership principle (Henrick et al., 2017; Lezotte et al., 2022). When trust-supporting labor becomes concentrated in one role, partnerships may benefit from explicit negotiation of expectations and reciprocity to support sustainability over time (Denner et al., 2019).
Implications, Limitations, and Conclusion
Our findings suggest several design and theoretical implications for monitoring and strengthening trust-supporting infrastructures in co-design RPPs. Design implications include the following recommendations: (1) schedule recurring work-cycle routines such as visit-debrief loops that support implementation sensemaking and relational check-ins rather than relying on ad hoc support; (2) maintain consistent researcher-teacher pairings over time, where feasible; (3) make decision rights and role expectations explicit by clarifying who decides what, and when, so respect for expertise is enacted through consequential influence rather than polite affirmation; (4) establish interactional working agreements that specify nonjudgmental responsiveness and follow-through, including norms for feedback uptake and response-time expectations, to support psychological safety and candor (Edmondson, 2003); (5) build explicit supports for teacher-teacher community and participation, treating within-cohort belonging as part of the trust infrastructure rather than an assumed byproduct; (6) establish an inclusive partnership identity through co-inquirer positioning and collective “we” language and design artifacts that legitimate contextual variation across teacher settings, particularly in cross-district partnerships; and (7) add periodic maintenance check-ins focused on shifting constraints and the distribution of relational labor to prevent trust work from becoming invisibly asymmetric across roles (Lezotte et al., 2022).
Theoretical implications include these recommendations: (1) extend analytic attention beyond teacher-researcher trust to include teacher-teacher trust; (2) examine trust as a cyclical, interactional process by tracing repeated collaborative moments and partners’ interpretations of those interactions across RPP activities; (3) frame trust as a multidimensional construct without treating dimensions as siloed; and (4) use directionality as an analytic lens to examine when trust work is reciprocal versus role-skewed.
These implications, however, should be interpreted in light of limitations. Our analysis draws heavily on interviews to capture partners’ interpretations across contexts and includes uneven observational coverage across participants, notably limited implementation-era observation/debrief data for Sarah due to geographic constraints. In addition, because teachers remained engaged in an ongoing partnership and researcher-partners served dual roles as co-authors, social desirability may have shaped interview accounts despite the use of an external interviewer and non-evaluative framing; findings should therefore be interpreted in light of this relational context. Additionally, because the co-design workshop was only partially recorded, our findings are limited in their capture of enacted interactional behaviors and rely in part on partners’ retrospective accounts of co-design interactions and dynamics. In some instances, however, the partial nature of the recordings also reflected responsiveness to teacher-participants’ preferences during workshops, a practice aligned with the relational trust central to the partnership. Finally, as a single case, our findings do not claim universal causal effects; instead, they offer a mechanism-based description of how trust-supporting infrastructures can be enacted and experienced in one co-design RPP, generating transferable design considerations that future work should test and elaborate in more diverse partnership contexts (Denner et al., 2019; Vetter et al., 2022).
Taken together, this study contributes a relational, cyclical, and multidimensional account of trust in co-design RPPs by identifying design-oriented mechanisms that operate across activity settings and trust dimensions and linking those mechanisms to observable trust enactments in co-design interactions. By making these mechanisms and enactments explicit, the study offers a more revisable account of trust-supporting partnership infrastructure, responding to calls for clearer empirical specification of what trust looks like in routine interaction and how it can be deliberately supported over time (Henrick et al., 2017; Lezotte et al., 2022).
Supplemental Material
sj-docx-1-ero-10.1177_23328584261448146 – Supplemental material for Unpacking Relational Trust Mechanisms in Teacher-Researcher Co-Design Partnerships: Insights from a Case Study
Supplemental material, sj-docx-1-ero-10.1177_23328584261448146 for Unpacking Relational Trust Mechanisms in Teacher-Researcher Co-Design Partnerships: Insights from a Case Study by Tugce Aldemir, Susan A. Yoon, Jooeun Shim and Katherine M. Miller in AERA Open
Footnotes
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by a grant from the U.S. National Science Foundation (DRL #1812738).
Open Practices
The data underlying this article are not publicly available because they derive from a small qualitative case study and contain contextual information that could permit deductive identification of participants or the study site. Data may be available from the corresponding author upon reasonable request.
Authors
TUGCE ALDEMIR is an assistant professor at College of Education and Human Development, Texas A&M University, College Station, Texas, USA; email:
SUSAN A. YOON is a professor at Graduate School of Education, University of Pennsylvania, Philadelphia, Pennsylvania, USA; email:
JOOEUN SHIM is an assistant professor of teaching at Teachers College, Columbia University, New York, New York, USA; email:
KATHERINE M. MILLER is a research associate at Concord Consortium, Concord, Massachusetts, USA; email:
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
