Reconfiguring Delphi: Reciprocity and Professional Reflection in Qualitative Research Design

Abstract

This article reflects on the adaptation of the Delphi method within a qualitative, social constructivist paradigm, illustrated through a study of resilience in early childhood education and care (ECEC) in Aotearoa New Zealand. Rather than using Delphi solely as a tool for generating expert consensus, the study reconfigures it as a reciprocal and dialogic process that amplifies practitioners’ situated knowledges and lived experiences. Two iterative rounds combined qualitative depth with selective quantitative tools, enabling the articulation, comparison, and refinement of diverse perspectives on resilience. Methodological decisions such as embedding participants’ contributions into subsequent rounds, incorporating flexible response formats, and sustaining iterative meaning-making fostered sustained engagement and supported reflective consideration of practice. These adaptations positioned participants as co-constructors of knowledge while generating insights relevant to professional practice. The study suggests the potential of Delphi not to impose fixed interpretations but to collaboratively generate nuanced understandings. It contributes to methodological debates by showing how Delphi can be reframed as a reciprocal process that can support professional reflection, highlighting the value of divergence, relationality, and contextual sensitivity in qualitative research design.

Introduction

The Delphi method has gained increasing traction across disciplinary domains (Beiderbeck et al., 2021; Khodyakov et al., 2023). Commonly understood as a mixed-methods approach, Delphi studies frequently lean towards the measurement of consensus among expert panels (Hsu & Sandford, 2007). This article takes a different orientation, offering a methodological reflection for scholars considering the Delphi method in qualitative research. Rather than seeking consensus as the sole priority, this study embraced the richness and contextual specificity of diverse viewpoints. Central to the design was an interest in identifying and articulating expert practitioners’ understandings of resilience emerging within distinct pedagogical, cultural, and institutional environments.

In this study, resilience was not operationalised as a fixed construct but engaged as an open site of inquiry from participants’ experiences. This orientation allowed participants to articulate different interpretations and to reflect on how their conceptions evolved across rounds. Reflection refers to how participants had the opportunity to revisit their own experiences and professional understandings of resilience through dialogue with others. It also guided my engagement as a researcher, as sustained dialogue with participants’ contributions informed and reshaped my conceptual framing of resilience. Reflexivity refers to the research-level awareness of how professional roles and researcher decisions shape interpretation and design (Braun & Clarke, 2019). In this article, reflexivity is embedded primarily at the level of research design. It informed decisions such as avoiding the imposition of a pre-established definition of resilience and allowing participants’ language and conceptual articulations to directly shape subsequent instruments. Reflexivity is further evident through my engagement with the Aotearoa New Zealand ECEC context and in the iterative adaptation of the study in response to participants’ input.

A key contribution of this article is its framing of the Delphi method as a reciprocal and dialogic research tool, one that enables researchers and participants to revisit, refine, and moderate their thinking over time. While prior qualitative Delphi studies acknowledge co-construction and multiple viewpoints, they less often examine how study designs can be intentionally oriented towards dissensus, reciprocity, and professional reflection in practitioner-led contexts. This article advances that conversation in three ways: first, by treating divergence not as a limitation but as a methodological resource that illuminates conceptual boundaries; second, by demonstrating how reciprocity can be enacted through concrete design decisions, including communication practices and embedding participants’ language into subsequent instruments; and third, by illustrating how Delphi functioned not only as a research tool but as a structured space for professional learning in ECEC settings in Aotearoa New Zealand. In this sense, the Delphi method created a space to support the emergence of professional insight. Participants were not positioned as passive informants but as active contributors whose voices shaped the design and direction of the study.

Context

The findings presented in this article are drawn from a doctoral study conducted in Aotearoa New Zealand that explored how resilience is understood and practised within ECEC. The research critically examined the concept of resilience through two stages: first, a Delphi study with practitioners and experts to gather perspectives on how resilience is interpreted within local contexts; and second, a multiple-case study in ECEC centres to observe and contrast how these interpretations are enacted in practice.

My cultural background as a Latin American scholar raised in Colombia initially led to an understanding of resilience as an individual or familial response to adverse social conditions. Undertaking research in Aotearoa New Zealand required re-examining these assumptions, particularly as local ECEC practice emphasises collective, relational, and culturally situated understandings of wellbeing. Entering this context as a visitor, I was mindful of being in a position of learning rather than authority, which included the deliberate decision to avoid imposing a singular definition of resilience. This orientation aimed to adopt an open, participant-led Delphi design in which practitioners articulated meanings of resilience from their own situated expertise, rather than responding to a predetermined conceptualisation.

Within this project, resilience is understood as a dynamic, negotiated, and socially mediated process shaped by the material and relational environments in which children learn. This view recognises that resilience is neither neutral nor universal: it is entangled with broader political and ideological discourses that influence how vulnerability, capability, and responsibility are framed within educational settings. In Aotearoa New Zealand, these tensions become particularly salient, as resilience is interpreted through cultural values embedded in Te Whāriki: He whāriki mātauranga mō ngā mokopuna o Aotearoa — New Zealand Early Childhood Curriculum (Ministry of Education, 2017) and te ao Māori, where wellbeing, belonging, and collective reciprocity offer alternatives to individualised understandings.

At the same time, the study recognises resilience as subject to critical scrutiny in Aotearoa New Zealand and other Indigenous contexts, where scholars have questioned the ways in which resilience discourses can individualise responsibility, obscure structural inequities, and normalise exposure to adversity (Penehira, 2014). These critiques highlight that resilience is neither neutral nor uncontested, and that Māori ways of knowing and being may position resilience quite differently from dominant Western framings.

While the present article concentrates on the design, a subsequent piece will examine substantive findings, including participants’ critical views on resilience in practice. It will also address their perspectives on current risks and limitations, including the tensions between tokenistic approaches and those grounded in culturally responsive practice, as well as the influence of governmental policies and regulations.

Reframing Delphi: From Consensus to Constructivism

The classical Delphi method appeared as a forecasting and decision-making procedure designed to obtain expert consensus through multiple iterative rounds of questionnaires. Originally developed for military and defence-related purposes, early iterations of the method relied on paper-based correspondence and followed a tightly structured format, with controlled feedback and anonymous participation to minimise the influence of dominant voices (Drumm et al., 2022; Linstone & Turoff, 1975).

Over time, the method has evolved considerably. It is now employed in a wide range of disciplines, from health and education to social sciences and technology, and is often adapted to incorporate qualitative input, online platforms, and more flexible designs. Its core features continue to include the anonymity of participants within an expert panel, the identification of consensus or dissensus, iterative rounds of questioning, and the provision of feedback to participants (Hasson et al., 2025).

Central to the Delphi method’s latest developments is a growing critical examination of the consensus among the panel, long considered a defining feature of the method, both in terms of how it is operationalised (for example, through response stability) and how it is interpreted by researchers (Powell, 2003; von der Gracht, 2012). These analyses have raised considerations regarding the methodological and ethical implications of encouraging convergence through iterative feedback, particularly when such processes contribute to participant attrition across multiple rounds and may discourage dissenting voices (Rowe & Wright, 1999). As a result, the legitimacy of consensus has become increasingly ambiguous, to the extent that, in some studies, it is no longer pursued as a primary objective (Diamond et al., 2014; Fletcher & Marchildon, 2014; Steinert, 2009). Alongside these concerns, there has been growing recognition of the social and contextual dimensions that shape participants’ responses, including their professional trajectories, institutional settings, and cultural frameworks (González-Sanmamed et al., 2019). Together, these insights have accentuated the adaptability of the Delphi method and its underlying philosophical flexibility, which are now widely seen as strengths that allow researchers to tailor the approach to diverse contexts and epistemological orientations.

From a social constructivist standpoint, Delphi is not merely a mechanism for extracting fixed knowledge from experts, but a dynamic process in which understanding is co-constructed across rounds. This orientation is supported by Keeney et al. (2011), who highlight the qualitative dimensions of Delphi and the importance of participant narratives. In this view, consensus reveals diverse professional realities and evolving perspectives, rather than a singular or reductive agreement. Accordingly, consensus is increasingly understood not as a definitive or objective representation of truth, but as a temporary stabilisation of viewpoints shaped through interaction and embedded in the social and contextual dynamics of a study (Gnatzy et al., 2011; Niederberger & Spranger, 2020). McIntyre-Hite (2016) similarly highlights the methodological flexibility of Delphi, aligning it with social constructivist approaches that value iterative refinement and the development of contextualised perspectives. This adaptability has proven especially valuable in studies addressing socially and ethically complex issues (Donohoe & Needham, 2009), where values, experiences, and interpretations actively inform outcomes (Jünger et al., 2017). Taken together, these contributions reaffirm the methodological legitimacy of Delphi where multiplicity of viewpoints and contextual meaning-making are not seen as limitations, but as central to the research process.

Methodological Framing: Why Delphi?

The decision to employ the Delphi method in this study of resilience was informed by both theoretical and pragmatic considerations. Positioned within a social constructivist paradigm, the study understands knowledge as contextually situated, socially mediated, and dynamically constructed through interaction and reflection (Cunliffe, 2008). Rather than identifying a single, consensus-based definition of resilience, this study aimed to gather a diversity of perspectives, tensions, and patterns of meaning from participants, allowing the iterative process to become a site of reciprocity where participants and I refined our thinking. The design also recognised my position as a newcomer to the Aotearoa New Zealand ECEC context, where imposing fixed Western categories could privilege traditional assumptions. Consequently, the study prioritised answers that developed from expert practitioner knowledge. A qualitative adaptation of the Delphi method was therefore well-suited to this purpose as it supported the generation of layered insights that evolve through iterative engagement, reflection, and refinement.

The Delphi technique was well suited to exploratory, concept-driven, and practice-oriented questions that sought to understand how experienced practitioners define resilience and identify the strategies they consider meaningful in their professional practice. Its iterative structure enabled participants to revisit their responses, considering others’ perspectives, allowing shifts in emphasis or reasoning to become visible. In this sense, the qualitative Delphi is well-suited to questions concerning the development, refinement, and contestation of expert understandings in context.

The orientation of the study was particularly important in the context of Te Whāriki, which places strong emphasis on cultural responsiveness, identity, and the agency of children, families, and communities in constructing their own learning pathways (Ministry of Education, 2017). The Delphi method enabled this research to respect and reflect that diversity, allowing expert participants to articulate views informed by their own realities. In doing so, the research did not aim to arrive at a prescriptive definition of resilience, but rather to illustrate its multiple meanings as they unfold within specific contexts.

Additionally, the Delphi method was chosen for its value as an exploratory and foundational phase within a broader, multi-stage design. Although I had conducted a literature review on resilience in early childhood, it could not capture the lived expertise of practitioners or the nuances of their daily interactions with children and the ways in which they support resilience. The literature review and Delphi study, therefore, complemented each other, combining up-to-date research insights with the iterative, contextually grounded input of experts in the specific settings of Aotearoa New Zealand.

These contributions added a responsive dimension to the analysis and supported my ongoing reflexive consideration of how instrument development and interpretation were being shaped by participants’ contextual insights. In turn, the Delphi outcomes operated as a guide for the second stage of the study, directing attention to the structural conditions that frame resilience in Aotearoa New Zealand. This encouraged the multiple case study phase to be designed with greater sensitivity to local values and relational dynamics that shape how resilience occurs in practice.

Finally, the method’s practical advantages also accounted for its selection. Although the Delphi method can be time-intensive for researchers, it is highly flexible as it does not require synchronous participant interaction. The asynchronous nature of participation allows experts to contribute at their own pace, enhancing accessibility and inclusion. This flexibility was particularly valuable in the Aotearoa New Zealand context, enabling participation from a geographically dispersed group of ECEC experts across both the North and South Islands.

Procedure

The Delphi study followed a two-round design. The procedure was as follows:

1. Recruitment: Eligible participants were qualified ECEC practitioners with a minimum of five years’ experience. Recruitment took place through professional networks, institutional invitations, and publicly available profiles.

2. Round One: Participants responded to three open-ended questions designed to provoke definitions, examples, and strategies related to resilience. Responses were collected over three weeks.

3. Data Analysis (Round One): Responses were coded inductively in NVivo to identify themes, complemented by deductive classification across individual, relational, institutional, and societal domains.

4. Round Two: Structured tasks (ranking, rating, categorisation) were introduced based on themes from Round One. Qualitative comment fields allowed participants to explain or expand their responses.

5. Data Analysis (Round Two): Median scores and interquartile ranges were calculated for ranked items. Qualitative justifications were analysed thematically to contextualise quantitative patterns.

Design and Implementation

Selection Criteria

Consistent with the understanding of knowledge as situated, relational and reciprocal, participants were required to hold professional qualifications and have a minimum of five years’ experience as ECEC teachers in the field. This condition aimed to ensure that contributors possessed substantial pedagogical and contextual understanding, shaped through daily interaction with children, families, and diverse communities. While practical experience was prioritised, the definition of expertise was intentionally broad, acknowledging the multifaceted nature of professional identity in the ECEC sector. Many participants also held roles as head teachers, managers, or researchers. These layered roles were seen not as disruptions from practice but as extensions of it, allowing participants to engage with open-ended questions, searching for critical thinking and reflective insights.

The study aimed for 40 participants who could draw on nuanced, practice-informed perspectives rather than abstract generalisations. Their situated understandings were crucial for generating contextually meaningful responses that could inform the second phase of the research and honour the diversity of Aotearoa New Zealand’s ECEC landscape. This inclusive yet experience-based definition of expertise was relevant considering the focus of the research. Resilience is not a fixed construct; it develops in complex, relational, and often unpredictable ways across settings and situations (Masten et al., 2023).

It is important to note that, as a result of these criteria and recruitment routes, the panel did not represent a standard perspective of ECEC practitioners. Rather, it comprised a purposive group of highly engaged and reflective experts, many of whom combined practitioner, leadership, and researcher identities. This composition was intentionally designed to create conditions in which conceptual thinking about resilience could be challenged and extended, rather than simply recorded as a static snapshot of everyday, habitual understandings. The Delphi phase, therefore, functioned less as a descriptive survey of sector-wide opinion and more as a structured space in which experienced practitioners could interrogate and elaborate their ideas.

Recruitment Process

Three strategies were used to recruit participants. First, initial contacts were made through professional networks and referrals from researchers in the ECEC field. These relationships enabled the identification of highly engaged practitioners and opened channels for wider dissemination. Second, formal invitations were extended to kindergarten associations and directors of ECEC learning centres, who were invited to circulate the opportunity among their teams. This approach broadened reach while retaining relevance. Third, potential participants were also identified through publicly accessible online resources such as research publications, professional profiles and government databases.

Including the invitations circulated at organisational levels, 120 profiles were identified as potential participants and were sent a personalised invitation. This included an email outlining the purpose of the study, an information sheet, and a link to register their interest. Some contacts were also approached in person at academic and sector-specific events. In total, 58 individuals registered their interest, exceeding the expected recruitment number initially set. Registrations indicating that participants’ professional experience was located mostly in primary education rather than ECEC were excluded at this stage. Following this review, 52 participants were formally invited to join the Delphi panel. These participants received the Round One survey link and provided their informed consent, following procedures approved by the institutional ethics committee.

Contextual Diversity Within the Delphi Panel

ECEC in Aotearoa New Zealand includes a diverse range of service types that reflect different philosophies, governance models, and community involvement. Kindergartens are teacher-led centres often affiliated with local associations. Playcentres are run cooperatively by parents and families, where they actively participate in their children’s education. Early learning centres, which include private and community-based services, vary widely in approach and structure, offering flexible hours and programmes tailored to family needs (Ministry of Education, 2023). The study aimed to include practitioners from a range of ECEC services, particularly kindergartens, playcentres, and early learning centres. Although Auckland represented a concentration of ECEC centres, participants from other regions and settings were also included.

The study also considered diversity in professional roles. All participants were experienced practitioners, although their current positions varied. The majority identified as current practitioners (55%), followed by centre managers or head teachers (22%), and researchers (18%). Notably, 5% identified their professional profile across the three roles. This professional heterogeneity enabled the capture of a broad spectrum of perspectives on the topic, varying by location, type of ECEC centre, and current position.

Round Structure

The Delphi process consisted of two rounds, each designed to balance openness with structure. Round One posed three open-ended questions phrased in a conversational tone to encourage genuine responses: 1) If you were to explain what resilience is to another person, how would you describe it? 2) What does resilience look like in a child? Can you share an example from your experience? 3) Based on your experience in ECEC, what are three strategies you would use to nurture resilience in children? No theoretical concepts or frameworks, aside from the term resilience, were introduced at this stage to avoid constraining participants’ interpretations. The questions were refined in consultation with research supervisors. Participants were given three weeks to respond, followed by an additional week allocated for analysis, reflection, and preparation for Round Two.

The Round Two questionnaire was informed by a thematic analysis (Braun & Clarke, 2022) of the Round One responses and a subsequent consultation process. It retained the three original areas of inquiry but introduced structured elements. Round Two invited participants to: (1) rate definitions of resilience on a scale; (2) classify scenario-based examples into conceptual categories; (3) reorder a set of strategies according to perceived importance; and (4) respond to an open-ended critical question on the risks and limitations of current resilience discourses in ECEC. Each item was complemented by an optional comment field, encouraging elaboration while respecting participants’ time constraints. Participants had three weeks to complete Round Two.

Although the possibility of Round Three was considered, it was determined that sufficient depth had been achieved within two rounds. This decision was based both on methodological reasoning and ethical considerations related to participant time and engagement. Time-tracking functionality within the survey platform enabled to monitor average completion times, ensuring alignment with estimated time commitments and demonstrating respect for participants’ contributions.

Designing to Sustain Participant Engagement

The study achieved high engagement across both rounds, an outcome that is often difficult to sustain in Delphi studies. Round One had an 82% response rate, while Round Two maintained a strong participation level at 78%. The survey platform generated individualised links that facilitated personalised follow-up and avoided redundant reminders. Participants who completed the survey early were not contacted again, while those who had not responded received a follow-up message. This tailored, relational approach to communication was intended to enhance engagement and reaffirm the study’s commitment to participant-centred research practices.

Participant engagement in this study was primarily driven by the participants’ genuine interest in the topic of resilience and their professional commitment to the ECEC sector. The pre-registration process served not only to establish eligibility but also to affirm participants’ willingness to engage meaningfully with the research under clearly defined expectations. No financial or transactional incentives were offered; instead, engagement was designed to be sustained through the perceived relevance of the topic, the clarity of the process, and the respect for participants’ time and expertise.

Time constraints are a commonly cited challenge in Delphi studies, particularly when engaging professionals in demanding roles. In response, the Round Two design intentionally combined time-efficient structures with opportunities for reflection. Each item began with a concise task followed by an optional open-ended question. This structure adapted different levels of engagement and allowed participants to contribute in ways that matched their availability, while still generating valuable qualitative insights.

The Round Two multimodal approach acknowledged participant diversity, recognising differences in professional roles as well as in cognitive preferences and ways of engaging with ideas. It further reinforced the reciprocal nature of the study by giving participants autonomy in their contributions. Because these items were optional, participants could choose whether to complete only the rating, ranking, or categorisation tasks or to invest additional time in extended written reflection, thereby exercising autonomy over the depth of their engagement.

Data Analysis

Round One

Round One responses were analysed using a combination of inductive and deductive strategies. I coded responses inductively in NVivo without predefined categories, identifying initial codes closely informed by participants’ own language. Coding began line by line, with descriptive labels created directly from the wording used in responses, and then repeatedly compared, merged, or separated through subsequent readings so that broader themes were built gradually rather than imposed in advance. These codes were then iteratively refined and organised into broader thematic clusters around definitions of resilience, concrete examples from ECEC settings, and strategies used to support children’s resilience.

Once inductive analysis had surfaced key areas of meaning, I added a deductive layer by organising responses into broader domains that reflected the level at which resilience was being interpreted in practice, such as individually, relationally, institutionally or within wider social conditions. Figure 1 visualises how inductive codes and deductive categories were integrated, illustrating the flow from participant responses to thematic clusters and then to conceptual frameworks. This sequential integration ensured that participant-led meanings were foregrounded while still enabling comparability and synthesis across established domains. In this way, inductive analysis preserved openness to participants’ voices, while deductive analysis provided a structure for dialogue with existing theory and for the design of Round Two.

Figure 1.

Integration of Participants’ Responses

Round Two

Round Two introduced quantitative and qualitative items. To assess alignment among panellists, descriptive statistics were calculated for each item, with median scores and interquartile ranges (IQRs) used to summarise central tendency and variability in participant rankings. Consensus was thus operationalised descriptively through these measures, treated not as a definitive endpoint but as one indicator of alignment or divergence across the panel. The median, being less influenced by extreme ratings than the mean, helped identify the central tendency of panel responses, while the IQR provided a clear indicator of consensus: a smaller IQR signalled a narrower clustering of ratings and stronger agreement, whereas a larger IQR indicated greater diversity (von der Gracht, 2012). Alongside these statistical measures, the targeted qualitative responses deepened the analysis by revealing the reasoning behind participants’ ratings and highlighting the nuances underlying quantitative patterns. For example, participants rated definitions of resilience on a scale and were then invited to justify their lowest-ranked choice, offering insight into areas of disagreement and added depth through reflective commentary.

Participants were also invited to categorise examples into broader domains that reflected where the examples seemed to be located (for instance, at an individual, relational, institutional, or wider societal level). Participants could also propose alternative categories if these domains did not align with their interpretations. This created space for critical engagement with the conceptual tools of the study itself. A third question asked participants to rank nine strategies according to perceived relevance, with an optional invitation to explain the selected order.

The final question asked participants to reflect on potential risks or limitations in how resilience is currently framed and promoted in ECEC. This question encouraged critical interrogation of resilience discourse and surfaced thoughts on its ethical, cultural, and ideological dimensions.

Analytically, Round Two drew on both within-participant and across-participant/round perspectives. The within-participant analysis examined each participant’s responses in depth, exploring their rationale, arguments, and conceptual framings, and situating these in relation to the panel’s aggregated patterns within the same round. The across-participant/round perspective considered how ideas evolved in response to group feedback and shifting contexts (Fletcher & Marchildon, 2014), confirming that Delphi processes benefit from examining both the individual’s reasoning and the group’s collective trajectory.

The analytical approach was consistent with the study’s epistemological stance. Rather than treating knowledge as fixed, the process recognised it as developing and situated, shaped through dialogic engagement. The Delphi method, particularly in its qualitative dimensions, provided a space for knowledge to evolve in response to others. This analytic change is evident in participants’ responses across rounds. The two examples below in Table 1 illustrate a broader trend: while Round One strategies to nurture resilience emphasised supporting children at an individual level, by Round Two, several participants had reframed resilience as relational, locating it in the quality of collective contexts.

Table 1.

Participants’ Response Evolution Across Rounds

Participant	Round one Q3: Proposed strategies	Round two Q3: Commentary to ranked strategies
20	“I would support a child to break down experiences or tasks into manageable parts so they can pace themselves”	“Because all social learning is bound in the context and relationships, and the calibre of these relationships sets the scene for nurturing healthy, resilient tamariki [children].”
35	“The most important aspect I see for our tamariki [children] is making sure that we meet their basic needs as best we can – are they hungry, tired, need time spent one-on-one to regulate themselves.”	“Relationships built on respect and trust are the essential foundation needed to allow any learning to occur.”

Taken together, these shifts can be interpreted to illustrate how the iterative Delphi process allowed participants to reconceptualise resilience from individualised strategies toward relationally grounded practices. This progression indicates Delphi’s value as a dialogic method that supports reflection and, at times, encourages awareness of how professional roles and institutional contexts inform interpretation.

Challenges in Data Analysis

Analysing a significant volume of data within a short timeframe posed a challenge, particularly between rounds. With only one week separating the two rounds, a deliberate choice intended to sustain engagement, I observed that most participants submitted their responses early in the cycle. This enabled an initial familiarisation process with the data before formal analysis began, easing the transition into Round Two. The aim was, at this time, not exhaustive interpretation but a meaningful synthesis, one that could inform and provoke reflective engagement in the design for the subsequent round.

A related challenge involved balancing thematic convergence with the preservation of divergence. This required navigating the inherent tension between honouring diverse viewpoints and shaping them into a structured format for further consideration. Rather than seeking forced consensus, the analysis sought to identify areas of overlap, resonance, or shared significance, while still respecting and preserving divergent perspectives.

Finally, the quantitative component, often perceived as a barrier for researchers with primarily qualitative training, was kept intentionally accessible. Measures such as the median and interquartile range (IQR) were used to summarise and compare participants’ rankings without imposing statistical complexity. This combination allowed me to present a clear, accessible picture of areas of alignment and disagreement, without oversimplifying the richness of the qualitative data. A quantitative and accessible approach also served as a methodological bridge, demonstrating how descriptive statistics can meaningfully support pattern recognition and panel synthesis in Delphi studies.

Ethics

Ethical approval for this study was obtained from the University of Auckland Human Participants Ethics Committee (Reference number UAHPEC28536). All participants received an information sheet outlining the purpose of the research, the voluntary nature of their involvement, and the right to withdraw their contributions up to two weeks after the completion of the final Delphi round. Informed consent was obtained electronically prior to participation.

Confidentiality was assured through the anonymisation of all responses, and participants’ identities were not disclosed in publications or reports. Data were collected through an online survey platform and stored securely on password-protected systems accessible only to me as the primary researcher.

In line with te Tiriti o Waitangi commitments and institutional ethics guidance, the study was designed with attention to equity and the potential implications for Māori participants. The research sought to uphold principles of partnership, participation, and protection by ensuring that participation was voluntary and could be withdrawn, inviting engagement from services working in multicultural settings, and emphasising relational understandings of resilience in the analysis.

Reciprocity and Reflection in the Delphi Process

Reciprocity is a foundational principle in ethical and relationally grounded research. It recognises participants not as data sources but as co-constructors of knowledge, and it asks what is returned intellectually, professionally, and relationally, through the research. In this study, reciprocity was operationalised through design decisions such as embedding participants’ Round One responses into subsequent instruments, tailoring reminders, and providing interim findings back to the panel.

The Delphi method, when designed with care, offers distinctive opportunities for reciprocity. Its iterative, feedback-oriented structure creates alternatives for participants to revisit their thinking, encounter additional perspectives, and refine or reposition ideas over time. In this study, reciprocity operated along two intertwined pathways: (1) my conceptual framing and subsequent study design were shaped by participants’ expertise; and (2) a design intended to provide participants with a structured reflective space, supporting professional inquiry into how resilience is understood and enacted in their settings.

Conceptualising Reciprocity in Practice

From the beginning, the Delphi process was framed for participants as a collaborative exploration rather than an exercise in agreement-seeking. Communications emphasised that divergence was welcome and analytically valuable. This orientation signalled respect for contextual knowledge and invited participants to speak from practice, culture, and community as well as from theory. The following extract from the first-round invitation email illustrates how reciprocity and openness were initially positioned:

We are now starting Round One, and I kindly invite you to complete the first questionnaire. In this round, you will be asked to provide your perspectives on three key open-ended questions. The questions are designed to capture a wide range of views and experiences for understanding resilience in the context of Aotearoa New Zealand.

Reciprocity was embedded at a procedural level. Participant feedback from Round One directly informed the framing of Round Two, and adaptations such as optional open text fields and flexible response lengths were implemented in response to panellist suggestions. These adjustments contributed to a more collaborative and responsive research environment and respected the varying capacities and time constraints of participants. Similarly, instruments for Round Two were directly built from Round One responses, using participants’ own words as prompts. This was designed so that participants could recognise contributions shaping the next stage, a tangible return on their input. The examples below, presented in Table 2, illustrate how individual responses informed the design of subsequent questionnaire items.

Table 2.

From Round One Responses to Round Two Questionnaire Items

Example	Round one participant’s response	Round two questionnaire design
First	Q2. Participant 24 response: “I have experienced children use creative strategies to make shelters with loose objects over their heads in a group to stay out of the rain, it is when they come together for a shared contribution to making the experience manageable for themselves.”	Q2, seventh item: “A group of children work together to build a shelter during the rain.”
Second	Q3. Participant 5: “Listening to the child, hearing and seeing what they are saying verbally and through their body language.”Q3. Participant 34: “A pedagogy of listening - this strategy involves being present with the child.”	Q3, first item: “A pedagogy of listening which involves acknowledging children’s right to express themselves, and showing an interest in their point of view and their feelings.”

By directly embedding participants’ words into the second-round instruments, reciprocity was operationalised in concrete methodological decisions. This suggests how Delphi, when designed responsively, can function as a collaborative space of professional reflection rather than a one-directional data collection tool.

Designing for Flexible Participation

Because time is a key barrier in practitioner-engaged research, the study incorporated dual-response formats: quick-entry items (ratings, rankings, categorisations) paired with optional open comment fields. This design respected participants who could contribute only briefly while still welcoming extended reflection from those with more capacity. It also allowed experts to decide where to invest depth.

Flexibility was also supported through communication practices. Individualised survey links enabled tailored reminders, and queries were answered promptly. These practices sought to balance efficiency with thoughtfulness, signalling to participants that their involvement was valued rather than administratively managed. For example, when participants had not yet completed the first-round survey, they received a reminder crafted to acknowledge their contributions and gently encourage further engagement. An excerpt is included below:

It’s been almost two weeks since we launched the first round of our Delphi panel, and it’s been encouraging to see so many thoughtful and enriching responses. Your insights into resilience and how it can be supported in ECEC are deeply appreciated and are already helping to shape a valuable collective perspective.

This is a kind reminder for those who haven’t had the chance to participate yet, there’s still time to contribute. The first round will remain open until Thursday, 24 April 2025.

You can access the survey here: [link]

Your contribution is important and will help ensure that diverse voices and experiences are reflected in this research.

This kind of communication modelled reciprocity in practice by showing appreciation for prior contributions, reassurance of ongoing value, and openness to diverse levels of engagement. Such practices supported high response rates across both rounds and helped maintain a sense of dialogue.

Participant Learning and Professional Reflection

Many participants shared additional professional implications, for example, noting how resilience is observed in children’s everyday interactions or how institutional expectations shape concepts such as coping and self-regulation. After receiving an initial summary of the Round Two findings, participant 32 commented, “Thank you for sharing these initial findings. They are very interesting, and I will be reflecting on them further in relation to my practice with children and EC student teachers.”

The Delphi method functioned as embedded professional learning. Participant 18 acknowledged this effect by noting that completing the two-round questionnaires had generated internal discussion within her centre, demonstrating that the exercise extended beyond individual responses into local professional dialogue. She also requested permission to use the questionnaire and her own responses in an upcoming team exercise, indicating her intention to engage the staff collectively in reflecting on their approaches to resilience.

Iterative Knowledge Production

The research process was characterised by iterative adaptation of survey instruments in response to participant feedback and by my responsiveness to their insights. One such moment came from Participant 33, whose reflection discussed the ideological underpinnings of resilience:

I think there is a risk if our reasons for valuing resilience go unexamined. For example, the interest in resilience as a positive attribute certainly seems to have grown under neoliberalism. Therefore, how much of the emphasis we place on it stems from economic drivers, such as the need for a work force that is ever increasing, working longer hours, in more complex and demanding environments? And, are these good enough reasons for the amount of attention it is getting?

This contribution, and others like it, challenged the usual assumptions about resilience by pointing to its relationship with political and economic ideologies. Such insights shaped the next stage of the study: I used them to guide site selection, refine the design of data collection tools, and improve interview questions. They also invited ongoing reflexive attention to how my own assumptions about resilience and adversity were being re-examined in light of participants’ perspectives in Aotearoa New Zealand.

Moving Beyond Extractive Research

An explicit objective of the study was to avoid an extractive model in which researchers collect data and withdraw. Several design decisions supported continuity: transparent communication about how data informed the next round; acknowledgement of participants’ work; and closing communications that reported back significant insights before the wider study outputs were produced. These practices aimed to create a sense of ongoing relationship, even within an asynchronous, survey-mediated format.

As part of the results communication plan, two weeks after the conclusion of the study, a six-page summary of results was produced. This document included additional insights related to the transitional stages of the study and highlighted key findings intended to support participants’ reflection. The report aimed to represent the range of participants’ perspectives, acknowledging the diversity of interpretations around resilience, as well as points of agreement and tension across the panel, thereby reflecting the iterative and deliberative nature of the Delphi technique.

Methodological Implications

Building on existing work that recognises Delphi’s constructivist and participatory potential, the adaptation presented in this study extends the literature in three methodological directions. First, it treats divergence as a productive outcome rather than a limitation, demonstrating how dissensus can be analytically generative. Second, it foregrounds reciprocity as integral to design quality, showing how participant contributions can shape subsequent rounds in tangible ways. Third, it documents how Delphi can operate as a structured space for ongoing professional reflection and learning, blurring boundaries between research and practitioner development. These implications extend beyond the immediate context of resilience in ECEC and are particularly significant for researchers considering Delphi in educational and social contexts.

Traditional Delphi designs often frame divergence as a limitation, positioning consensus as the ultimate goal. In this study, however, divergence emerged as a source of conceptual richness. Across two rounds, shifts in participants’ responses highlighted how resilience was understood not only at an individual level but also as relational and systemic. The persistence of multiple viewpoints did not weaken the study but deepened its explanatory power, revealing tensions that can inform both theory and practice. For researchers, this suggests that Delphi can be used not only to identify common ground but also to map conceptual boundaries and highlight productive disagreements.

The conceptual shifts observed across rounds had concrete implications for the wider research project. The Delphi phase generated a nuanced scenario of resilience discourses in ECEC. This scenario illustrated tensions and boundary zones between, for example, child-focused self-regulation narratives, relational narratives, and more critical perspectives on policy and neoliberal expectations. Such insights enabled the subsequent multiple case study phase to pose more targeted questions, such as: Under what conditions do individualised framings become visible, contested, or reworked in situ? These are questions that could not have been formulated with the same precision without the iterative, dialogic work undertaken with the Delphi panel.

Reciprocity was an ethical orientation and a methodological resource embedded within the design. Round One responses were directly incorporated into Round Two instruments, giving participants tangible evidence of their influence. Communication practices reinforced this reciprocity, with tailored reminders and interim summaries demonstrating respect for participants’ contributions. By operationalising reciprocity in these ways, the study illustrates how Delphi can move beyond extractive models of data collection, instead creating iterative cycles of giving back. This points to a broader methodological implication: reciprocity should be seen not as peripheral to rigour but as integral to the quality of Delphi research.

Finally, the study revealed that Delphi can serve as a site of professional learning. Participants reported that their engagement prompted reflection within their centres, influencing conversations with colleagues and shaping their thinking about resilience in practice. In this sense, the Delphi functioned as more than a tool for gathering data: it created a structured opportunity for practitioners to interrogate their assumptions, engage with diverse perspectives, and revisit their own positions. This underscores the potential of Delphi to blur boundaries between research and professional development, offering benefits that extend beyond the immediate production of knowledge.

Taken together, these implications highlight the generative potential of Delphi when approached from a constructivist perspective. Divergence, reciprocity, and professional reflection each expand the method’s capacity to contribute not only to academic knowledge but also to ethical and practice-oriented research relationships.

Limitations

This doctoral study was exploratory in scope and invited participation from a range of ECEC services across Aotearoa New Zealand. Although information about participants’ ethnic backgrounds was not collected in this phase, their contributions reflected a range of cultural influences shaped by the relational and multicultural contexts of ECEC practice in Aotearoa New Zealand, as these were negotiated and expressed through expert professional understandings of resilience. Future Delphi studies could incorporate ethnic and cultural demographic information to enable a more nuanced examination of whose cultural understandings are being recognised within ECEC professional communities.

Conclusion

This study illustrates how the Delphi method can be adapted within a qualitative, constructivist paradigm. By treating divergence as productive, operationalising reciprocity as a design practice, and recognising Delphi as a space for professional reflection, the research offers an example of how Delphi might move beyond a tool for consensus-building to become a relational and dialogic process. Although the study is situated in the context of ECEC in Aotearoa New Zealand, the methodological insights may hold relevance for researchers in education and the social sciences who wish to work with multiplicity, reflective inquiry, and collaborative engagement, and where reflexive awareness can be fostered through iterative dialogue. In this way, the study contributes not only to understanding resilience in ECEC but also to ongoing debates about the possibilities of Delphi in qualitative research.