Write a reflexivity note that explains analytical decisions

A reflexivity note should help a reader understand how the researcher became part of the research process. It is most useful when it links a position, relationship or assumption to a particular decision, then explains what that decision made easier to see, harder to see or necessary to qualify.

That is different from writing a list of demographic characteristics, declaring “I may be biased,” or describing personal growth as you might in an assessed reflective assignment. Those statements can matter, but they do not yet show how the study changed because this researcher asked these questions, met these participants, developed these codes or chose this interpretation.

This guide uses a decision ledger to make that chain visible. The worked study, people, quotations and decisions are wholly fictional. Adapt the process to your methodology and to your institution, ethics approval and reporting requirements; do not treat the headings below as a universal reflexivity template.

Start with the method you actually used

Reflexivity is not one standard paragraph that can be pasted into every qualitative project. Its purpose and language depend on what the methodology assumes about knowledge, interpretation and the researcher’s role.

In Braun and Clarke’s reflexive thematic analysis, analysis is understood as skilled, situated and interpretive work. Their guidance on doing reflexive thematic analysis presents phases as guides rather than rigid rules. Their later reporting guidance stresses methodological coherence and reflexive openness. A note written for that approach should not claim that the researcher eliminated subjectivity through neutral coding or consensus.

Other qualitative approaches may place different weight on bracketing, researcher–participant relationships, team interpretation, field position or an audit trail. Begin by naming the approach and checking its assumptions. Then ask what a reader needs to know to understand the decisions you made within it.

If you are still deciding what your analytic outputs mean, use the separate guide to distinguish a code, topic summary and theme. A reflexivity note explains how your situated choices influenced the analysis; it does not replace the analysis itself.

Distinguish four records that can overlap

Researchers may be asked for a positionality statement, reflexive note, research journal or decision log. Institutions and journals use these labels differently, so follow the definition that governs your project. A useful working distinction is:

A positionality statement describes relevant social, professional, disciplinary or experiential locations from which the researcher approaches the work.
A reflexivity note examines how those locations, assumptions, relationships and responses shaped research choices and interpretations.
A research journal is a chronological working record. It may contain reactions, questions and emerging interpretations that later support a reflexivity note.
An audit or decision trail records what changed, when, why and with what evidence. It can make reflexive claims traceable.

These records can support one another. The danger is substitution. A list of identities is not automatically reflexive. A diary entry is not automatically an analytical explanation. A version history can show that a code changed without explaining why the researcher noticed one meaning and not another.

The complete accepted manuscript by Lazard and McAvoy treats reflexivity as critical evaluation of the connections among research questions, decisions and conclusions. It also places reflexivity across topic selection, design, conduct and writing, rather than only at the end of analysis.

Build a seven-field decision ledger

Use the ledger while the study is in progress. Add an entry when a choice could materially affect what evidence is generated, how material is interpreted or how claims are presented.

On a narrow screen, scroll the table sideways to read every column.

Field	Question to answer	Weak entry	More useful entry
Stage	Where in the study did this occur?	Analysis	Early coding after interviews 1–3
Decision	What did you choose or change?	I changed the codes	I separated “quiet observation” from “reluctance to participate”
Plausible alternative	What else could you reasonably have done?	None	I could have retained one broad participation code
Situated influence or assumption	What in your position, experience, relationship or expectation may have shaped the choice?	I am biased	As a workshop facilitator, I initially treated visible speaking as the main sign of engagement
Consequence	What did the choice make more or less visible?	The analysis improved	The distinction made listening and private note-taking visible, but the data still could not establish learning
Response or limit	What did you do, and what uncertainty remained?	I stayed objective	I recoded earlier extracts and tested the distinction against a non-speaking participant’s account; observational limits remained
Evidence trail	Where can the reasoning be checked?	My notes	Journal entry 12, codebook v3 and theme map dated 18 September

The “plausible alternative” field matters because it prevents hindsight from making every decision look inevitable. The “consequence” field matters because it moves the note beyond biography. The “response or limit” field stops a corrective action from being presented as if it erased the original influence.

Worked example: make one decision chain visible

The Harbourline Online Writing Circle, its participants, quotations and records are wholly fictional. They were created for this guide and are not research data.

A fictional researcher studies how first-generation postgraduate students decide whether to contribute during optional online writing workshops. The researcher previously facilitated similar workshops and values active discussion. Early interview prompts and field notes repeatedly describe students who speak, ask questions or share text on screen.

After the third interview, one participant says:

“I rarely unmuted, but I wrote down how other people explained their paragraphs. I tried the same check on my chapter later.”

The researcher notices that the early coding treated limited speech as reluctance or disengagement. That interpretation may reflect the researcher’s facilitation experience rather than the participant’s account.

The decision should be documented as a chain rather than converted into a declaration that the researcher “had bias.”

On a narrow screen, scroll the table sideways to read every column.

Stage	Decision and alternative	Situated influence	Consequence	Response, remaining limit and trail
Question design	Added “What were you doing when you were not speaking?” rather than asking only why participants stayed silent	Facilitation experience had made audible contribution feel like the default form of participation	Opened space for listening, private annotation and later application; still relied on self-report	Used a neutral follow-up in later interviews; interview-guide v2 and pilot memo 04
Early coding	Separated “quiet observation” from “reluctance to contribute” rather than retaining one low-participation code	The researcher initially read workshop behaviour through a facilitator’s participation goals	Prevented silence from being assigned one meaning; risked creating a distinction supported by few extracts	Recoded interviews 1–3 and looked for contrary cases; journal entry 12 and codebook v3
Theme development	Reframed “barriers to speaking” as “managing the visibility of participation” rather than organising the theme only around confidence	The initial analysis privileged an improvement narrative in which speaking more was the desirable endpoint	Connected speaking, chat, listening and delayed application through a shared concern with visibility; did not establish which practice improved writing	Checked the central concept across all extracts and retained a clear evidence limit; theme map v4
Reporting	Reported the facilitator background beside the analytic decision, not as a detached identity list	Professional familiarity offered contextual sensitivity but also normalised programme assumptions	Readers can see both the interpretive resource and the risk of over-reading participation	Added a bounded reflexivity paragraph and avoided outcome claims; methods draft v5

This ledger does not prove that the revised interpretation is correct. It shows why the interpretation changed, how the researcher tested it and where the evidence still stops.

Turn the ledger into a concise note

A published note rarely needs every journal entry. Select decisions that help a reader evaluate the path from question to conclusion. A useful paragraph can follow this sequence:

name the relevant position or prior assumption;
locate the stage and decision it influenced;
explain what became visible or obscured;
describe the response or test; and
state the remaining limit.

For the fictional study, a concise note might read:

The researcher had previously facilitated online writing workshops and initially treated audible contribution as the clearest sign of participation. After an interviewee described learning through listening and later private revision, the researcher separated quiet observation from reluctance to contribute, recoded the earlier interviews and tested the distinction across the dataset. This decision supported a theme about managing the visibility of participation rather than a simple confidence deficit. The revised analysis makes less visible forms of engagement easier to examine, but interview accounts cannot establish whether those practices improved writing outcomes.

The paragraph does not claim neutrality. It identifies a situated starting point, a traceable analytical change and an unresolved evidential boundary.

Do not copy this wording into an unrelated project. Its logic depends on the fictional research question, relationship and analysis. Use your own records to recover the decisions that actually occurred.

Track reflexivity before, during and after data generation

A final note is stronger when it is built from contemporaneous records. Add brief entries at decision points such as:

narrowing the research question;
defining inclusion or recruitment routes;
writing, piloting or revising interview questions;
deciding what to record as contextual field notes;
responding to rapport, discomfort, disagreement or disclosure;
noticing an early code that reflects a disciplinary expectation;
combining, separating or rejecting candidate themes;
selecting extracts for the report;
deciding what not to claim; and
revising language after supervisory or team discussion.

For interview work, connect the ledger to the versioned process used to draft an interview guide with open and focused questions. Record why a prompt changed and what the change may enable or suppress. A polished final guide alone cannot show that reasoning.

Do not manufacture a clean retrospective trail. If a decision was not recorded at the time, say that the explanation was reconstructed from dated materials and identify the limit. A candid incomplete trail is more useful than invented certainty.

Include decisions that did not flatter the analysis

Reflexivity becomes decorative when it reports only thoughtful improvements. Look for moments where your preferred explanation failed, a participant resisted your framing, a familiar concept crowded out alternatives, or a quotation was especially persuasive because it matched your expectations.

Ask:

Which participant or extract was hardest to place?
Which code felt obvious because of my training?
Which account did I initially call irrelevant?
What did I expect participants to value?
Did rapport make one interpretation feel safer or more compelling?
Which alternative was considered and rejected?
What remains ambiguous after the response?

A contradictory case does not automatically invalidate an interpretation, and it should not be hidden to protect a neat story. Explain whether it led you to narrow, divide or reject a claim. The companion code, topic-summary and theme diagnostic shows how contradictory extracts can test a candidate analytic pattern.

Handle team analysis without inventing consensus

If several researchers contributed, record who made which interpretation and how differences were handled. Do not imply that discussion removed subjectivity or produced one objectively correct code unless that claim fits the declared methodology and procedure.

A team entry might identify:

the interpretation proposed by each analyst;
the different disciplinary or experiential standpoints that informed the disagreement;
the material each interpretation explained or overlooked;
the decision ultimately taken and why; and
the uncertainty that remains.

Within reflexive thematic analysis, Braun and Clarke’s work on reflecting on reflexive thematic analysis foregrounds researcher subjectivity rather than treating it as contamination to be eliminated. Team discussion can deepen interpretation, expose assumptions and improve coherence; it is not automatically a reliability statistic.

If you used a codebook, consensus process or another form of thematic analysis, describe that method accurately instead of borrowing reflexive-TA language.

Protect participants and collaborators

A reflexivity note may involve sensitive information about relationships, field access or the researcher’s own position. Transparency does not require disclosing identities or confidential events that the report does not need.

Before including an entry, ask whether it reveals:

a participant’s identity through a rare role or incident;
confidential institutional information;
safeguarding or health information beyond the approved reporting scope;
a colleague’s private reaction;
details that change the consented use of material; or
unredacted text from a protected research journal.

Use role-level descriptions where they are sufficient. Keep identifiable working notes in approved storage. Apply the project’s retention and access controls. Do not paste participant material, confidential field notes or protected research data into an unapproved online tool to generate a reflexivity statement.

Diagnose common weak notes

On a narrow screen, scroll the table sideways to read every column.

Weak pattern	Why it is insufficient	Repair
“I acknowledge that I am biased.”	Names no decision, mechanism or consequence	Identify the assumption, where it entered the study, what it shaped and how you responded
A list of identity categories	Positionality is presented without analytical relevance	Connect only relevant positions to concrete relationships, choices or interpretive possibilities
“Member checking ensured accuracy.”	Turns a procedure into a guarantee and may blur participant feedback with one final truth	State what feedback was requested, how it influenced interpretation and what it could not establish
“The team agreed on all themes.”	Hides interpretive differences and may import a reliability logic that does not fit the method	Describe material disagreements, the reasoning used and the selected approach
A polished paragraph with no trail	Readers cannot tell whether the account reflects actual decisions	Retain dated journal, codebook, guide or theme-map references in the working record
Reflexivity used to excuse an unsupported claim	Researcher transparency cannot create missing evidence	Narrow the claim and state the evidential limit separately

Use a twelve-step drafting process

Name the methodology or analytical approach.
Check what that approach assumes about the researcher’s role.
Gather dated journals, guide versions, codebooks, memos and theme maps.
Mark decisions that changed evidence generation, interpretation or reporting.
For each decision, record one plausible alternative.
Identify the relevant assumption, position, relationship or response.
Explain what the chosen path made more and less visible.
Record the action taken to examine or respond to that influence.
State the limitation that the action did not remove.
Point to an evidence trail without exposing confidential material.
Select the few chains readers need to evaluate the study.
Check the final note against the method and the claims actually reported.

A supervisor, co-author or qualitative-methods adviser can challenge a vague chain. Ask them to point out where a position is named without a consequence, a correction is claimed without evidence, or a method-specific assumption is missing.

Final check

Before submitting the note, confirm that it:

names the method and does not borrow incompatible quality claims;
distinguishes positionality from decision-linked reflexivity;
covers relevant stages rather than only end-stage writing;
identifies concrete choices and plausible alternatives;
explains consequences for what became visible or obscured;
records responses without pretending they restored neutrality;
includes remaining uncertainty or limits;
is traceable to genuine study records;
protects participants and collaborators;
avoids fabricated hindsight; and
does not use reflexivity to justify a claim the evidence cannot support.

If the chain from position to decision to consequence is missing, return to the ledger. The purpose is not to perform self-awareness. It is to make consequential interpretive work clearer and more open to evaluation.

If you want support checking the coherence of a note based on your own study records, research support can help you examine the reasoning and wording. The service does not invent decisions, fabricate an audit trail or produce assessed research under false authorship.