Rapid qualitative analysis: steps and trade-offs

Scope: interview and focus group data in applied and evaluation research, analysed with the summary-matrix form of rapid qualitative analysis. All example data are invented.

Short answer. Rapid qualitative analysis (RQA) is a family of techniques for producing qualitative findings in time for a decision. A widely used form, and the one this guide covers, is Hamilton-style summary-matrix analysis. Each interview or focus group is summarised under predefined domains, usually taken from the interview guide, instead of being coded line by line. The summaries are then compared in a case-by-domain matrix. This form is enough when the questions are known in advance, the data are semi-structured and a decision needs results soon. It is not enough when you need a segment-level audit trail or a full interpretive or theoretical analysis. In every case the report must say what was reduced.

The steps in brief:

1. State the question, the deadline and why a rapid approach is justified.
2. Define neutral domains, usually from the interview guide.
3. Build a summary template, pilot it and calibrate it.
4. Summarise each interview or group under every domain.
5. Transfer the summaries into a case-by-domain matrix.
6. Read across the matrix, synthesise, check key points against the data and report the reduction.

Where the method comes from

RQA developed in health services and implementation research, where findings have to reach a programme team while the programme is still running. The summary-matrix version was presented by Alison Hamilton in a US Veterans Affairs cyberseminar in 2013. She later discussed rapid approaches in an introduction to qualitative methods in implementation research (Hamilton, 2013; Hamilton & Finley, 2019). The approach builds on matrix analysis, in which summarised data are arranged in tables so that cases and topics can be compared systematically (Averill, 2002).

Rapid techniques vary. A critical review described adaptations to both data collection and analysis, and found six main reasons for using them: to reduce time, to reduce cost, to collect more data, to improve efficiency, to improve accuracy, and to stay closer to what participants said (Vindrola-Padros & Johnson, 2020). Some rapid designs code instead of summarising: one evaluation team coded detailed interview notes directly into a construct-by-site matrix (Nevedal et al., 2021). PARRQA, a consensus framework, sets out how to plan rapid projects and how to assess and report their rigour. Its five phases run from design to synthesis (Kowalski et al., 2024).

Does it give the same answers?

At least three studies have compared rapid with in-depth analysis of the same kind of data. All had focused questions and a practical purpose:

• In an evaluation of an academic detailing programme, a template-and-matrix rapid analysis gave findings consistent with a later in-depth analysis of the same transcripts, with fewer resources and in time for ongoing implementation (Gale et al., 2019).
• In a comparison across two cohorts of an evaluation, a rapid deductive approach took 409.5 analyst hours against 683 for the traditional one, needed no transcription and still met the evaluation objectives; interpretation took the same time in both (Nevedal et al., 2021).
• In an NHS service evaluation, two teams analysed the same focus group and interview data, one with rapid analysis and one with thematic analysis using the Framework Method. The rapid analysis took 100 hours against 126.5 in total and much less time for data management (43 against 116.5 hours), but interpretation and write-up took longer (52 against 8 hours). Rapid findings overlapped with 79% of the thematic findings, but only about half the recommendations matched (55% and 59%, depending on the direction of comparison). The authors suggest that differences between the teams may explain the long interpretation time (Taylor et al., 2018).

These results broadly support RQA under these conditions, with modest time savings in the third study: focused questions, structured data and a practical decision. They do not show that RQA can replace interpretive or theory-building analysis.

Is RQA the right choice? A quick check

PARRQA notes that rapid methods are best suited to projects with a narrow scope or a focused research question (Kowalski et al., 2024).

The steps

1. Write down the question and why you are working rapidly

State the decision the findings will inform, when they are needed, and why a rapid approach is justified. PARRQA asks teams to record this because it sets the scope of every later step (Kowalski et al., 2024).

2. Define the domains

A domain is a broad topic that every case will be summarised under. Domains usually start from the questions in the interview guide (Kowalski et al., 2024). Give them neutral labels: "Introduction of the pilot", not "Poor communication". A label that contains a finding decides the answer in advance. Add a domain for relevant material the guide did not anticipate, and review whether the domains still fit as analysis goes on.

3. Build and pilot a summary template

The template lists the domains and says what to record under each: the positions people took, who took them, short verbatim quotations with their location, contradictions, and anything not discussed. PARRQA recommends a focused template of about two to three pages and says it should be pilot-tested (Kowalski et al., 2024). If several people write summaries, they should first summarise the same transcript and compare: early summaries vary between too brief, too detailed and too interpretive, and calibration settles the right level.

4. Summarise each case

Read the whole transcript, then fill in the template. Stay close to what was said and give turn numbers or timestamps so that every point can be checked. For focus groups, record who said what, and whether a point was made by one person, contested, or only agreed with ("Mhm. That's exactly it."). Agreement shows alignment with an earlier turn but is not a separate account; the focus group guide explains why.

5. Build the matrix

Transfer the summaries into a matrix with cases (participants or groups) as rows and domains as columns. A spreadsheet or word processor is enough: PARRQA notes that not using specialist software is not a sign of weak rigour (Kowalski et al., 2024). A senior analyst should review the first rows early, before inconsistencies spread.

6. Read across, synthesise and report

Read down each column (how cases differ on one domain) and along each row (how one case hangs together). Synthesis can stay descriptive or move towards patterns and interpretation (Kowalski et al., 2024). In either case, findings should answer the questions and point back to cells and quotations. Check key points against the transcripts or recordings, then state the reduction (see below).

Worked example

Invented example. Rovina Regional Hospital (fictional) pilots a new rota on three wards: twelve-hour shifts and a shared team of nurses who move between wards as needed. After three months, a moderator runs two focus groups with the same topic guide: FG-A with eight ward nurses (N1–N8) and FG-B with six ward managers (W1–W6). Management wants to know how the pilot works, what helps and hinders it, and what changes are acceptable before rollout.

Excerpts (turns are numbered; A-13 is FG-A, turn 13):

[A-13] N3: On the noticeboard. One day it was just up there. Nobody asked us,
           and then they were surprised that we swap shifts among ourselves.
[A-14] N5: It suits me, actually. I have a small child, and twelve-hour shifts
           mean more days off. If only we knew more than a week ahead.
[A-16] N3: […] Someone from the shared team comes in, doesn't know the
           patients, and I have to show her everything anyway. So I'm doing
           two jobs.
[A-17] N6: Mhm. That's exactly it.

[B-31] W2: The nurses complain, but overtime has gone down. That's a fact.
[B-32] W4: It's gone down on the spreadsheet. When they swap a shift between
           themselves, nobody records it as overtime.
[B-33] W2: But that's not the rota's fault.

Domains from the brief: D1 Introduction of the pilot, D2 Working with the shared team, D3 Workload and overtime, D4 Conditions for rollout.

Part of the matrix:

A synthesis finding: both groups mention informal shift swaps. One nurse mentions them right after saying staff were not consulted (A-13). One ward manager says that because of them, recorded overtime may not show the real hours (B-32). Evidence: A-13, B-32. Limits: one speaker per group; D1 and D2 were not discussed in the managers' excerpt, so the groups cannot be compared there.

How far the analysis can go. During synthesis, an analyst could draw the informal swaps, the extra induction work and the unrecorded hours into a higher-level concept such as invisible work that keeps the rota running. This implementation has no separate inductive coding stage that would develop such a concept systematically across all the data, so report how you derived it and what the rapid design limits. For an in-depth account of what the rota does to nursing work, choose a fuller method (see inductive vs deductive coding).

RQA, the Framework Method and thematic analysis

RQA and the Framework Method both use matrices. The difference is the indexing step: in the Framework Method all the data are indexed against a framework before charting (Gale et al., 2013), while summary-matrix RQA summarises straight from the transcript into the template. A domain summary is also not a theme. In reflexive thematic analysis a theme is a pattern of shared meaning built around a central concept, and a summary of what people said about a topic is not one (Braun & Clarke, 2021). For a programme evaluation with criteria and standards, see the Framework Method guide.

What the report must say about the reduction

The methods section should state:

• that each case was summarised by domain and not coded segment by segment, and that no coding template or codebook was built;
• where the domains came from and whether they changed during analysis;
• who wrote the summaries, how they were calibrated, and whether key points were checked against transcripts or recordings;
• that a gap in the matrix may come from the summary rather than from the transcript;
• that saturation is not claimed, and that counts of how many people said something are approximate;
• how any higher-level concepts were derived in synthesis, and what the method cannot show, for example causal effects of a programme (interviews give perceived effects).

Example wording: "We used rapid qualitative analysis. Each focus group was summarised under four domains taken from the brief, and the summaries were compared in a group-by-domain matrix. Transcripts were not coded line by line. Key findings were checked against the transcripts. Matrix gaps reflect topics not discussed or not captured in summaries."

When this is not the right approach

• The research question calls for an in-depth interpretive account, or aims at theory.
• The questions are not yet known, and finding them is the point of the study.
• The findings will be used as evidence in a dispute, an audit or a publication that needs a segment-level trail.
• Language, interaction or narrative structure is the object of study. Summaries remove exactly that.
• You need population-level or cross-case claims that the number or comparability of cases cannot support.

For a decision table across all approaches, see how to choose a qualitative analysis approach.

Running a rapid analysis in the app

SurveysAI Qualitative Analysis Workflow offers a Rapid qualitative analysis that follows the summary-matrix logic. It derives analytic questions and domains from your brief, then summarises each group under every domain: positions with the speakers who hold them, quotations, contradictions, and topics not discussed. It then compares the groups in a group × domain matrix with a synthesis and writes a main report with an executive summary. The report's method statement says that the data were not coded segment by segment and that no coding template was built. Program code checks quotations word for word against the transcripts. An AI critic reads key outputs against the rules of the method. It is the same AI model that wrote them, not a human and not a second model, so this is not independent validation. The app produces draft analytic outputs; you remain responsible for reading the summaries against your transcripts and for the interpretation.

You need transcripts with speaker labels and a brief; an interview guide, field notes and speaker details are optional.

Create an account, and upload your interview or focus group transcripts. If the analysis is not preselected, choose Rapid qualitative analysis in the start dialog.

References

• Averill, J. B. (2002). Matrix analysis as a complementary analytic strategy in qualitative inquiry. Qualitative Health Research, 12(6), 855–866. https://doi.org/10.1177/104973230201200611
• Braun, V., & Clarke, V. (2021). One size fits all? What counts as quality practice in (reflexive) thematic analysis? Qualitative Research in Psychology, 18(3), 328–352. https://doi.org/10.1080/14780887.2020.1769238
• Gale, N. K., Heath, G., Cameron, E., Rashid, S., & Redwood, S. (2013). Using the Framework Method for the analysis of qualitative data in multi-disciplinary health research. BMC Medical Research Methodology, 13, 117. https://doi.org/10.1186/1471-2288-13-117
• Gale, R. C., Wu, J., Erhardt, T., Bounthavong, M., Reardon, C. M., Damschroder, L. J., & Midboe, A. M. (2019). Comparison of rapid vs in-depth qualitative analytic methods from a process evaluation of academic detailing in the Veterans Health Administration. Implementation Science, 14, 11. https://doi.org/10.1186/s13012-019-0853-y
• Hamilton, A. B. (2013, December). Qualitative methods in rapid turn-around health services research [Cyberseminar]. VA HSR&D National Cyberseminar Series: Spotlight on Women's Health.
• Hamilton, A. B., & Finley, E. P. (2019). Qualitative methods in implementation research: An introduction. Psychiatry Research, 280, 112516. https://doi.org/10.1016/j.psychres.2019.112516
• Kowalski, C. P., Nevedal, A. L., Finley, E. P., Young, J. P., Lewinski, A. A., Midboe, A. M., & Hamilton, A. B. (2024). Planning for and assessing rigor in rapid qualitative analysis (PARRQA): A consensus-based framework for designing, conducting, and reporting. Implementation Science, 19, 71. https://doi.org/10.1186/s13012-024-01397-1
• Nevedal, A. L., Reardon, C. M., Opra Widerquist, M. A., Jackson, G. L., Cutrona, S. L., White, B. S., & Damschroder, L. J. (2021). Rapid versus traditional qualitative analysis using the Consolidated Framework for Implementation Research (CFIR). Implementation Science, 16, 67. https://doi.org/10.1186/s13012-021-01111-5
• Taylor, B., Henshall, C., Kenyon, S., Litchfield, I., & Greenfield, S. (2018). Can rapid approaches to qualitative analysis deliver timely, valid findings to clinical leaders? A mixed methods study comparing rapid and thematic analysis. BMJ Open, 8(10), e019993. https://doi.org/10.1136/bmjopen-2017-019993
• Vindrola-Padros, C., & Johnson, G. A. (2020). Rapid techniques in qualitative research: A critical review of the literature. Qualitative Health Research, 30(10), 1596–1604. https://doi.org/10.1177/1049732320921835