Qualitative depth. Big Qual breadth.

Big Qual work can move between the breadth of a large collection and the depth of a situated account. A useful starting question makes both movements possible: what would you need to compare across records, and what would you need to understand within a case?

Write a short analysis brief

1. 01
   State the purpose and intended readers of the work.
2. 02
   Name the source, time window, selection process and unit of analysis.
3. 03
   Explain how you will select records for close reading and revisit exceptions.
4. 04
   Decide what evidence would support a claim, and what the material cannot establish.

The breadth-and-depth tradition connects broad exploration, purposive selection and fuller interpretation of cases. The NCRM resources on page 16 offer methodological grounding. SAM provides workspaces for carrying out and documenting these analytical choices.

Prepare a working copy of your material. Retain stable identifiers and a route back to the original source. For interviews, keep the interview or case identifier beside each excerpt. For surveys, preserve the question wording and the response identifier. For public discussion, retain the source URL and collection date.

For sensitive material, use the access, consent, anonymisation and retention arrangements agreed for your project. Review the actual deployment and account arrangements before uploading. Prepare teaching material separately from working research data.

The companion CSV contains 24 fictional excerpts from 12 invented interviews. Each interview has two excerpts and a stable case identifier. Use it to practise the workflow, then adapt the research question and codebook to your own study.

First reading

Together, the excerpts suggest a distinction between interpersonal kindness and coordination across services. A first memo might ask who is doing the work of connecting information, and what changes when a named contact becomes available.

In SAM, upload the teaching CSV, select the text column and preserve interview_id in the dataset. Open Classic Coding and read each excerpt alongside the other excerpt from that case. Record your initial questions before applying a broad coding scheme.

Start with a small set of codes linked to the question. The companion codebook offers five candidates. You can revise, combine or replace them as the analysis develops.

A worked decision

SYN-CARE-07-A can receive Continuity and coordination because the speaker describes carrying information between services. Self-advocacy is a possible second reading, but the excerpt does not say that they challenged a decision or actively pressed for a need. Keep that distinction in the boundary note.

In Classic Coding, use Qualitative and review Codes and scheme. Choose a preset, your custom codes, or a combined scheme. Add definitions or upload a codebook using the current sample format. Review any mapping of extra columns, and keep the complete original codebook with the project.

Sentiment and stance

Sentiment addresses evaluative tone using the selected scheme. Stance addresses a position towards an explicit target. For SYN-CARE-06-B, a stance lens could ask about keeping a telephone option alongside online forms. Define that proposition before choosing categories. Avoid inferring a position on the wider service from a narrow preference.

Extend the review in Adaptive Coding Studio

Bring the selected coding approach into the Studio. Review code definitions and records, develop learning examples, and examine a candidate profile on separate evaluation material. Approve the version you choose; then select its application scope and inspect the resulting coding.

Review excerpts grouped by code, then return to the interview pairs. Ask whether the same code describes different experiences. Practical access includes timing, travel, caring responsibilities and digital arrangements; each may have a different relationship to the encounter itself.

An example analytical memo

Across the invented accounts, access and the experience of the consultation can vary independently. SYN-CARE-03-A describes convenient timing with a rushed conversation; SYN-CARE-11-B describes a helpful encounter after difficulty attending. These contrasts suggest examining the pathway into care as well as the encounter.

1. 01
   Gather the supporting and contrasting record IDs.
2. 02
   Read the surrounding account and preserve the speaker’s distinctions.
3. 03
   State the interpretation in terms of this material and this sampling design.
4. 04
   Export coded texts and the codebook; retain a memo explaining how the interpretation developed.

Results can provide further views of the corpus. Narrative Synthesis can help organise an account around evidence anchors; follow those anchors back to the source. Narrative Timeline needs usable dates and topic clustering. The undated teaching CSV is intended for close reading and synthesis, rather than a timeline.

A public video gives commenters a shared point of reference. Begin by watching or reading enough of that source to understand what the discussion responds to. Record the video, your question and the collection date before collecting comments.

The companion webinar example uses “Dealing with the death of a teammate // DOCTOR Covid-19 Vlog #31”. Its URL is https://www.youtube.com/watch?v=oQCgbNeoPl4. It provides a collection starting point; the examples on pages 10-11 are fictional comments written for teaching.

Comments are a self-selecting discussion around this video. Deleted, unavailable or unreturned material may be absent. A collection limit is a retrieval boundary, and a returned row is not automatically a distinct participant.

1. 01
   Open SAM Collect and choose YouTube Comments. Enter one public video URL.
2. 02
   Choose the requested limit. The inspected interface offers 50, 100, 200, 500, 700 or 1,000. Check the account’s collection allowance.
3. 03
   Run Collect for Review when available, then inspect the returned records, source fields and gaps.
4. 04
   Use Save Snapshot to retain a selection, or Save to Analyse to add the material to My Corpora. Record what you retained.

YouTube Videos & Transcripts is a separate source choice. Use Comments for this question. If collection is unavailable, read the source status and allowance message before changing the design. Keep a durable project record of the provenance and selection: Collect displays a 30-day source-material notice.

Comment A offers condolence and recognition. Comment B combines concern with an expectation of organisational support. These are possible interpretations to discuss, rather than model answers.

Read a sample independently before comparing SAM’s reading with researcher coding. Where readings differ, inspect the text, the target and the code definitions. Preserve uncertainty when the comment is too brief or its reference is unclear.

C may extend care to colleagues and families. D introduces comparison and a question about subsequent support. A codebook can record those differences instead of grouping every sympathetic response into one category.

1. 01
   Review a varied initial set: short and long responses, common language and less frequent perspectives.
2. 02
   Use Topics, sentiment or a relevant Results view to find areas for further reading. Check how the displayed grouping was produced.
3. 03
   Return to source text and available conversation context. Read references such as “that” or “afterwards” in relation to the video.
4. 04
   Revise code definitions, review affected records and export the coded material with a dated memo.

If your export retains reply relationships, use them when interpreting exchanges. If it does not, record that limit and inspect the source where appropriate. Do not reconstruct a conversation from row order alone.

Choose views that answer the research question and are supported by the dataset. Preserve denominators: records, comments, interviews and people are different units. A code may occur more than once in a case, and a record may carry several codes.

For visual material, open the dedicated Visual Coding workspace in Results, or inspect the media galleries associated with Tone of Voice and Framing. Apply your own tags and inspect captions and images in context. Record how visual and textual readings informed each other.

Pathways can examine reviewed event sequences. Interviews can explore meanings and experiences. Parliament Written Q&A can show how an issue is raised and formally answered. OpenAlex Works and UKRI Gateway to Research can add publication and funded-project context.

Use quality evidence precisely

Researcher agreement, AI performance against an adjudicated reference, and repeatability between AI runs answer different questions. Report the relevant sample, missing decisions and procedure with each measure. Use the method-contract version displayed with the project.

For a small study, individual comparisons and qualitative review may be the evidence you can support. If SAM marks the overall Coding Quality Score unavailable, describe the available evidence and the unmet requirements. Repeated readings of the same material do not create more distinct records.

1. 01
   Question: can a reader see what the analysis was designed to understand?
2. 02
   Material: are sources, boundaries, dates, omissions and selection decisions recorded?
3. 03
   Context: can coded records be traced to the relevant case, question or source?
4. 04
   Definitions: are code boundaries, the sentiment scheme and any stance target explicit?
5. 05
   Variation: have you read contrasting cases and material that challenges the current interpretation?
6. 06
   Review: are researcher, AI-assisted and independent comparison activities described accurately?
7. 07
   Claims: do statements match the source, sample and units actually analysed?
8. 08
   Outputs: are excerpts, captions, identifiers and export versions checked for the intended audience?

Save a short decision trail

A useful memo can be a paragraph: the question at this stage; the material you inspected; the change in interpretation; the records supporting it; and the next reading you plan. Keep these notes with the codebook and exports.

Explore in SAM

Classic Coding introduces close reading, codebooks, sentiment and stance. Adaptive Coding Studio follows reviewed examples through learning, evaluation, approval, application and review. Explore both workflows in the product walkthroughs on the SAM website.

The Methods & Practice companion series connects a synthetic Pathways example, a separate interview teaching dataset, and a bounded Parliament collection. The in-app manual is prepared separately for future authenticated deployment.

Methodological reading and source documentation

NCRM: “Big qual: the breadth-and-depth method of working with large amounts of qualitative data” explains the breadth-and-depth approach: https://www.ncrm.ac.uk/news/show.php?article=5721

NCRM: Big Qual: A Guide to Breadth-and-Depth Analysis, and the authors’ resources: https://www.ncrm.ac.uk/news/show.php?article=5813 and https://bigqlr.ncrm.ac.uk/resources-3/

Parliament Written Questions, example 109583, answered 13 February 2026: https://questions-statements.parliament.uk/written-questions/detail/2026-01-29/109583

UKRI: A guide to Gateway to Research: https://www.ukri.org/publications/gateway-to-research-guide/a-guide-to-gateway-to-research/