Qualitative Coding Tool

Export and back up

The CSV and Markdown exports list every passage under its code. The project file holds everything, so you can restore it on another device.

How to Use the Qualitative Coding Tool

1. Add a document. Paste a transcript or notes and give it a title, or choose Try a sample. Anonymize it first if it names participants — the Text Anonymizer does that in your browser.
2. Build the codebook. Add a code, optionally inside a parent (Barriers › Cost), and keep its radio button selected: that is the code you are applying.
3. Code the text. Select words with the mouse or keyboard and the chosen code is applied at once, or switch that off and confirm each selection. The + beside a paragraph codes the whole paragraph. Overlapping passages are fine; each code gets its own passage.
4. Read by code, then export. Filter the excerpts by code, with or without its sub-codes, add a memo to any passage, and download the excerpts as CSV or Markdown. Download the project file as a backup.

What Coding Is For

A code is a short label for what a passage is about or does. As Saldaña's The Coding Manual for Qualitative Researchers describes it, a code is usually a word or short phrase that captures what a stretch of data is about. Coding is how a pile of transcripts becomes something you can compare: once every passage about money carries the same code, you can read them side by side, count them, and ask how they differ between participants.

Codes are not themes. In Braun and Clarke's reflexive thematic analysis, codes are the smaller building blocks, and themes are patterns of shared meaning you construct from them by grouping and refining. The nested codebook supports that move: put Cost and Time under Barriers, review the combined excerpts, and decide whether Barriers is a theme or only a heading.

Reading the Numbers

The number beside each code is the count of coded passages, and the figure in brackets includes its sub-codes. A passage coded twice with the same parent counts twice in the parent's total, so read the bracketed number as how much material sits under this heading, not as the number of distinct quotations. Text coded is the share of the document's characters covered by at least one passage, with overlaps counted once.

Counts describe your coding, not your participants. Five passages about cost can come from one talkative interviewee or from five people, and in qualitative work how often something is said is rarely the evidence you rely on. Use the excerpt list to check whose voice a code rests on.

Worked Example

Try a sample loads a short interview about commuting to university with five codes (Barriers, Cost, Support, Staff, Peers) and five coded passages. Three passages sit under Cost, so Barriers shows 0 (3): none coded to it directly, three in total. Support shows 0 (2). The coded passages cover 61.3% of the text. Filtering to Support with sub-codes included gives:

Storage, Exports and Limits

• Local only. The project is kept in your browser's local storage on this device. It is not uploaded, not synced, and gone if you clear site data or use a private window — so download the project backup (JSON) after each session and restore it with Restore a project backup.
• Text is fixed once added. Passages are stored as character positions, so the text cannot be edited afterwards. To correct a typo, add the corrected document and recode it, or fix the text before you add it.
• One code per passage. To apply two codes to the same words, apply each; they are two passages with the same range.
• Exports. The CSV has one row per passage (document, code path, excerpt, start and end character, memo) and opens in Excel or Google Sheets. The Markdown groups excerpts under their code as nested headings. A separate CSV lists the codebook with counts.
• What it is not. There is no audio or PDF coding, no team projects, and no inter-coder statistics. For a large project use a full CAQDAS package such as NVivo, ATLAS.ti or MAXQDA; this tool suits a handful of transcripts, a pilot, or coursework.

Frequently Asked Questions

Where is my data stored?

In your browser's local storage, on this device only. Nothing is sent to CitationEasy or anyone else. Because it lives in the browser, clearing site data, switching browsers or using a private window loses it, so download the JSON project backup regularly.

Can I use this for thematic analysis?

Yes, as a workspace for the coding stage. Code the transcripts, nest related codes, and read the excerpts gathered under each. Grouping codes into themes, naming them and judging whether they hold up is your analytical work. See Braun and Clarke's 2006 paper "Using thematic analysis in psychology" (Qualitative Research in Psychology, 3(2), 77–101) for the six phases.

How many codes should I have?

There is no right number. Many projects start with dozens of fine-grained codes and merge or nest them as patterns emerge. If a code has one or two passages, ask whether it belongs under a broader code; if a code has hundreds, ask whether it is doing any analytical work.

Can I code a passage with more than one code?

Yes. Select the same words twice with different codes chosen. Each application is its own passage, so the excerpt appears under both codes and counts once in each.

How do I get my coding into Excel or my write-up?

Use Excerpts (CSV) for a spreadsheet with one row per passage, or Excerpts (Markdown) for a document organized by code. Both include the document title and character positions so you can trace a quotation back to its source.

Can two people code the same transcript?

Not in the same project: there is no shared workspace. Each coder can work on their own copy, then compare the exported CSVs. For agreement statistics use the Cohen's Kappa or Krippendorff's Alpha calculators on the two sets of codes.

Should I anonymize before coding?

Usually, yes, especially if you will export excerpts or share the project file. Anonymize first, keep the key file separately, and code the anonymized text.