Qualitative material becomes codable when you can make a defensible decision about what a segment means, what it belongs to, and why that decision matters for the research question. “Codable” does not mean that a transcript is perfectly tidy or that every sentence must receive a label. It means the material is sufficiently interpretable, contextualised, and ethically usable for systematic analytic attention.
Start with a clear analytic unit
Before importing material, decide what you will treat as an analytic unit. This might be a sentence, a conversational turn, an answer to an interview question, a paragraph, a visual region, or a time range in a recording. The right unit is determined by the question and method—not by what is easiest to select on screen.
For example, a study of how teachers justify classroom decisions may code a complete explanation, rather than individual words. A visual study may code a bounded area of a drawing together with the child’s verbal explanation. A conversation analysis project may require turns, pauses, overlaps, and timestamps. In each case, keep enough surrounding context to prevent a short excerpt from changing meaning.
Use four readiness checks
- Relevance: The material can inform at least one research question.
- Context: You can identify who produced it, when, in what setting, and what comes before or after the segment when that matters.
- Interpretability: The material has enough clarity to support a reasoned coding decision. Ambiguity can itself be coded; unintelligibility should not be disguised as analysis.
- Ethical permission: You have the necessary rights, consent, anonymisation plan, and security arrangements for the intended analysis.
Prepare the material without flattening it
Preparation is not a neutral cleaning exercise. Correcting a transcription, translating an idiom, cropping an image, or removing a speaker label can alter the evidence available to the analysis. Preserve an original copy, document material changes, and keep links between transformed material and its source. For audio and video, decide whether timing, delivery, or non-verbal action is analytically relevant before reducing the recording to plain text.
What to record at project setup
A small amount of project information pays off later: the research aim, questions, data sources, participant or case context, language, analytic approach, and any important methodological decisions. These details help you interpret patterns responsibly and make later reporting less dependent on memory.
Common misconception: everything must be coded
Not necessarily. A defensible analysis explains how decisions were made about relevance, not how many highlights were created. Routine greetings, administrative discussion, corrupted recording, or material outside the analytic focus may remain uncoded. Record the reason where the omission could affect how readers understand the dataset.
Questions researchers often ask
Can a spreadsheet be coded?
Yes, when rows or fields represent analytically meaningful cases or responses. Preserve the row metadata so comparisons remain interpretable.
Can a scanned PDF be coded?
Yes. A usable text layer supports text selection; otherwise, a visual region can be coded. Check OCR quality before relying on extracted text for searching or quotation.
Further reading
- COREQ reporting guidance for interview and focus-group reporting.
- Byleron One Help Center for supported sources, imports, coding, and project setup.