A code is a decision, and every decision has to be accountable
Coding looks mechanical and is not. Every code you apply is a judgment about what a piece of text means, and a committee is entitled to ask why you made it. The work here is building a frame that is consistent, documented, and yours.
Get your free quote
We reply within 1 business hour, no delays.
No card details required · 100% confidential · On time, or it's free
Inductive, deductive, or both, and you have to say which
An inductive frame is built from the data. A deductive frame is brought to the data from theory or from prior research. Most real projects use both, which is fine and common, but it has to be stated and the boundary has to be clear. Examiners notice when a candidate claims a purely inductive analysis and then applies a framework from the literature.
The second decision is granularity. Code too finely and you end up with four hundred codes and no analysis. Code too coarsely and you lose the distinctions that make the findings interesting. There is no correct number, but there is a correct process for arriving at yours, and it involves testing the frame on a subset before committing to it.
The third is what happens when the frame changes partway through, which it always does. Changing it is not a problem. Failing to go back and recode consistently is, and so is failing to record that it happened.
- A codebook with definitions, inclusion rules and example quotes
- The frame tested on a subset before it is applied to everything
- Changes to the frame recorded, with recoding done properly
- A clear, stated path from codes to categories to themes
What a defensible codebook contains
The document that answers most viva questions before they are asked
- Each code named, defined, and distinguished from its neighbours
- Inclusion and exclusion rules, so application is repeatable
- An example quote for each, drawn from your own data
- A dated log of changes and the reasoning behind them
- How codes group into categories, and categories into themes
Four things that cost points in the coding chapter
Codes that are topics, not meanings
Labeling a passage Funding records what it is about. It does not record what is being said about it. Topic codes produce descriptive themes, which is the most common criticism of qualitative chapters.
A frame that drifted
Applying a code differently in interview twelve than in interview two. Almost universal, entirely fixable, and only a problem if it is not caught and corrected before submission.
Intercoder agreement misused
A kappa statistic quoted without saying what was double-coded, by whom, or what happened to disagreements. In interpretivist work, agreement figures may not be appropriate at all, and claiming them can undermine your own position.
No trail from code to claim
The findings chapter asserts a theme and the reader cannot see how the data supports it. The fix is structural rather than cosmetic, and it is much cheaper to build in than to retrofit.
Building a frame that survives examination
Read before you code
Whole transcripts, without coding, taking notes. Coding from the first line of the first interview produces a frame shaped entirely by whoever you happened to interview first.
Draft the frame on a subset
Two or three transcripts, coded openly, then reviewed. This is where you find out whether your codes are meanings or topics, and it is much cheaper to find out now.
Define every code in writing
A name, a definition, what it includes, what it excludes, and an example. If you cannot write the definition, the code is not yet a code.
Test consistency
Recode an early transcript after you have coded several more. Where you now code differently, the frame has drifted and needs reconciling. Do this deliberately rather than discovering it at the end.
Group upward, keeping the trail
Codes into categories, categories into themes, with the reasoning recorded at each step. This record is what turns your analysis from a set of assertions into an argument.
You code your data. We teach and review
- We do not code your transcripts for you.
- We do not supply a ready-made coding frame as though it were yours.
- We do not write your analysis or findings chapter.
- We do not act as a second coder for an agreement statistic you then report as independent.
- We teach the method, review your codebook and your coding, and tell you what an examiner will question.
The second-coder point matters. If you report intercoder agreement, the reader assumes two researchers on your team coded independently. A paid tutor doing it is a different thing and would have to be disclosed as such. Read the full policy.
Frequently Asked Questions
There is no right number and it varies enormously by dataset and method. What matters is whether each code is doing distinct work and whether you can define it. If two codes cannot be told apart from their definitions, they are one code. If a code applies to almost everything, it is probably a topic rather than a meaning.
A code labels what a piece of data means. A category groups codes that belong together. A theme makes an analytic claim that runs across the data and answers something about your research question. Most weak qualitative chapters stop at categories and call them themes.
It depends on your methodology. In positivist or post-positivist content analysis, intercoder reliability is often expected. In reflexive thematic analysis, the method's own authors argue against it, because the analysis is understood as interpretive rather than something to be replicated. Doing the wrong one for your stated position is worse than doing neither.
No. Frames change in almost every project, and a frame that never changed usually means the data was forced into it. What matters is that you go back and recode consistently, and that the change and its reasoning are recorded.
Yes, and that is often the most useful session. We review the codebook and a sample of coded transcripts and tell you where the frame is inconsistent, where codes are topics rather than meanings, and what a committee will push on.
Yes, with one difference worth planning for. Focus group data carries interaction as well as content, and analyzing only what was said while ignoring how the group negotiated it is a common weakness examiners pick up on.
Send the codebook and a couple of coded transcripts
We will tell you where the frame is inconsistent, which codes are doing no work, and what stands between what you have and a defensible analysis.
Get Help With My Coding