Data analysis guidance
Academic integrity
This guidance supports planning and revision of your own work. It does not offer ghostwriting, invented sources, or a guarantee that a journal will publish a manuscript.
Before any software, write what you are looking for and what would surprise you. Quantitative work needs a reason for each comparison. Qualitative work needs a trace from an extract to a theme, so a reader could disagree with you by looking at the same words. Keep the raw material secure and separate from what you are willing to publish.
Do not let a tool choose the claim. A chart, a code, or a model is an instrument. If you cannot explain the step without the brand name, you do not yet understand the step. Report missing data and small numbers plainly.
A practical sequence
- Write the question the analysis answers.
- Keep a log of decisions, including discarded codes or tests.
- Tie each claim to a table, extract, or count a reader can see.
- List what the material cannot support.
Examples
- “Eight of the twelve essays received a comment on grammar and none on the argument. That count does not show what learners did next.”
