Data analysis help begins with the research question and the evidence you collected. The goal is to choose a transparent analysis, check its assumptions and explain the result accurately. Software output is evidence to interpret, not a conclusion by itself.
A defensible analysis workflow
01
Prepare
Document variables or transcripts, missing information, exclusions, coding decisions and changes from the original plan.
02
Analyse
Use descriptive, inferential or qualitative techniques that answer a stated question or objective and meet the relevant assumptions.
03
Interpret
Explain direction, size, uncertainty and practical meaning; compare the finding with literature without extending beyond the design.
Quantitative BIM data
Questionnaire or project data normally need a codebook and cleaning record before analysis. Start by describing the sample and variables. Then choose relationship or difference tests only where the research question, measurement level, design and assumptions support them. Report effect size or uncertainty where appropriate, not only whether software labels a result significant. A cross-sectional association does not by itself demonstrate causation.
Qualitative BIM data
Interview, observation or document analysis should be traceable from source material to codes, categories and themes. Compare accounts, search for evidence that challenges an early interpretation and distinguish a recurring theme from an isolated comment. Quotations illustrate an analysis; they do not replace it. Explain how your position, sampling and coding process shape the interpretation.
Present findings so a reader can audit them
- Tables: use descriptive titles, units, sample sizes and notes so the table can be understood without guessing.
- Charts: choose a form that matches the comparison, label axes and avoid decorative effects that distort scale.
- Themes: name each theme as an analytical idea, define it and support it with a pattern of evidence.
- Narrative: state the result, explain what it means for the objective and acknowledge uncertainty or alternative explanations.
What to send for an analysis review
Send an anonymised description of the dataset, research question, objectives, instrument, codebook and analysis attempted. For qualitative work, explain the sample, interview guide, transcription status and current coding method. Do not send identifiable participant data or confidential project material in an initial enquiry.
Describe your data and the result you need to understand.
Sources used to shape this guidance
Use the edition required by your university and check each source against your actual research context. These links provide methodological or professional guidance; they do not replace peer-reviewed literature for your topic.
Government Functional Standard GovS 010: Analysis: Current expectations for planning, producing and communicating robust analysis.
The AQuA Book: Government guidance for proportionate quality assurance of analytical work.
Government Analysis Function questionnaire guidance: Design and testing principles that affect the quality of questionnaire data before analysis.