Useful statistical analysis starts with the research question and the actual structure of the data. Variables, assumptions, sample size and measurement quality should be reviewed before choosing tests.
From question to analysis plan
Define which outcome answers the objective, which variables are compared and which analyses are necessary.
Data preparation and quality
Missing data, outliers, units and coding can change conclusions. Preliminary review helps document decisions and limitations.
Data visualisation
Figures reveal distributions, trends, dispersion and anomalies and later communicate magnitude and uncertainty clearly.
Interpret without overstatement
A p-value does not replace effect size, uncertainty or context. Conclusions should stay within the design.
Possible deliverables
How the work is organised
The objective, available documents and data are reviewed first; scope, deliverables, missing information and review points are then defined. Support may cover one phase or connect with methodology, laboratory work, analysis and reporting.
Frequently asked questions
Can an existing test choice be reviewed?
Yes. Its fit with the objective, variables, assumptions and design can be reviewed.
Can completed analyses be reviewed?
Yes. The link between objective, test, table, figure and interpretation can be checked.
Does non-significance mean there is no effect?
Not necessarily. Sample size, uncertainty, power and effect magnitude also matter.