Statistical consulting for researchers

Most statistical problems are created long before the data are analysed. A study that is too small to detect its effect, an outcome measured on the wrong scale, or a design that confounds the question cannot be rescued by a clever test afterwards. We work from the beginning, helping you choose a study design that can answer your question and a sample size with the power to find what you are looking for, so the analysis at the end has something real to find.

The other common failure is reaching for a familiar test that does not match the data. We choose the analysis on the structure of your data and your question, whether that is a regression, a mixed model, a generalised linear model, or a simpler comparison, and we explain the trade-off so you understand why. This is general research statistics; where the numbers sit inside a systematic review or meta-analysis, our systematic review statistics service handles the review-embedded pooling, heterogeneity, and subgroup work instead.

1. 1

   Design and sample size

   We advise on the study design that fits your question and calculate the sample size and power so the study can detect the effect you care about.

2. 2

   Analysis plan

   We write a pre-specified analysis plan that names the primary and secondary analyses, the model, and how missing data and assumptions will be handled.

3. 3

   Model selection and execution

   We choose and run the right test or model in R, Stata, SPSS, or SAS, check the assumptions, and deliver the script so the work is reproducible.

4. 4

   Interpretation, reporting, and review

   We interpret the output in plain language, write the methods and results, and draft the statistical part of your response to reviewers.

You receive the analysis, the figures and tables, and a methods-and-results write-up you can drop into your paper or thesis, each one accompanied by the script or syntax so the work is reproducible. Where effect sizes are central we report them properly, and our explainer on effect sizes shows how we present them. If your project is heading towards pooling evidence across studies, our meta-analysis services carry the quantitative synthesis, while health and clinical projects with trial, survival, or epidemiological designs are better served by our biostatistics consulting specialism. To sanity-check a single estimate before you commit, our free confidence interval calculator works on your own numbers.

• A sample-size and power calculation with the assumptions stated
• A pre-specified analysis plan naming the primary and secondary analyses
• The analysis run with assumptions checked and diagnostics reported
• Plain-language interpretation of what each result means
• A drafted methods and results section in your required style
• The script or syntax and package versions for full reproducibility

We quote a fixed fee per project rather than billing by the hour. The drivers are the complexity of the design and model, how much of the work sits at the planning stage versus analysis and write-up, and whether reviewer responses are likely. Tell us your design, your data, and the question you need answered and we will return a fixed quote.