When to hire a statistician
The earlier a statistician is involved, the cheaper good statistics become. Advice at the design stage can fix problems that no analysis can repair later: an outcome measured on the wrong scale, a comparison group that cannot answer the question, or a sample too small to detect any plausible effect. That said, most people contact us at one of four moments:
- Planning. Writing a protocol, grant or ethics application that needs a sample size justification and an analysis plan.
- Analysis. Data are collected and the team needs someone to run, or check, the models.
- Revision. A journal reviewer has questioned the statistics, and the reply needs to be technically airtight. Our response to reviewers support covers the wider letter.
- Second opinion. A supervisor, collaborator or AI tool produced an analysis and you are not sure it holds up.
Consulting areas and typical questions
| Area | Examples of what we help with |
|---|---|
| Study design | Randomisation schemes, stratification, cluster designs, choosing primary and secondary outcomes |
| Sample size and power | Calculations for comparisons of means and proportions, regression, repeated measures and cluster-randomised designs, including design effects |
| Biostatistics | Survival analysis, diagnostic accuracy (sensitivity, specificity, ROC curves), agreement statistics, adjusted odds and risk ratios |
| Survey statistics | Weighting, complex sampling designs, scale validation, non-response |
| Longitudinal and multilevel data | Mixed-effects models, generalised estimating equations, growth curves |
| Social science and economics | Panel regression, instrumental variables, mediation and moderation, structural equation models |
We work in R, Stata, SAS and SPSS, and can match whatever your team already uses so the code stays maintainable after we step away.
How an engagement runs
Every consultation starts with a scoping conversation. We ask what decision the analysis is meant to support, who will read the results, and what constraints apply, such as a pre-registered plan, a reporting guideline like CONSORT or STROBE, or a fixed submission date. You then receive a written scope and quote.
During the work, the consulting statistician keeps a decision log so that anyone on your team can see why a model was chosen, how missing data were handled and which sensitivity analyses were run. At the end, you get the code, the output and a methods write-up that matches what was actually done. For larger projects we can schedule regular check-ins across several months instead of a single handover.
If you are a doctoral researcher rather than a research team, our dissertation statistics help is set up around thesis rules and supervision.
What statistical consulting costs
We do not publish fixed prices because the effort varies so much. A power calculation for a two-arm comparison may take an hour or two; a reanalysis of a multi-site longitudinal dataset can take weeks. Cost depends mainly on how clean the data are, how complex the model is and how much written explanation you need. Send a short description through our quote form and you will have a free, itemised estimate within a few hours, with no obligation.
For grant applications, we can usually turn around a sample size justification and short analysis plan within a few working days, which helps when a funding deadline is close.
Independence and transparency
A consultant is only useful if they are willing to say an analysis does not support a conclusion. We report what the data show, recommend against analytical choices that look like p-hacking or selective reporting, and document every step so results can be reproduced. Where our contribution is substantial, we discuss acknowledgement or co-authorship using recognised criteria such as the ICMJE guidelines, so there are no surprises at submission.
What you get
- A short scoping call and written summary of the statistical questions to solve
- Sample size or power calculations with stated assumptions you can cite
- A statistical analysis plan suited to a protocol, grant or pre-registration
- Analysis with reproducible code and a methods paragraph you can adapt
- Independent review of an existing analysis, with prioritised recommendations
- Support drafting statistical responses to peer reviewers
What we don't do
- We advise and analyse; authorship decisions follow your field's criteria and are discussed openly at the start
- We do not provide regulatory submissions requiring a sponsor's named trial statistician
- We cannot certify results we have not been able to reproduce from the underlying data
How it works
Describe your project
Level, deadline and what you need help with.
Agree a written scope
What's included, the price and the timeline.
Work in stages
Check-ins so you can steer the work.
Review and ask
Explanations and follow-up questions included.