Statistician for physiotherapy researchers: repeated-measures ANOVA, effect size, reliability (ICC), ROC curves and reviewer responses. PhD theses and Q1/Q2.
Statistics for physiotherapy that understands small samples and real designs
PhD theses, Q1/Q2 articles and reviewer responses in physiotherapy, rehabilitation and manual therapy. Small-N clinical trials, repeated-measures ANOVA, mixed models, reliable change index, single-case (SCED), inter-rater ICC and meta-analysis.
Your designs have very specific obstacles: small samples, functional scales, individual change. There's method for all of them.
Your 40-60-patient RCT is real, but the reviewer demands statistical power, ANCOVA with baseline and CONSORT reporting you don't handle. Intent is right, reporting falls short.
You have VAS, functional scales or ROM measured at several time points and ran traditional ANOVA. The reviewer asks for mixed models due to follow-up loss or non-sphericity.
You have a single-case (SCED) or pilot with 8-12 patients. You think it can't be published: turns out it can: there is specific statistics for N=1 designs and clinical pilots.
Mixed ANOVA, Greenhouse-Geisser correction, specific contrasts and when a mixed model is preferable to classical ANOVA.
For longitudinal data with follow-up loss or imbalance. Individual trajectories, random effects per patient and correct CI computation.
Clinically meaningful individual change, minimal important change, responders vs non-responders with psychometric justification.
AB, ABA, multi-baseline designs. Specific statistics: Tau-U, NAP, PND. How to report for journals demanding meta-evidence.
Correct ICC model (1,1 vs 2,1 vs 3,1), standard error of measurement and clinically meaningful cut-offs.
If reviewers ask for power, I help with G*Power or pwr in R and write the analysis up so it holds.
No clock running on your tab. We agree on scope, timeline and a fixed price upfront: no inflated hours, no surprise invoice at the end. Before delivery, every analysis goes through a careful review on my side: I rerun the code, re-check the assumptions and review tables and figures line by line. That's the part I obsess over: that what you receive is ready to defend without errors surfacing in peer review. Bounded projects (RCT, pilot, reviewer response) fit this model with no surprises.
Illustrative example: individual pre-post trajectories on a functional scale (e.g. ODI) with the MIC band, effect size and the proportion of patients exceeding clinically relevant change.
Free resources on the techniques that show up most in physiotherapy theses and articles.
Small-N RCT, single-case, functional scale validation, reviewer response… Fill the form and I respond within 24h with a tentative proposal.
Biostatistics and statistical consulting for health research · All services
Yes, with the right analytical design. Key: ANCOVA with baseline instead of simple ANOVA, CIs alongside p-values, effect size (Cohen's d with 95% CI), intent-to-treat and full CONSORT reporting. Small N doesn't kill your article; a naive analysis does.
Traditional ANOVA works if your data is complete and spherical. With drop-out or non-sphericity (the usual case), a mixed model is methodologically superior and current reviewers prefer it. I can deliver both analyses so you see the difference and argue your choice.
Yes, in specific journals and with serious methodology. There are statistics designed for N=1 designs (Tau-U, NAP, PND) that give effect size and significance. The key is replicable design and transparent reporting.
Yes, those are the ones I see weekly. The statistics any journal asks for functional scales (MIC, RCI, test-retest reliability) are the same regardless of instrument.