Seeking Independent Statistical Reviewer for Bayesian Ordinal Mixed-Effects Model (Pre-Submission Methodology Review)

Jobs
1

Hi all,

I’m looking for an independent statistician to review the methodology of a completed research protocol before ethics committee submission and journal review. This is a methodology review engagement, not a full analysis service — the model has already been specified and validated via simulation; I need an experienced second opinion on the statistical design.

Scope of the study: A comparative evaluation of three large language models’ surgical guidance quality, using a Bayesian ordinal (cumulative-logit) mixed-effects model fit with brms/cmdstanr, with crossed random effects (rater, scenario, scenario×model, rater×model, response-level). Sample size was determined via Monte Carlo power simulation across multiple variance scenarios.

What I need reviewed:

  • Model specification and random-effects structure
  • Prior choice and sensitivity
  • The multiple-comparison decision rule and its consistency with the power simulation
  • Overall statistical soundness before submission

Required background: Experience with Bayesian statistics and Stan/brms; prior work with clinical/medical data is a plus. An ASA PStat® or RSS CStat credential is a plus but not required.

Deliverable: A signed, dated review letter/report (with your CV or a brief statement of relevant experience) summarizing your assessment, suitable to reference during ethics committee submission and peer review.

I’d rather not post full protocol details publicly — happy to share the complete methodology document privately with serious applicants under a short confidentiality understanding.

To apply, please send me a DM with:

  1. A brief note on your relevant Bayesian/Stan experience
  2. Your CV or relevant background (papers, GitHub, etc.)
  3. Your estimated rate/timeline for a review of this scope

Thanks!

2

Hi Muharrem,

Are the protocol and brms/Stan simulation study LLM generated?

3

Hi,

Thanks for asking — happy to be transparent about this. The study design, the choice of the Bayesian ordinal mixed-effects model, and the statistical decisions (prior specification, variance-scenario assumptions, etc.) are mine. I used an AI assistant to help draft the protocol text and write/run the brms/Stan simulation code under my direction — which is exactly why I’m looking for an independent statistician to give it a human second opinion before it goes to the ethics committee.

Happy to answer more specific questions if you’d like to take a closer look.

Best,
Muharrem