Being a statistical consultant fundamentally hinges on effective communication. It’s about ensuring a clear, mutual exchange of ideas, requiring active participation from both the consultant and the client. For a consultation to truly succeed, it's crucial that clients articulate their expertise just as effectively as the consultant provides clear, actionable statistical guidance. In this talk, I'll share insights from my early experiences as an academic statistical consultant, highlighting the strategies and approaches that enhance the effectiveness of consultations. We'll examine the process from both perspectives: how clients can prepare and engage in a way that maximises their benefits, and the best practices I’ve adopted as a consultant to offer the most useful advice. This discussion aims to shed light on achieving a productive consultation for both parties involved and serve as a useful insight for those who may be engaging in statistical consultation in the future.

Sample code for MCMC for hierchical modelling in R winbugs

This document will help you to prepare for the training school on Bayesian methods being held in Larissa, Greece, 22 - 24 April 2024. The training school involves both introductory didactic teaching and hands-on practical sessions on Bayesian methods for true prevalence estimation & diagnostic test evaluation. The following core topics will be covered: - Bayesian Markov Chain Monte Carlo (MCMC) - Apparent & true prevalence estimation - Test evaluation without cut-offs - Test evaluation with Hui-Walter models Primary learning objectives of the school are that participants (i) become familiar with methods for true prevalence estimation and diagnostic test evaluation (ii) understand the basic principles of MCMC and where Bayesian Latent class models can be applied.

Tidying data with dplyr, using custom ggplot functions, preforming parametric and non parametric means tests among three or more groups.