Experimental Design

Author

Vicki Hodgson, Martin van Rongen

Published

August 2, 2023

Overview

This one-day course is designed to complement training in statistical analysis, and focuses on how to design effective experiments while bearing in mind the planned analysis.

Included topics:

  • Setting a good research question
  • Choosing & defining variables
  • Confounding variables
  • Independence & pseudoreplication
  • Revisiting statistical power
  • Case study examples, for discussion
Learning Objectives
  • Feel confident designing experiments with statistical analysis in mind
  • Understand common “pitfalls” that occur when designing experiments, and how to avoid or combat them
  • Apply these skills to at least one case study example

Prerequisites

Knowledge of core statistical concepts, including the statistical inference framework, linear modelling and power analysis, are required for the course. We recommend that students have attended the Core Statistics course or an equivalent.

Some of the course materials have been created using R; users may wish to follow along by copying the code themselves. If so, knowledge of statistical analysis in R is preferred.

Authors

About the authors:

  • Vicki Hodgson
    Affiliation: Bioinformatics Training Facility, University of Cambridge
    Roles: writing - original draft; conceptualisation; coding; creation of synthetic datasets
  • Martin van Rongen
    Affiliation: Bioinformatics Training Facility, University of Cambridge
    Roles: writing; conceptualisation; coding