Power Analysis in research studies
Overview
Power and Sample size estimation represent one of the most important decisions researchers make. Intervention studies which enrolled too many patients might expose people to unnecessary risks. Intervention studies enrolling too few patients fail to detect the desired effect under the study and they expose patients to possible and needless risks. Accurate power assessment must be an integral part of any grant proposal and research study plan.
This course is aimed at researchers in any field who design or conduct research studies. Examples of power/sample size and sensitivity analysis for a variety of design and model parameters will be discussed, including examples of multiomics studies.
The course will be held on-site over two afternoons on 15 and 16 October 2024, 14:00 - 17:00.
Successful completion of the assignments will be awarded with 0.5 ECTS by the Doctoral School to PhD students of University of Luxembourg.
Learning objectives
- Conceptual understanding of statistical power in research projects
- Learning misconceptions and pitfalls
- Knowing the key factors influencing power
- Gain experience with tools/software for conducting power analyses
- Reporting power assessments into research study plans
Requirements
Prior Knowledge
No prior requirements are needed, however this course assumes you are comfortable with math and inferential statistics.
Material
A laptop is required to follow the examples and assignments. A few examples will require a recent version of R (4.2.0 or later) and RStudio installed. Please look at install tutorial to set it up prior to the course. If you’re not familiar with R, you can skip the examples and the installation.
Agenda
Day 1
- Motivations and principles of power and sample size estimation in research studies
- Main factors influencing power
- Practical applications using G*Power and pwr R package
- How to write power and simple size section for grant proposals and peer-reviewed articles
Day 2
- Simulations for complex study designs and multiple testing correction methods for omic experiments
- Power assesment in multiomics experiments
- Power assesment in RNAseq experiments
- Practical applications using MultiPower and PROPER R packages
Registration
Registration is open with limited places. Therefore, it is confirmed on a first-come, first-serve basis.
Address
The workshop will be held at Roudeneck BTL building on 4th floor, meeting room 04.020 University of Luxembourg, Belval campus 1, Boulevard du Jazz Esch-sur-Alzette
Instructor and Contact
Veronica Codoni (LCSB, Elixir-Luxembourg)
