Sample size and power calculation
Friday, 17 July (Melbourne City Centre) 9:00 am to 12:30 pm
One of the decisions to be made when planning a research proposal, which can have an enormous impact on the results of the study, is the size of the sample to be drawn. Sample size calculation and power analysis rely on a number of practical and statistical matters, which are not always easy to master. The study power (capacity to detect a treatment effect of a given size) depends on factors such as the type of outcome (and its variability), the study design, the statistical hypotheses, the targeted effect size, the level of significance, and the sample size. Statistical software can help, but we still need to provide the software with right input.
This workshop will present an introduction to power analysis and sample size calculation with a special emphasis on concepts. We will revise concepts underlying hypothesis testing (sampling distributions, statistical hypothesis, level of significance, power, and the type of errors) to help identify the key factors that determine the statistical power. We will then define the input needed to calculate sample size for some common statistical methods and discuss the criteria to select or estimate this input. We will show how the complexity to define the input increases as we move to more advanced statistical methods. Ethical, practical and cost considerations will be discussed.
Data simulation and real scientific studies will be used to exemplify problems associated with sample size determination.
Who should attend?
PhD students and early career researchers within the Faculty of Health.
A working knowledge of sampling distributions, confidence intervals and significance tests for means and proportions as covered in “Workshop: An introduction to statistical reasoning” would be advantageous.
Steve Bowe, PhD M Med Stats B Ed(maths), Senior Researcher Biostatistics
Mohammadreza Mohebi, PhD, Senior Researcher Biostatistics
Liliana Orellana, PhD, Associate Professor Biostatistics