Session S65a (Wednesday, 9:30am, Thomas 201)

P703: Avoiding the spectre of no significant difference: Determining necessary sample size via power analysis

Xiaoying Xu, Jennifer Lewis, Sachel Villafane (University of South Florida, USA)

Power analysis enables researchers to answer the question: How many research participants are needed to achieve a significant difference when there is a real treatment effect? This is an important question for researchers to consider before beginning a study, preventing a potential failure to find significance because of lack of power. Statistically, the power for a t-test or ANOVA is the ability to detect a significant group difference when this difference is real. Four factors that can increase statistical power will be discussed: 1) increased alpha level, 2) larger sample size, 3) larger effect size, and 4) a one-tailed test rather than a two-tailed test. In addition to discussing the conceptual basics, this presentation will demonstrate how to use power analysis software to determine the minimum number of participants for various scenarios. This tool will help researchers polish research plans earlier and avoid engaging in work likely to be inconclusive.


A sample text widget

Etiam pulvinar consectetur dolor sed malesuada. Ut convallis euismod dolor nec pretium. Nunc ut tristique massa.

Nam sodales mi vitae dolor ullamcorper et vulputate enim accumsan. Morbi orci magna, tincidunt vitae molestie nec, molestie at mi. Nulla nulla lorem, suscipit in posuere in, interdum non magna.