Overview

Autumn 2017
9:00am–11:45am Wednesdays
220 Lazenby Hall

Instructor: Jolynn Pek
Office: 228 Lazenby Hall
Phone: 614 292 4940
Emailpek.5@osu.edu
Office hours: Monday 9:00am–10:00am

Course Description

Power analysis is an important component to research design. In the context of limited resources, which should be responsibly expended, power analysis allows researchers to determine a range of sample sizes which would presumably provide adequate power to detect statistical significance of effect sizes. Underlying concepts to understanding and computing power will be discussed in relation to current methodological concerns in psychological science (e.g., replicability and reproducibility).

Students will learn how to determine an effect size and accompanying error variances required to compute power. Additionally, various approaches to computing power will be introduced such as using canned software (e.g., G*power, R pwr package, and SAS PROC POWER) as well as implementing Monte Carlo simulations to estimate power in R. Additionally, misconceptions, limits, and abuses of power analysis will also be discussed.

This course is appropriate for graduate students in psychology, education, communications, related social sciences, public health, and business. Evaluation will be based on three components: (a) Class attendance and participation, (b) a limited number of homework sets, and (c) power analysis project.

 

Course Objectives

This course is designed to provide students with basic concepts regarding how to compute power, and how to conduct power analyses for common research designs and analyses. Students would also become familiar with ongoing discussions of power in relation to replication.