Magazine
How to Get the Most Out of Your Correlation Analysis (303777)
*Stephen Warwick Looney, Augusta University
Keywords: Spearman correlation, Kendall coefficient, Fisher z-transformation, Sample Size, Confidence Interval, Hypothesis Test
Correlation analysis is one of the most commonly used statistical methods. Despite its widespread use, however, there are numerous challenges that have been overlooked by many data analysts, including issues of statistical vs. practical significance and appropriate sample size and power calculations. Failure to use the optimal methods for planning and performing correlation analyses can lead to sample sizes that are too small to adequately address statistical and practical significance. Using less than optimal methods for performing the analyses can lead to incorrect conclusions regarding the true value of the correlation. In this course, many of the challenges associated with correlation analysis are illustrated with real-world examples, and up-to-date approaches and methods for dealing with them are described. Implementing these methods using SAS and R will also be discussed.