Research Note: Standard Scores Based on the Median and Inter-quartile Range

Recently when asked to help with statistical analyses of the Medicare Provider Analysis and Review (MEDPar) Data, I encountered a common statistical and analytical problem with this national database, which is also common to many national and health databases. This problem was the problem of extremely skewed distributions (g1 and/or g2 typically exceeding 20), which creates numerous difficulties in comparing and interpreting values and results both between and within variables, never mind statistically testing them.