A Simple Method to Control Positive Baseline Trend Within Data Nonoverlap

Abstract

Nonoverlap is widely used as a statistical summary of data; however, these analyses rarely correct unwanted positive baseline trend. This article presents and validates the graph rotation for overlap and trend (GROT) technique, a hand calculation method for controlling positive baseline trend within an analysis of data nonoverlap. GROT is validated for controlling positive baseline trend and validated socially by visual analysis agreement. The flexibility and generality of GROT is demonstrated by using it with two alternative slope calculations: White and Haring’s bi-split and Tukey’s tri-split. In addition, GROT is presented as a technique that can be adapted for any non-overlap effect size method; examples here include the original percent of nonoverlapping data and newer nonoverlap of all pairs. examples here include the original percent of nonoverlapping data and newer nonoverlap of all pairs. Caution is urged to control baseline trend only when it is pronounced and reliable. GROT moves the field forward as a robust technique suitable for both visual and statistical analysis.