Observation Oriented Modeling Versus Tests of Significance: Case Studies in Analyzing Experimental Data

• In Event: Poster Session for Quantitative Methods and Statistical Theory

Abstract

The aim of this paper is to share with education researchers a new method of data analysis. Using data from three original experiments as case studies, this paper compares the analysis of results using traditional parametric techniques to that of a novel form of data analysis, Observation Oriented Modeling (OOM). As a software package and methodological framework, OOM offers researchers the ability to examine patterns at the level of the unit of analysis without the calculation of a single mean value. This method conceptualizes and presents effect size as the percent of persons or cases that match a predetermined or algorithm-detected pattern in data. This enables a more direct analysis of variance in terms of individual, organizational and community characteristics.

Authors

• Mark J. Garrison, West Texas A&M University

• Maria C. Van Remmen, Buffalo Public Schools