Exploratory Data Analysis

3 credits · Graduate School

Exploratory Data Analysis (EDA) is a modern statistical paradigm developed by John Turkey in the 1970's. EDA emphasizes fitting mathematical models to data as preliminary to the traditional hypothesis testing approach used in confirmatory analyses. Hallmarks of EDA include graphical methods, residual analysis, robust/resistant statistical methods, and data re-expression/transformation. But EDA is actually a whole philosophy of data analysis, and includes treatment of ethics and propriety in research. In this class we study EDA, as it has developed over the past four decades. We also do a great deal of EDA. An "EDA Portfolio" is developed by each student of different data analysis and graphical analysis projects. Included within the course is treatment of "big data" and data mining approaches, and also discussion of the current "replication crisis" and its emphasis on Questionable Research Practices (QRP's); EDA provides a certain type of prescriptive treatment of QRP's. Prerequisites: PSY-GS 8861 and PSY-GS 8870 or equivalent. [3]