Correlation or Causation?: The Sorry State of Inference in Empirical Modeling

14 Pages Posted: 16 Mar 2017

See all articles by Xiaojing Dong

Xiaojing Dong

Santa Clara University - Marketing

John M Heineke

Santa Clara University - Leavey School of Business

Date Written: august 12, 2016

Abstract

For decades, statistical methods, many based upon the “general linear model,” have been used to do estimation and test hypotheses in the social and natural sciences, in medicine, and in the private sector. These tools have become increasingly sophisticated and are often paired with powerful open source data analytic software. We now regularly see mathematical/statistical output combined with data visualizations that are truly mind-boggling and, once in a while, thought provoking. But an increasing number of papers and studies appear to have little statistical validity, in which the line between causality and correlation is often non-existent. This is a danger sign not only in science and medicine but also to companies who unwittingly rely on such results for forecasting and business strategy. Could it be true that researchers and analysts who learn ever more powerful analytical methods lack even a basic understanding of the limitations of these methods? The purpose of this short, non-technical paper, which relies heavily upon examples, is to shed some light on the underlying statistical issues. The ideas here are certainly not original with us and have been raised for a number of years across multiple disciplines.

Keywords: Causation, Correlation, Overfitting, p-Hacking, Endogeneity, Identification, Simultaneity, Measurement Error, Data Generation Processes, Inconsistency, Bias

Suggested Citation

Dong, Xiaojing and Heineke, John M, Correlation or Causation?: The Sorry State of Inference in Empirical Modeling (august 12, 2016). Available at SSRN: https://ssrn.com/abstract=2933053 or http://dx.doi.org/10.2139/ssrn.2933053

Xiaojing Dong

Santa Clara University - Marketing ( email )

Santa Clara, CA 95053
United States

John M Heineke (Contact Author)

Santa Clara University - Leavey School of Business ( email )

500 El Camino Real
Santa Clara, CA 95053
United States
408-554-4346 (Phone)
408-554-2331 (Fax)

Register to save articles to
your library

Register

Paper statistics

Downloads
67
Abstract Views
307
rank
342,868
PlumX Metrics