Internet website http://www.foundationwebsite.org, updated 23 October 2020. Books and articles on Planetary Management and other topics from Tucson, Arizona. Copyright © 1999-2022 Joseph George Caldwell. All rights reserved.
Days 4 and 5: Causal Inference and Matching.
There is at present a serious crisis in the field of research design and analysis. The problem is that many researchers do not know how to properly construct a research design involving matching, and they often incorrectly analyze data from a research design that includes matching. This situation, which has existed for a number of years, has been documented in peer-reviewed journals. If matching is done incorrectly, the statistical power of the research design to detect effects of interest may be substantially reduced, or the sample size required to detect effects of a specified size with a specified probability may be substantially increased. When matching is done incorrectly in biomedical applications, lives may be lost or research funds may be wasted. The above link is to lecture notes on a course that describes how to perform matching in a way that takes into account a causal model.