Causal Effect Model Using Matching Methods in Fisher's Experiment Design and Rubin's Causal Model with Small Participants in a Control Group in Quasi-Experimental Data

• Activity Number: 65
• Type: Contributed
• Date/Time: Sunday, August 2, 2009 : 4:00 PM to 5:50 PM
• Sponsor: Biometrics Section
• Author(s): Gideon D. Bahn*+ and Kathleen Ruroede
• Companies: Loyola University Chicago and Marianjoy Wheaton Franciscan Healthcare
• Keywords: Fisher's experimental design ; Rubin's causal model ; matching method ; quasi-experimental data ; small unit in a control group ; comparing IRFs and SNFs

Abstract

In medical research, randomization is often impossible when participants are diagnosed with a symptom that requires them to be treated in a particular treatment. When two groups of participants are being compared, a serious problem arises as the participants drop out from the study, making the sample size small and/or different between the groups. The present study attempts to offer ways to draw a causal inference in a quasi-experimental study mentioned above, importing the idea of matching method from the Fisher's experimental design and Rubin's Causal Model (RCM). The results suggest that both methods provide similar results. Fisher's design fits better for small data while RCM, for larger data by using propensity scores to balance the matching sets. In order to illustrate, a part of the data funded by American Medical Rehabilitation Providers Association is utilized.