What is it about?
Sometimes a finding that a program or policy has no effect is due more to how the research is structured than to a true deficiency in the program. Impacts often vary from one person to another. When researchers focus on average treatment effects, they impose an assumption that may not be correct and that may produce misleading results. This is demonstrated through simulations, in which average treatment effects fail to capture the known impacts.
Featured Image
Why is it important?
Researchers are often taught that randomized control trials are the gold standard for research, and that the use of average treatment effects (ATEs) is the best way to analyze RCT data. Based on this approach, many programs are judged to be ineffective. However, ATEs may not provide the best model. Programs may be canceled or have their funding lowered when they actually are effective for many people.
Perspectives
I've learned that the answers you find depend on the questions you ask. This is the problem with imposing a standardized research methodology such as ATEs; it puts the focus on a single research question, which may not the right question to ask.
Dr Bradford W Chaney
Westat Inc
Read the Original
This page is a summary of: Reconsidering Findings of “No Effects” in Randomized Control Trials, American Journal of Evaluation, March 2015, SAGE Publications,
DOI: 10.1177/1098214015573788.
You can read the full text:
Contributors
The following have contributed to this page
