Subject Rational Expectations Will Contaminate Randomized Controlled Medical Trials

27 Pages Posted: 7 Jul 2016 Last revised: 11 Dec 2017

See all articles by Gilles Chemla

Gilles Chemla

Imperial College Business School; CNRS ; Centre for Economic Policy Research (CEPR)

Chris Hennessy

London Business School

Multiple version iconThere are 2 versions of this paper

Date Written: June 2016

Abstract

We develop a rational expectations model of placebo effects. If subjects in seemingly-ideal single-stage RCTs form rational beliefs about breakthroughs based upon personal physiological responses, mental effects differ across medications received, treatment versus control. Consequently, the average cross-arm health difference becomes a biased estimator of the mean non-placebo physiological effect. Constructively, we show: bias can be altered through choice of control; high-efficacy controls mitigate upward bias; and unbalanced panels may be preferred since bias approaches zero as treatment probability approaches zero. Consistent with experimental evidence, our theory implies outcomes within-arm and cross-arm differences can be non-monotone in treatment probability.

Keywords: Bias, Rational Expectations, RCTs

Suggested Citation

Chemla, Gilles and Hennessy, Christopher, Subject Rational Expectations Will Contaminate Randomized Controlled Medical Trials (June 2016). CEPR Discussion Paper No. DP11360. Available at SSRN: https://ssrn.com/abstract=2804538

Gilles Chemla (Contact Author)

Imperial College Business School ( email )

South Kensington Campus
London SW7 2AZ, SW7 2AZ
United Kingdom
+44 207 594 9161 (Phone)
+44 207 594 9210 (Fax)

CNRS ( email )

Dauphine Recherches en Management
Place du Marechal de Lattre de Tassigny
Paris, 75016
France
331 44054970 (Phone)

Centre for Economic Policy Research (CEPR)

London
United Kingdom

Christopher Hennessy

London Business School ( email )

Sussex Place
Regent's Park
London, London NW1 4SA
United Kingdom

Here is the Coronavirus
related research on SSRN

Paper statistics

Downloads
0
Abstract Views
340
PlumX Metrics