TY  - JOUR
T1  - EMergence of large treatment effects from small trials
AU  - Batterham AM, Hopkins WG
Y1  - 2013/02/27
N1  - 10.1001/jama.2012.208828
JO  - JAMA
SP  - 768
EP  - 769
VL  - 309
IS  - 8
N2  - 
The authors concluded that large, nominally statistically significant effects in single small trials should not be trusted because these findings may be a result of a combination of luck and statistical artifact. Our concern is that Pereira et al have created the misleading impression that small studies are inherently flawed. Sample size does not bias the outcome of a study. The culprit is what happens afterward: publication bias.

SN  - 0098-7484
M3  - doi: 10.1001/jama.2012.208828
UR  - http://dx.doi.org/10.1001/jama.2012.208828
ER  -