Thursday, November 28, 2013

Policy: Twenty tips for interpreting scientific claims

William J. Sutherland, David Spiegelhalter & Mark Burgman, 20 November 2013
This list will help non-scientists to interrogate advisers and to grasp the limitations of evidence, say William J. Sutherland, David Spiegelhalter and Mark A. Burgman.
Calls for the closer integration of science in political decision-making have been commonplace for decades. However, there are serious problems in the application of science to policy — from energy to health and environment to education…….Of course, others will have slightly different lists. Our point is that a wider understanding of these 20 concepts by society would be a marked step forward.
1.       Differences and chance cause variation.
2.      No measurement is exact.
3.      Bias is rife.
4.      Bigger is usually better for sample size.
5.      Correlation does not imply causation.
6.      Regression to the mean can mislead.
7.      Extrapolating beyond the data is risky.
8.     Beware the base-rate fallacy.
9.      Controls are important.
10.  Randomization avoids bias.
11.   Seek replication, not pseudoreplication.
12.  Scientists are human..
13.  Significance is significant.
14.  Separate no effect from non-significance.
15.   Effect size matters.
16.  Study relevance limits generalizations.
17.   Feelings influence risk perception.,
18.  Dependencies change the risks.
19.  Data can be dredged or cherry picked.
20. Extreme measurements may mislead.
To Read More……

No comments: