by Bruce Abramson
Real Clear Politics
Every now and then, I’m lucky enough to meet someone who “follows the science.” I count on such folks to teach me some science that I do not yet know. Being scientifically literate, I like to start by asking them some basic questions:
How are key data terms defined? How are data collected and reported? What theories guided the design of the models that process the raw data? What studies validated the models? How sensitive are the models to variations in inputs? How well do the models perform using historical data? Do the models have a track record at prediction — and if so, how well have they done? What alternative hypotheses were considered? How were the hypotheses tested?