A long pondered but only lately realized blog about economics, politics, evaluation, econometrics, academia, college football and whatever else comes to mind.
A list of euphemisms, drawn from the published literature, for results that do not obtain statistical significance at conventional levels.
Of course, it makes little sense to set up an arbitrary, binary cutoff level for the p-value, but that is, sadly, the world that we live in as we await the casual bayesian revolution.
The evaluation was performed by Mathematica Policy Research on behalf of the US Department of Labor. I would say (and I might be biased as I provided some comments at various points) that the authors have done the best they could with the available data and variation. We could, of course, have better data and variation if we really wanted to.