Interesting how many of these "science isn't working anymore" stories you see lately. First Silberman complains that the placebo effect is "getting stronger." Then this article by Lehrer suggesting that "the scientific method" is "not not working." Both of these are silly theses, and not likely to be taken seriously by scientists. The essays always deal with clinical trials in diseases whose symptoms are extremely subjective. There's no mystery, there. It's just hard to pin down all of the variables and change only one in a clinical trial. When you consider how difficult it is to get consent to participate in a clinical trial from a newly diagnosed schizophrenic, you begin to comprehend the problem. Science is still working just fine.

That was indeed a very interesting article about the so-called decline effect. The one thing I think they missed is that they posited regression towards the mean as an explanation, said that couldn't be so in many cases as the studies were based on solid data, then refuted their own refutation by saying in fact positive effects were more likely to appear in studies with smaller numbers of subjects.

I don't think there's anything wrong with the scientific method. What's wrong is the way we play around with it, for example the well-known ploy of describing results as a "trend" when they fall below the level of statistical significance. Non-replicability is always going to happen because of regression towards the mean, but rigor in data collection plus a policy shift in journals so that negative results are as easy to get published as positive ones would minimise it.

A: Light and locomotive are not mutually exclusive possibilities. When the answer cannot be known, I prefer to hope for the light and try to prepare for the train (as much as one possibly can).

Obviously this is a lead-up to the Change, where all technology fails and is replaced by magic. Stephen Boyett and Steve Stirling were right, then, they just had the wrong dates . . .

I was reminded of

http://en.wikipedia.org/wiki/The_Practice_Effect

for some reason.