One of the greatest science news last year was the Science article “Estimating the reproducibility of Psychological Science”, an attempt to repeat the results from the 100 psychological studies published in leading scientific magazines.
in the majority of cases, the researchers got the same result as in the original study. It may be due to a number of reasons. Often made psychological experiments in small groups, in which chance plays a pretty big role. Researchers have a tendency to publish studies showing positive results and ignore the trials where the effect had been expected did not occur. You can often find excuses to ignore some different results. Pure cheats can also occur.
There is also the entire scientific areas which largely works with interviews and observations, in a way that makes it almost never seen scientists, consciously or unconsciously, gives a distorted picture of the results.
the same kind of errors seems to arise again and again, but we really learn something from it. Sometimes, as with the Riksbank’s forecasts, it can have serious consequences because it affects investment decisions and the like. In other cases, the more subtle effects on the world we are building.
In some areas, it is going to improve the situation; the progress probably some soul-searching among psychology researchers. Other things we may not know anything about, at least not with the methods we have today. Riksbank should not give forecasts beyond their reasonable accuracy for perhaps a month or two ahead.
A reasonable thought is that the state should not fund research and forecasts that most give uncontrollable or misleading results. But we can not wait for us to researchers and forecasters within a discipline to lead the purges, they have everything to lose by saying that what they have been doing is not working. And those who are not experts are of course difficult to safely determine what is worthless and what is just clutter.
Perhaps State and some interested private parties assemble a nonsense Commission, which evaluates the various fields in the dialog with the experts. The showing stable results gets the green light, those who should be able to do it with a little tougher scrutiny is given a yellow light and those who work with methods that never seems to give reliable knowledge or meaningful forecasts a red light. Then universities, government, media, banks and draw their conclusions from it.
Human beings are not infallible and there is no limit to his ability to fool themselves and others. But we might at least be able to avoid having to see the same stupid things over and over again?
No comments:
Post a Comment