« NRA convention sets new record | Main | Mass killing in Australia »
A growing problem...
especially when doctors study crime (gun crime), funded by grants from the antigun CDC:
"Many studies' results cannot be reproduced, scholars warn." They found that, of 53 medical studies, only 6 had reproducible results. "Not all irreproducible research is progressive advocacy; not all progressive advocacy is irreproducible; but the intersection between the two is very large. The intersection between the two is a map of much that is wrong with modern science."
From the executive summary: "Many common forms of improper scientic practice contribute to the crisis of reproducibility. Some researchers look for correlations until they find a spurious "statistically signicant" relationship. Many more have a poor understanding of statistical methodology, and thus routinely employ statistics improperly in their research. Researchers may consciously or unconsciously bias their data to produce desired outcomes, or combine data sets in such a way as to invalidate their conclusions. Researchers able to choose between multiple measures of a variable often decide to use the one which provides a statistically signicant result. Apparently legitimate procedures all too easily drift across a fuzzy line into illegitimate manipulations of research techniques."
Hat tip to Joe Olson.
4 Comments | Leave a comment
My Dad once told me there is an old saying:
"Figures don't lie, but liars figure".
Interesting that results of the Kleck study on defensive gun use were very similar to the CDC study on the same subject.
Kleck was (and still is) ridiculed for his results - the CDC merely suppressed theirs.
Another factor in this desperate search for anything that looks "statistically significant" is the academic environment in which many researchers live and breathe. In order to advance or get tenure, one needs a record of publications in professional journals. Journals will hardly ever accept a paper that doesn't show "statistically significant" results, so researchers continue to torture their data by transformations and other methods until they get those results. They then develop a story that coincides with the analyses, and (as long as it doesn't offend the sensibilities of their colleagues or the journal editors) submit it for publication. Anything that doesn't have that veneer of statistical rigor or that provokes opposition from the establishment is simply shelved.
Having spent more than 40 yrs in academia, I can vouch for the falling intelligence and growing ignorance of the professoriate and the students. Since many of today's faculty have gone through their educational years after AA and EO laws went into effect, the dumbing down of education at the post secondary level has left us with the ignorati running things.
Many folks believe in "peer" review which is another BS practice since no true review occurs. Some group of professors read the papers of other professors and grant or deny publication. No testing. No proper evaluation. Often with "I know him/her. He/She's a great researcher."
I've watched idiot faculty using testing methods based on standard chemical methods complain that the standard method didn't work without understanding that the automated method was based on the standard method.
I've watched professors change their statistical processing after the experiment is complete.
But this is nothing new. The original 1954-6 research telling the world that sat fats were bad was falsified/selectively done. So for 60+ yrs, diets have been based on a lie.
"Nothing new under the sun" as the wisest man to ever live once said.