In many fields of research right now, scientists collect data until they see a pattern that appears statistically significant, and then they use that tightly selected data to publish a paper. Critics have come to call this p-hacking, and the practice uses a quiver of little methodological tricks that can inflate the statistical significance of a finding. As enumerated by one research group, the tricks can include:
- “conducting analyses midway through experiments to decide whether to continue collecting data,”
- “recording many response variables and deciding which to report postanalysis,”
- “deciding whether to include or drop outliers postanalyses,”
- “excluding, combining, or splitting treatment groups postanalysis,”
- “including or excluding covariates postanalysis,”
- “and stopping data exploration if an analysis yields a significant p-value.”
Add it all up, and you have a significant problem in the way our society produces knowledge.
A number of studies have spotted a worrisome trend: although the number of scientific journals and articles published is increasing each year, the rate of papers being retracted as invalid is increasing even faster. Some of these are being retracted due to obvious ethical lapses—fraudulent data or plagiarism—but some past studies have suggested errors and technical problems were the cause of the majority of problems.
A new analysis, released by PNAS, shows this rosy picture probably isn’t true. Researchers like to portray their retractions as being the result of errors, but a lot of these same papers turn out to be fraudulent when fully investigated. If there’s any good news here, it’s that a limited number of labs 38, to be exact are responsible for a third of the fraudulent papers that end up being retracted.
Right now, if you want to read the published results of the biomedical research that your own tax dollars paid for, all you have to do is visit the digital archive of the National Institutes of Health. There you’ll find thousands of articles on the latest discoveries in medicine and disease, all free of charge.
A new bill in Congress wants to make you pay for that, thank you very much. The Research Works Act would prohibit the NIH from requiring scientists to submit their articles to the online database. Taxpayers would have to shell out $15 to $35 to get behind a publisher’s paid site to read the full research results.
Here’s an idea: See if your English instructor or your Speech / Communications instructor would allow you to dance in order to explain your research.
You see the American Association for the Advancement of Science has put on the third annual Dance Your Ph.D. competition wherein a Ph.D thesis is explained with a dance routine.
Here is one of the finalists. Enjoy.
(found via Boing Boing)
Welcome and welcome back! The Fall semester is underway–classes, tests, papers, research. Tired just thinking about it? Let the library help! We have study guides and books to help you with every class: Math, English, Science, Nursing…the list goes on and on. Don’t let yourself get behind. Let the library get you started on the right foot!