What information can you trust?

The world of scientific research is competitive: grant money and livelihoods depend on getting research published and journals select articles that make splashy news. At the moment, a greater number of post doctorates are competing for fewer jobs and grant resources. What this has led to, in the words of one researcher, is an “epidemic of fraud.” The “fraud” in this case consists of such deceptions as leaving out contradictory data, cutting corners, or even making up data.

Example: You may have heard of “mindless eating,” the idea that the unconscious decisions we make about food can have profound effects on our diet and weight. The man responsible is Brian Wansink, a Cornell University professor, who became quite well known for his experiments such as the one showing that serving bowl size affects food consumption. He has influenced consumer behavior and even national policy. But he was forced to resign after thirteen of his papers were retracted because of data irregularities and gross statistical errors. In the words of Cornell, he committed “academic misconduct in his research and scholarship, including misreporting research data.”

In the field of psychology, one researcher started a “reproducibility project:” he selected one hundred published psychological experiments and tried to reproduce the findings but was only able to reproduce fewer than half of them. In the field of genomics, Stanford’s John Ioannidis (very trustworthy) found that only a tiny fraction of papers on that topic stood the test of time. For more examples, you can check out the blog, Retraction Watch, which strives to post every single academic retraction and keeps a top ten list of the most highly cited retracted papers. When I looked at it, it had already listed thirty-three retracted papers on the coronavirus.

Now, for myself, I select only those studies that support my own biases, which are always correct, of course.

For an introduction to this blog, see I Just Say No; for a list of blog topics, click the Topics tab.