Science

n2doc

(47,953 posts) Fri Feb 27, 2015, 12:27 PM Feb 2015

Why one science journal wants to publish failed studies

If science were a TV show, it would be a lot more like Modern Family than Breaking Bad: the promising and positive results from studies are usually on display, not the dark, ambiguous underbelly of failed and inconclusive research.

This is because research suffers from what's known as "publication bias": not all studies that are conducted actually get published in journals, and the ones that do tend to have positive (i.e. statistically significant) findings. This is a big problem because it means we have a biased picture of what's going on in science, and many researchers waste their time and funding repeating work that's already been done. The issue is so severe right now that some have wondered whether negative results are disappearing entirely from some countries and fields of science.

Now, one journal is trying to correct publication bias: PLoS One this week launched its "Missing Pieces" collection of negative, null, and inconclusive studies — in other words, a celebration of the seamier side of botched and boring experiments that usually never sees the light of day.

Articles in the collection include one that failed to find that women's support groups had a significant impact on postpartum disorder in Bangladeshi mothers (despite promising findings in similar research in India) and a study that could not replicate four previous experiments on the "depletion model" of self control, an increasingly popular idea that posits that self control is a limited resource that runs out in people.

more

http://www.vox.com/2015/2/27/8119957/publication-bias

2 replies

= new reply since forum marked as read

Highlight:

Why one science journal wants to publish failed studies (Original Post) n2doc Feb 2015 OP

Much of the pharmaceutical industry depends upon publication bias. hunter Feb 2015 #1

There is some strange stuff out there jakeXT Feb 2015 #2

hunter

(38,326 posts)

1. Much of the pharmaceutical industry depends upon publication bias.

Reply to n2doc (Original post)

Fri Feb 27, 2015, 01:20 PM

Feb 2015

They don't want to hear that their expensive new medicines are ineffective or no better than inexpensive alternatives, going so far as to make it nearly impossible to study cannabis, mild opiates, and certain other classes of drugs that might prove far superior to the products they are selling.

jakeXT

(10,575 posts)

2. There is some strange stuff out there

Reply to hunter (Reply #1)

Fri Feb 27, 2015, 02:41 PM

Feb 2015

The problems that weren't reported were sometimes egregious. One paper, for example, said all patients reported improvement, but in fact, the FDA found that one patient had a foot amputated two weeks after receiving the treatment.

In another case, the entire clinical trial was considered unreliable by the FDA - but the published paper didn't mention that.

In another, researchers falsified data, which led to one patient’s death.

Data on these violations are not readily available. So it's impossible to say how often tainted data are published and how often the violations are noted, Seife said.

http://www.reuters.com/article/2015/02/09/us-trial-violations-idUSKBN0LD25B20150209

Reply to this discussion