Slow & Nonexistent Scientific Self-Correction in Psychology

Self-correction in psychology proceeds at a snail's pace, or not at all...

Posted May 15, 2015

Self-correction is one of the key characteristics of science that differentiates it from other ways of attempting to understand the world. Religious tomes are the infallible word of (the) God(s).  There is, however, nothing inherently infallible about scientific claims.  The world is not filled with "essences" just because Aristotle said so.  The Sun does not revolve around the Earth just because it looks that way.

ALISON
Source: ALISON

So how is psychology doing?  Self-correction is different from "finding stuff out." Psychology has found out quite a lot about people that is true, important, and has held up well.

Self-correction only becomes an issue when science gets something wrong. Unfortunately, correction in psychology is often so slow it is hard to detect its very occurrence. (Thus my comparison to plate tectonics—the Earth's continents move unbelievably slowly, undetectable to the naked eye, but move they do, sometimes leading to things such as earthquakes).

One can find information on Psychology's dismal record of self-correction here:

http://bps.stanford.edu/?page_id=2367

(note that there are links at the top of the page for different types of self-correction, includijng failed replications, retractions, and alternative explanations).

It is a mostly sordid record of:

1. Famous Published Paper

followed by

2. One or more corrective publications that get mostly ignored.

You can see for yourself why this is a sorry state of affairs.  Go to Google Scholar:

https://scholar.google.com/schhp?hl=en

And plug in any famous entry under "Original Article."  We want only entries that have been highly influential  because these are the most important to correct if wrong.  So, look for original articles that have at least 500 citations.

The first such original article with over 500 citations is:

Bargh, J.A., Chen, M., & Burrows, L. (1996).  Automaticity of social behavior: Direct effects of trait construct and stereotype activation on action.  Journal of Personality and Social Psychology, 71, 230-244.

It has been cited over 3500 times.  

The three complete or partial failures to replicate have been cited, respectively:

99, 203, and, 174 times.

The last paper, by Doyen et al, was the most problematic for the original, and it was published in 2012.  So, how many times has the original Bargh et al paper been cited, say, since 2013**?  761 times!!! more than all three failures put together!

** In Google Scholar, look at the left panel.  About halfway down.  You can get the number of citations in any given date range.  Under "custom range" I just entered 2013-2015.

Lee Jussim
Source: Lee Jussim

So much for self-correction...  

Try this yourself, say, for other famous, influential papers on the page such as Darley & Gross, 1983, Steele & Aronson, 1995, or any of the others.

Is psychology a self-correcting science?  You tell me.  Do you know of any scientific research in psychology that:

1. Was published

2. Is highly cited and widely believed

3. Has been refuted by failed replications, retractions, or alternative interpretations

and

4. Continues to be cited as if the original finding is the truth, the whole truth, and nothing but the truth.

Examples and comments welcome.