Tags

, , , ,

In his recent column in Nature Jelte M. Wicherts considers the circumstances of the “disastrous case of prolonged misconduct” by Diederik Stapel, a scientist from netherlands.
Stapel had become famous for his research on “aspects of social behaviour such as power and stereo­typing”. At least 30 of his papers in peer-reviewed journals are fake (read here).
For Wicherts one major problem seems to be that, especially in the field of psychological research, sharing “research data with colleagues, even co-authors (…) is no common practice.” Most researchers in psychological research “simply fail to document their data in a way that allows others to quickly and easily check their work. It is not unusual for data that are shared to list variables only as VAR00001 through VAR00019, with no further explanation.”
The article is very “straight forward” and worth reading, if you have the privileged access to the Nature-Journal. ;( 
I’ve been working for several years in the field of electrophysiological research, where measured brain potentials in the µV-range are knitted with complex perceptual phenomena, like gestalt-perception. The will to measure such constructs/processes leads to fragile designs, but produces a massive amount of data, with high variability, which make them more look like a noisy epiphenomenon of the experimental session than like being caused by it.
The art of developing scientific articles with these data is excellently supported by the normality to take the data, chase them through your “hope so”- algorithms until “hope so” turns out to “be so”. In your article you explain how the algorithms work in principle, but you don’t append them. Now you only need to produce a pdf-format and your data are sufficiently separated from any data base.
And your data? Lock them up in your institutional safe, for safety and legal reasons only.
Lit.: Stapel, D, Psychology must learn a lesson from fraud case, Nature, 480, 7 (01 December 2011), doi:10.1038/480007a
Advertisements