SCIENCE TAKES A DIVE

likely the worst brain imaging study ever published, researchers interested in football penalty-taking behaviour discarded 86% of their data, did 165 statistical tests, then mis-interpreted the results. in ten thousand words  

original article: Slutter et al., 2021 (Frontiers in Computer Science), reported in: The Daily Mail by Joe Pinkstone on 7th May 2021

this story was in episode 11 #penalty #choke #football #FNIRS #QRP


the error bar says

ever-obsessed with the failure of English football players to score during penalty shoot-outs, the Daily Mail reports a Dutch study of 22 football players taking penalties

the researchers attempted what many scientists would think impossible - recording high-quality brain imaging data while people are running around & kicking balls

finding that the motor cortex is involved in playing & the prefrontal cortex in thinking about football, the authors noted it is "feasible... to detect neurological clues relevant to anxiety and performance in the field"


are these the neural correlates of football?

no

there's good reason no-one has attempted this research before - it's not possible

as exemplified by our previous story, brain imaging is a difficult, labour-intensive technique that involves training, patience & competence. you can't just stick a scanner on someone's sweaty forehead & send them out to play games.

well, you can, but the result would be dribble.

to achieve the unachievable, the authors first removed one of their 22 participant's data for technical failure. this is quite common.

then they ignored 2 of the 10 scanned parts of the brain because the data were noisy. OK.

'outliers' were then removed. OK, fairly typical.

then one of the two sets of data signals were removed for being, you know, not useful.

& 59% of the remaining data was removed for being poor quality.

in total, the researchers removed 86% of their data. what little data remains for analysis looks strange.

the analysis itself was 165 different statistical tests, under three analytical approaches with two different ways of correcting for all of these comparisons.

this work is ground-breaking in that we just don't have words to describe how bad this study is. there are so many red flags that the authors, participants, reviewers & editors should all have been sent off long before any words were sacrificed on these seventeen pages.

i mean what's the [BLEEP] point?

the only redeeming feature in this 10000-word dump of an article is that the authors followed the "4-4-2" method of placing the brain scan devices on the head. i hope that does something to ease the pain for our listeners.


conclusion

balls

rating

The Daily Mail: fudge - scientific story distorted