Science is a messy, error fraught business, which is why reproducibility is so essential. Unfortunately, that doesn't appear to be one of psychology's strong suits, according to a massive analysis published yesterday in Science.
A years-long effort to reproduce more than 100 psychology studies across three leading journals paints a pretty dismal picture. When re-tested by independent research psychologists, the conclusions of more than 60 studies on personality, relationships, learning, and memory, turned out to be far less whelming. Strongly significant findings often became weaker, while weakly significant findings became non-existent.
http://gizmodo.com/a-lot-of-published-psychology-results-are-bullshit-1727228060
[Source]: The New York Times
(Score: 0) by Anonymous Coward on Sunday August 30 2015, @08:03AM
Houston, we have a problem!
Notice how he said "80%" and not all. Not sure where that number came from, but still.
Furthermore, generally, the more complex whatever you're studying is, the more difficult it is to study it properly. Reproducing the studies to see if you get the same result is more simple than doing original and good science in the first place, as there is less that can go wrong. That is why this is more believable than the types of studies that reach arbitrary and subjective conclusions about how violent video games make people more violent, how X makes people "callous" towards Y, etc.
(Score: 4, Informative) by Phoenix666 on Sunday August 30 2015, @11:38AM
As someone trained in the social sciences, I have on occasion thought that physical scientists rather take the easy way out in studying natural phenomena that can be pinned down, controlled, and whose behaviors can be reproduced. People are devilishly difficult to study, because they are complex and moving all the time. They're conscious, and consciously or unconsciously mess with your study. As such it's much more complex, much more multi-variate a system than what, say, chemists tackle.
Washington DC delenda est.
(Score: 2) by c0lo on Sunday August 30 2015, @12:58PM
If the "social scientists" would deal with populations on the order of 1023, I reckon the social sciences would be as accurate as chemistry (which is not that accurate as many people think, those equilibrium points for the reversible reactions are correct just statistically and, anyway, empirically determined).
https://www.youtube.com/watch?v=aoFiw2jMy-0 https://soylentnews.org/~MichaelDavidCrawford
(Score: 2) by maxwell demon on Sunday August 30 2015, @02:59PM
Actually I've once heard that some chemical reactions won't work if the chemist cleans his equipment too thoroughly.
The Tao of math: The numbers you can count are not the real numbers.
(Score: 0) by Anonymous Coward on Sunday August 30 2015, @01:39PM
I think Dr. Fancyfree [wikia.com] says it best regarding this entire field: I'm going to do something that goes against all my training as a therapist... I'm going to CURE you!!!
(Score: 0) by Anonymous Coward on Sunday August 30 2015, @04:06PM
It's complex, but that is no reason to lower our standards. Bad science is bad science. I'm not saying you were necessarily saying that, though.
(Score: 2) by Nerdfest on Sunday August 30 2015, @05:17PM
Hah. Try quantum physics.
(Score: 2) by Phoenix666 on Sunday August 30 2015, @05:25PM
Touche, nerdfest.
Washington DC delenda est.
(Score: 2) by hendrikboom on Tuesday September 01 2015, @01:38PM
The theoretical calculations of various properties of fundamental particles have often been orders of magnitude more accurate than the observations the theories were originally based on, as evidenced by later, more accurate observations.
No one knows why.
(Score: 0) by Anonymous Coward on Sunday August 30 2015, @06:29PM
It is not specific to social sciences. [washingtonpost.com] There is tons of pressure to bias results in the hard sciences too.
IIRC there was a recent story about how requiring scientists to formally specify their testing regimen up front, before starting the experiment, drastically cut down the 'success rate.' The theory being that the urge to comb through the data after the fact, looking for novel results was just too strong.
(Score: 2) by c0lo on Sunday August 30 2015, @12:47PM
He's trying to align more with the classics than with Sturgeon [wikipedia.org].
From the same page and as a word of caution for those disgusted by 80% crap of everything:
https://www.youtube.com/watch?v=aoFiw2jMy-0 https://soylentnews.org/~MichaelDavidCrawford