3 ms·

Someone's anecdotal experience and laymen theorizing is next to useless compared to data. Likewise for cherry picking studies out of a large body of literature

Someone's anecdotal experience and laymen theorizing is next to useless compared to data. Likewise for cherry picking studies out of a large body of literature. > This strict version has since been rejected for second-language acquisition, as some adult learners have been observed who reach native-like levels of pronunciation and general fluency. Huh? This doesn't make the research any less conclusive, or support the GP's point. It's like saying the existence of players like Muggsy Bogues makes the thesis "height is extremely important for profession baseketball" less conclusive.

2109302 10y ago
You don't cite any data, you are just waving a degree that anyone can obtain. But the "unruly hackers" here are always on the side of the status quo ...
coldtea 10y ago
>Someone's anecdotal experience and laymen theorizing is next to useless compared to data. Data are next to useless without proper interpretation of what they mean. And even before that, we must assume that the data are correctly collected and relevant. And something like 80% of research papers have been found in meta-studies to be non reproducible, to the extend that scientists speak of a "reproducibility crisis". So, unless we're talking about hard core sciences, like physics and the like, where things are testable with actual experimental apparatus and hard numbers (and even there, I wouldn't go into speculative stuff like String theory), I'd take any society/psychology/physiology related study with lots of grains of salt. Up until the 80s it was "scientific fact" that being gay was a abnormal mental condition -- supported by scientific papers and literature. Or take the dietary (the official scientific dietary advice) over various decades, and how it made a u-turn regarding fat, etc. Computer Scientists that are driven to science should not take into much account pseudo-sciences and half-sciences which much of cognitive studies, psychology, sociology etc are.

2109302 10y ago

You don't cite any data, you are just waving a degree that anyone can obtain. But the "unruly hackers" here are always on the side of the status quo ...

coldtea 10y ago

>Someone's anecdotal experience and laymen theorizing is next to useless compared to data. Data are next to useless without proper interpretation of what they mean. And even before that, we must assume that the data are correctly collected and relevant. And something like 80% of research papers have been found in meta-studies to be non reproducible, to the extend that scientists speak of a "reproducibility crisis". So, unless we're talking about hard core sciences, like physics and the like, where things are testable with actual experimental apparatus and hard numbers (and even there, I wouldn't go into speculative stuff like String theory), I'd take any society/psychology/physiology related study with lots of grains of salt. Up until the 80s it was "scientific fact" that being gay was a abnormal mental condition -- supported by scientific papers and literature. Or take the dietary (the official scientific dietary advice) over various decades, and how it made a u-turn regarding fat, etc. Computer Scientists that are driven to science should not take into much account pseudo-sciences and half-sciences which much of cognitive studies, psychology, sociology etc are.