This topic has been archived, and won't accept reply postings.
not surprised that this damning indictment of science hasn't been posted..
http://www.bbc.co.uk/programmes/b04f9r4k
http://www.bbc.co.uk/programmes/b04f9r4k
In reply to malk:
It's true. It's all nonsense, and lies. For example, the science that they say underpins mobile telecommunications and portable microprocessors is all bunkum and so the device I'm typing this on doesn't actually exi
It's true. It's all nonsense, and lies. For example, the science that they say underpins mobile telecommunications and portable microprocessors is all bunkum and so the device I'm typing this on doesn't actually exi
In reply to malk:
Bit of an part of the forum to post it up in really.
Anyway, it's nothing really new. If you have even had a short passing experience of academia then this is hardly surprising. I'm not sure what you're trying to say.
I guess you're someone who has a thing against 'science'?
Bit of an part of the forum to post it up in really.
Anyway, it's nothing really new. If you have even had a short passing experience of academia then this is hardly surprising. I'm not sure what you're trying to say.
I guess you're someone who has a thing against 'science'?
In reply to malk:
Does this mean that Dr John Ioannidis findings will also become less true over time...AAAAaarrggghh..
<disappears into infinite loop....>
Does this mean that Dr John Ioannidis findings will also become less true over time...AAAAaarrggghh..
<disappears into infinite loop....>
In reply to malk:
Yes old understandings are replaces with new better understandings. Thats what science is. Looking for the least wrong thing. Or do you think that all science is some kind of lie and actually its magic all the way down?
Yes old understandings are replaces with new better understandings. Thats what science is. Looking for the least wrong thing. Or do you think that all science is some kind of lie and actually its magic all the way down?
In reply to malk:
Probably because it's not a damning indictment of science. Though as usual it is a damning indictment of irresponsible science journalism.
http://www.economist.com/news/science-and-technology/21598944-sloppy-resear...
http://scienceblogs.com/insolence/2007/09/24/the-cranks-pile-on-john-ioanni...
> not surprised that this damning indictment of science hasn't been posted..
Probably because it's not a damning indictment of science. Though as usual it is a damning indictment of irresponsible science journalism.
http://www.economist.com/news/science-and-technology/21598944-sloppy-resear...
http://scienceblogs.com/insolence/2007/09/24/the-cranks-pile-on-john-ioanni...
In reply to off-duty:
improper research etc-listen to the link
improper research etc-listen to the link
In reply to malk:
I'd rather read the source paper:-
http://www.plosmedicine.org/article/info%3Adoi%2F10.1371%2Fjournal.pmed.002...
> improper research etc-listen to the link
I'd rather read the source paper:-
http://www.plosmedicine.org/article/info%3Adoi%2F10.1371%2Fjournal.pmed.002...
In reply to Ste Brom:
Are we human or are we dancer?
My sign is vital, my hands are cold
And I'm on my knees looking for the answer
Are we human or are we dancer
I think it's actually about a reindeer, Rudolph's colleague...
> Or are they dancers?
Are we human or are we dancer?
My sign is vital, my hands are cold
And I'm on my knees looking for the answer
Are we human or are we dancer
I think it's actually about a reindeer, Rudolph's colleague...
In reply to malk:
Why aren't you surprised?
> not surprised that this damning indictment of science hasn't been posted..
Why aren't you surprised?
In reply to malk:
Based on its publication date of 2005 presumably "Why Most Published Research Findings Are False" has since fallen victim to its own prognosis and has been proved false?
Based on its publication date of 2005 presumably "Why Most Published Research Findings Are False" has since fallen victim to its own prognosis and has been proved false?
In reply to mattrm:
That's a very black and white conclusion to draw, especially for someone with experience of academia.
This research might well be nothing new, but it does suggest that a bit of modesty might be appropriate among scientists in an age where phrases like "the science says…" or "the science is settled…" are deemed to carry such weight in political debates.
> I guess you're someone who has a thing against 'science'?
That's a very black and white conclusion to draw, especially for someone with experience of academia.
This research might well be nothing new, but it does suggest that a bit of modesty might be appropriate among scientists in an age where phrases like "the science says…" or "the science is settled…" are deemed to carry such weight in political debates.
In reply to Mike Stretford:
there are a lot of vocal supporters of everything science on ukc..
those that express doubt are swiftly dispatched
> Why aren't you surprised?
there are a lot of vocal supporters of everything science on ukc..
those that express doubt are swiftly dispatched
In reply to malk:
Hmm. Or you could read the variety of links I have posted above.
The only certain thing about science is it's uncertainty.
There does appear to be a degree of confusion between the "scientific method" and "science" itself. Personally I blame those liberal arts degrees
> there are a lot of vocal supporters of everything science on ukc..
> those that express doubt are swiftly dispatched
Hmm. Or you could read the variety of links I have posted above.
The only certain thing about science is it's uncertainty.
There does appear to be a degree of confusion between the "scientific method" and "science" itself. Personally I blame those liberal arts degrees
In reply to malk:
That Newton chap was a famous scientist and got it all wrong, didn't he? Later corrected by that Einstein chap, whose corrections themselves were all wrong. How are we to believe anything science tells us when even our most famous scientists got it so badly wrong?
Alternatively ... well-conducted science postulates and demonstrates hypothesised behaviours. Many times the behaviours shown are a good fit for the model proposed and the science proves useful in describing the physical world. Sometimes the science isn't well-conducted and sometimes the results may be misleading outliers. Often the model proposed and 'demonstrated' is later refined to take into account more subtle influences. Sometimes the science is conducted by charlatans.
Eventually, given enough attention and scientific rigour, such issues have been and will continue to be resolved and scientific understanding will progress further. Apart from deliberately misleading influences such as big pharma and scientology I fail to see the problem.
That Newton chap was a famous scientist and got it all wrong, didn't he? Later corrected by that Einstein chap, whose corrections themselves were all wrong. How are we to believe anything science tells us when even our most famous scientists got it so badly wrong?
Alternatively ... well-conducted science postulates and demonstrates hypothesised behaviours. Many times the behaviours shown are a good fit for the model proposed and the science proves useful in describing the physical world. Sometimes the science isn't well-conducted and sometimes the results may be misleading outliers. Often the model proposed and 'demonstrated' is later refined to take into account more subtle influences. Sometimes the science is conducted by charlatans.
Eventually, given enough attention and scientific rigour, such issues have been and will continue to be resolved and scientific understanding will progress further. Apart from deliberately misleading influences such as big pharma and scientology I fail to see the problem.
In reply to john arran:
Can't disagree with any of that. Although you missed out the fact that plenty of good, well-conducted science also gets refuted. And I'm not sure Newton is the best example, because his work had few political implications. I'd wager that the more politically relevant the research, the more prone it is to error. Ioannidis touches on this. That programme that Malk links to is well worth a listen.
Can't disagree with any of that. Although you missed out the fact that plenty of good, well-conducted science also gets refuted. And I'm not sure Newton is the best example, because his work had few political implications. I'd wager that the more politically relevant the research, the more prone it is to error. Ioannidis touches on this. That programme that Malk links to is well worth a listen.
In reply to stu7jokes:
That would mean it either wasn't well-conducted, its observed result was an outlier or the refutal was unfounded.
The study of science isn't immune to human frailty and never has been. Doubtless people will continue to misrepresent science, the scientific method and scientifically obtained results for personal, idealogical or corporate gain for a very long time to come. The science however will remain unchanged.
> Can't disagree with any of that. Although you missed out the fact that plenty of good, well-conducted science also gets refuted.
That would mean it either wasn't well-conducted, its observed result was an outlier or the refutal was unfounded.
> And I'm not sure Newton is the best example, because his work had few political implications. I'd wager that the more politically relevant the research, the more prone it is to error. Ioannidis touches on this. That programme that Malk links to is well worth a listen.
The study of science isn't immune to human frailty and never has been. Doubtless people will continue to misrepresent science, the scientific method and scientifically obtained results for personal, idealogical or corporate gain for a very long time to come. The science however will remain unchanged.
In reply to john arran:
Not necessarily an outlier, because most research is not replicated, so there's no body of work of work from which it is possible to be an outlier. It is also a simple product of the arbitrary thresholds science uses to determine statistical significance.
Of course. But your concluding sentence betrays a confusion between the material world and the discipline of science. It is only the former that remains unchanged. My concern is that it is the scientific establishment, not the liberal arts degrees, that misrepresent science. Which is to be expected in a world that grants it so much political power, but is not helpful. All of which is not to knock science, which is the best method we have for understanding the material universe. But it's worth recognising that most of what science discovers turns out not to be true. Have you listened to the programme?
> That would mean it either wasn't well-conducted, its observed result was an outlier or the refutal was unfounded.
Not necessarily an outlier, because most research is not replicated, so there's no body of work of work from which it is possible to be an outlier. It is also a simple product of the arbitrary thresholds science uses to determine statistical significance.
> The study of science isn't immune to human frailty and never has been. Doubtless people will continue to misrepresent science, the scientific method and scientifically obtained results for personal, idealogical or corporate gain for a very long time to come. The science however will remain unchanged.
Of course. But your concluding sentence betrays a confusion between the material world and the discipline of science. It is only the former that remains unchanged. My concern is that it is the scientific establishment, not the liberal arts degrees, that misrepresent science. Which is to be expected in a world that grants it so much political power, but is not helpful. All of which is not to knock science, which is the best method we have for understanding the material universe. But it's worth recognising that most of what science discovers turns out not to be true. Have you listened to the programme?
In reply to stu7jokes:
I haven't had time to listen to the programme but it's quite undeniable that if science 'discovers' something that turns out not to be true then the scientific study leading to that discovery was at best questionable and most likely flawed. By the way statistical significance thresholds are far from arbitrary, but as with any sampling there remains a remote possibility of freak results (which would become outliers were the study to be replicated repeatedly).
I haven't had time to listen to the programme but it's quite undeniable that if science 'discovers' something that turns out not to be true then the scientific study leading to that discovery was at best questionable and most likely flawed. By the way statistical significance thresholds are far from arbitrary, but as with any sampling there remains a remote possibility of freak results (which would become outliers were the study to be replicated repeatedly).
In reply to john arran:
OK, I'm happy with that. So let's rephrase it: most scientific findings are at best questionable and most likely flawed. The problem remains.
Statistical significance is most certainly arbitrary. The probability of a result occurring by chance must be lower than 5% for significance. Why not 6 or 7%? It is merely convention. Again, nothing wrong with that in itself, but an inevitable consequence is that 1 in 20 statistically significant effects are not true. That's not a remote possibility. I suspect it's more than 1/20, though, because, as you say, science is subject to the same human frailties of any other endeavour. So, for example, published papers will tend to go with the statistical test that supplied the statistically significant result rather than those that turned up less interesting results - naughty but effective.
> I haven't had time to listen to the programme but it's quite undeniable that if science 'discovers' something that turns out not to be true then the scientific study leading to that discovery was at best questionable and most likely flawed.
OK, I'm happy with that. So let's rephrase it: most scientific findings are at best questionable and most likely flawed. The problem remains.
> By the way statistical significance thresholds are far from arbitrary, but as with any sampling there remains a remote possibility of freak results (which would become outliers were the study to be replicated repeatedly).
Statistical significance is most certainly arbitrary. The probability of a result occurring by chance must be lower than 5% for significance. Why not 6 or 7%? It is merely convention. Again, nothing wrong with that in itself, but an inevitable consequence is that 1 in 20 statistically significant effects are not true. That's not a remote possibility. I suspect it's more than 1/20, though, because, as you say, science is subject to the same human frailties of any other endeavour. So, for example, published papers will tend to go with the statistical test that supplied the statistically significant result rather than those that turned up less interesting results - naughty but effective.
Post edited at 21:50
In reply to stu7jokes:
Not sure we're disagreeing on much except perhaps semantics. Statistical thresholds are carefully chosen to reflect a precise quantity of uncertainty which tallies with people's concept of reassuringly likely. 95% is probably most common but 99% is also often quoted. But even at 95% that's still a small part of the story when it comes to explaining this reported/alleged incidence of false results. Most of the rest will be bad science, however well intentioned.
Not sure we're disagreeing on much except perhaps semantics. Statistical thresholds are carefully chosen to reflect a precise quantity of uncertainty which tallies with people's concept of reassuringly likely. 95% is probably most common but 99% is also often quoted. But even at 95% that's still a small part of the story when it comes to explaining this reported/alleged incidence of false results. Most of the rest will be bad science, however well intentioned.
In reply to john arran:
It's detail, not semantics. The point is that most research findings are not true, for a variety of reasons. And that is well worth bearing in mind in an age of evidence-based policy making. It might be obvious to academics, but I don't think it's obvious to the electorate. And when science is widely regarded as the arbiter in so many policy debates, there is little incentive for the scientific establishment to clarify things. Sorry to bang on about it, but that programme is very interesting.
It's detail, not semantics. The point is that most research findings are not true, for a variety of reasons. And that is well worth bearing in mind in an age of evidence-based policy making. It might be obvious to academics, but I don't think it's obvious to the electorate. And when science is widely regarded as the arbiter in so many policy debates, there is little incentive for the scientific establishment to clarify things. Sorry to bang on about it, but that programme is very interesting.
In reply to malk:
Interesting programme.
Those who see this as an indictment of science are, however, completely wrong. The programme gives a series of examples of pieces of research based on essentially unexplained correlations: since the human body and mind are too complex to understand mechanistically we rely on correlations.
This makes medical research quite slippery: a feature exploited by the pharmaceutical industry of course.
But medical research looking for correlations is not synonymous with science. Most branches of science make little use of unexplained correlations, they look for models and explanations for specifically chosen and isolated phenomena.
Our trust in medicine can and should be called into question: it's mostly based on correlation, and subject to all the potential bias explained well in the programme. But it's a massive crime of science misreporting to then pile the whole of science into this correlation basket and claim that therefore it's all false.
Our understanding of the laws of nature is not false, our understanding of how biology works is not false, our understanding that underpins all of the engineering and technology is not false. Basic science is not based on unexplained correlations. However, our understanding of which drug works best for depression, what the risk factors are for any cancer you might name, how genetics and environment interact to influence personality, etc, are all deeply flawed just as the programme explains.
Interesting programme.
Those who see this as an indictment of science are, however, completely wrong. The programme gives a series of examples of pieces of research based on essentially unexplained correlations: since the human body and mind are too complex to understand mechanistically we rely on correlations.
This makes medical research quite slippery: a feature exploited by the pharmaceutical industry of course.
But medical research looking for correlations is not synonymous with science. Most branches of science make little use of unexplained correlations, they look for models and explanations for specifically chosen and isolated phenomena.
Our trust in medicine can and should be called into question: it's mostly based on correlation, and subject to all the potential bias explained well in the programme. But it's a massive crime of science misreporting to then pile the whole of science into this correlation basket and claim that therefore it's all false.
Our understanding of the laws of nature is not false, our understanding of how biology works is not false, our understanding that underpins all of the engineering and technology is not false. Basic science is not based on unexplained correlations. However, our understanding of which drug works best for depression, what the risk factors are for any cancer you might name, how genetics and environment interact to influence personality, etc, are all deeply flawed just as the programme explains.
Post edited at 22:37
In reply to Jon Stewart:
Totally agree.
I am back on the wine after being told that more than two glasses a year or whatever will lead to DEATH
Totally agree.
I am back on the wine after being told that more than two glasses a year or whatever will lead to DEATH
In reply to stu7jokes:
It's not that they aren't factually 'true' (in general) but they aren't complete and, of course, are highly likely to be overinterpreted in support of current theory.
But compared to just about anything else that's probably OK. Don't let's beat ourselves up about it too much.
> (In reply to john arran)
>
> It's detail, not semantics. The point is that most research findings are not true, for a variety of reasons.
>
> It's detail, not semantics. The point is that most research findings are not true, for a variety of reasons.
It's not that they aren't factually 'true' (in general) but they aren't complete and, of course, are highly likely to be overinterpreted in support of current theory.
But compared to just about anything else that's probably OK. Don't let's beat ourselves up about it too much.
In reply to malk:
Sensationalist headline there - "Everything we know is wrong.". To the point of being a lie no less.
Papers come and papers go, they come to all manner of conclusion, some right, some wrong. Merely getting a paper published, or a single study being done, only reflects the work and conclusions of the people participating - who may have had various flaws in their work.
One would be a fool to take the result of a single study as gospel, or paper for that matter.
As part of the continue improvement of the body of human knowledge and understanding, we put forward all manner of ideas or investigations. Some wrong, some right. We slowly converge on increasing understanding, though not without various confusions along the way.
There there is that problem that research dollars are hard won and results are often demanded. This like in any other field can lead to a corruption of sorts.
Modern society wouldn't function if "science is wrong". Buildings wouldn't stay standing, water wouldn't flow out of taps, turds wouldn't flush away, lightbulbs wouldn't light up, cars wouldn't move, etc...
There's an amazing amount of understanding in even the apparently simplest of things, like my frying pan.
Sensationalist headline there - "Everything we know is wrong.". To the point of being a lie no less.
Papers come and papers go, they come to all manner of conclusion, some right, some wrong. Merely getting a paper published, or a single study being done, only reflects the work and conclusions of the people participating - who may have had various flaws in their work.
One would be a fool to take the result of a single study as gospel, or paper for that matter.
As part of the continue improvement of the body of human knowledge and understanding, we put forward all manner of ideas or investigations. Some wrong, some right. We slowly converge on increasing understanding, though not without various confusions along the way.
There there is that problem that research dollars are hard won and results are often demanded. This like in any other field can lead to a corruption of sorts.
Modern society wouldn't function if "science is wrong". Buildings wouldn't stay standing, water wouldn't flow out of taps, turds wouldn't flush away, lightbulbs wouldn't light up, cars wouldn't move, etc...
There's an amazing amount of understanding in even the apparently simplest of things, like my frying pan.
This topic has been archived, and won't accept reply postings.
Loading Notifications...