Re: Re: Why Most Published Research Findings Are False

[Guest guest]

> I have to disagree " that it's not scientifically correct... [if]it's

> merely statiscal. " Merely? Statistics is the basis for modern

> physics. And it's not scientific? Here's Hawking on the

> point.

I don't think the questionable credibility in the conclusions in most

published papers is due to the inherent uncertainty of quantum

mechanics to which you refer. In my opinion, one has to question the

conclusions of most studies because the reader has no way to verify:

the quality of the raw data, the quality of the data acquisition, all

the miscellaneous confounding factors involved in the data process,

the data analysis, etc. Once an article has been written and

published (i.e., polished and edited) only the most glaring

deficiencies will be obvious to the reader. If the results are only

evident upon statistical analysis then these associated

uncertainties, that I mentioned, should cause the reader to be highly

skeptical since there is no way to compensate (statistically) for

problems or errors in the whole experimental process. It seems that

in science, the closer one examines any particular area or study, the

murkier things become (i.e., what seems simple and concrete is

actually quite complicated) so (as our moderator often says)

" moderation " is the best approach.

Tom

[Guest guest]

> I have to disagree " that it's not scientifically correct... [if]it's

> merely statiscal. " Merely? Statistics is the basis for modern

> physics. And it's not scientific? Here's Hawking on the

> point.

I don't think the questionable credibility in the conclusions in most

published papers is due to the inherent uncertainty of quantum

mechanics to which you refer. In my opinion, one has to question the

conclusions of most studies because the reader has no way to verify:

the quality of the raw data, the quality of the data acquisition, all

the miscellaneous confounding factors involved in the data process,

the data analysis, etc. Once an article has been written and

published (i.e., polished and edited) only the most glaring

deficiencies will be obvious to the reader. If the results are only

evident upon statistical analysis then these associated

uncertainties, that I mentioned, should cause the reader to be highly

skeptical since there is no way to compensate (statistically) for

problems or errors in the whole experimental process. It seems that

in science, the closer one examines any particular area or study, the

murkier things become (i.e., what seems simple and concrete is

actually quite complicated) so (as our moderator often says)

" moderation " is the best approach.

Tom

[Guest guest]

When I went to school, scientific stuff didn't change from day to day.

If the human system truly worked as described we wouldn't be having this conversation.

Surely, they have to use stats to get into the ballpark, but MANY associations do not mean anything when it comes to application.

They give you clues what might be wrong.

Like the imaginary values in a quadratic solution, we have to guess whether they apply. If you take seriously all the articles on a given subject you'll not a scientific answer. Many times I give them a laugh test or a logic test rating.

And stats? Well look up Simpson's Paradox. Outcomes can change using the same data.

http://en.wikipedia.org/wiki/Simpson's_paradox

"It appears that the two sets of data separately support a certain hypothesis, but, considered together, support the opposite hypothesis"

How can I believe the data? Until they do a meta-analysis or something, 10 yrs later. That's why a number of people have to die before we know for sure the drugs/chemicals work.

Regards.

Re: [ ] Re: Why Most Published Research Findings Are False

> I have to disagree "that it's not scientifically correct... [if]it's> merely statiscal." Merely? Statistics is the basis for modern> physics. And it's not scientific? Here's Hawking on the> point.I don't think the questionable credibility in the conclusions in most published papers is due to the inherent uncertainty of quantum mechanics to which you refer. In my opinion, one has to question the conclusions of most studies because the reader has no way to verify: the quality of the raw data, the quality of the data acquisition, all the miscellaneous confounding factors involved in the data process, the data analysis, etc. Once an article has been written and published (i.e., polished and edited) only the most glaring deficiencies will be obvious to the reader. If the results are only evident upon statistical analysis then these associated uncertainties, that I mentioned, should cause the reader to be highly skeptical since there is no way to compensate (statistically) for problems or errors in the whole experimental process. It seems that in science, the closer one examines any particular area or study, the murkier things become (i.e., what seems simple and concrete is actually quite complicated) so (as our moderator often says) "moderation" is the best approach.Tom

[Guest guest]

When I went to school, scientific stuff didn't change from day to day.

If the human system truly worked as described we wouldn't be having this conversation.

Surely, they have to use stats to get into the ballpark, but MANY associations do not mean anything when it comes to application.

They give you clues what might be wrong.

Like the imaginary values in a quadratic solution, we have to guess whether they apply. If you take seriously all the articles on a given subject you'll not a scientific answer. Many times I give them a laugh test or a logic test rating.

And stats? Well look up Simpson's Paradox. Outcomes can change using the same data.

http://en.wikipedia.org/wiki/Simpson's_paradox

"It appears that the two sets of data separately support a certain hypothesis, but, considered together, support the opposite hypothesis"

How can I believe the data? Until they do a meta-analysis or something, 10 yrs later. That's why a number of people have to die before we know for sure the drugs/chemicals work.

Regards.

Re: [ ] Re: Why Most Published Research Findings Are False

> I have to disagree "that it's not scientifically correct... [if]it's> merely statiscal." Merely? Statistics is the basis for modern> physics. And it's not scientific? Here's Hawking on the> point.I don't think the questionable credibility in the conclusions in most published papers is due to the inherent uncertainty of quantum mechanics to which you refer. In my opinion, one has to question the conclusions of most studies because the reader has no way to verify: the quality of the raw data, the quality of the data acquisition, all the miscellaneous confounding factors involved in the data process, the data analysis, etc. Once an article has been written and published (i.e., polished and edited) only the most glaring deficiencies will be obvious to the reader. If the results are only evident upon statistical analysis then these associated uncertainties, that I mentioned, should cause the reader to be highly skeptical since there is no way to compensate (statistically) for problems or errors in the whole experimental process. It seems that in science, the closer one examines any particular area or study, the murkier things become (i.e., what seems simple and concrete is actually quite complicated) so (as our moderator often says) "moderation" is the best approach.Tom

[Guest guest]

Thanks, Tom,

Like looking at the Okinawans, or other group, and saying their diet is better, is totally unsubstantiated, IMO, just because they might have a few more centenarians than the U.S. There are many predisposing factors ignored, like do they drive to work, build high rise buildings, have to meet schedules.

And they can ignore the fact they might have a higher rate of stroke, while presenting an article on the goodness of eating fish for CVD. The writers don't hafta integrate the various diseases. So taken as a whole I can decide NOT to follow the recommendation to eat fish, eg.

You might not want to follow a rec to eat garlic for something if you're taking warfarin.

Now in the case of CR, it's a slam dunk, because we probably have been eating too much anyway, as long as we don't get obsessed with losing weight to some STATISTICAL norm.

Regards.

Re: [ ] Re: Why Most Published Research Findings Are False

> I have to disagree "that it's not scientifically correct... [if]it's> merely statiscal." Merely? Statistics is the basis for modern> physics. And it's not scientific? Here's Hawking on the> point.I don't think the questionable credibility in the conclusions in most published papers is due to the inherent uncertainty of quantum mechanics to which you refer. In my opinion, one has to question the conclusions of most studies because the reader has no way to verify: the quality of the raw data, the quality of the data acquisition, all the miscellaneous confounding factors involved in the data process, the data analysis, etc. Once an article has been written and published (i.e., polished and edited) only the most glaring deficiencies will be obvious to the reader. If the results are only evident upon statistical analysis then these associated uncertainties, that I mentioned, should cause the reader to be highly skeptical since there is no way to compensate (statistically) for problems or errors in the whole experimental process. It seems that in science, the closer one examines any particular area or study, the murkier things become (i.e., what seems simple and concrete is actually quite complicated) so (as our moderator often says) "moderation" is the best approach.Tom

[Guest guest]

Thanks, Tom,

Like looking at the Okinawans, or other group, and saying their diet is better, is totally unsubstantiated, IMO, just because they might have a few more centenarians than the U.S. There are many predisposing factors ignored, like do they drive to work, build high rise buildings, have to meet schedules.

And they can ignore the fact they might have a higher rate of stroke, while presenting an article on the goodness of eating fish for CVD. The writers don't hafta integrate the various diseases. So taken as a whole I can decide NOT to follow the recommendation to eat fish, eg.

You might not want to follow a rec to eat garlic for something if you're taking warfarin.

Now in the case of CR, it's a slam dunk, because we probably have been eating too much anyway, as long as we don't get obsessed with losing weight to some STATISTICAL norm.

Regards.

Re: [ ] Re: Why Most Published Research Findings Are False

> I have to disagree "that it's not scientifically correct... [if]it's> merely statiscal." Merely? Statistics is the basis for modern> physics. And it's not scientific? Here's Hawking on the> point.I don't think the questionable credibility in the conclusions in most published papers is due to the inherent uncertainty of quantum mechanics to which you refer. In my opinion, one has to question the conclusions of most studies because the reader has no way to verify: the quality of the raw data, the quality of the data acquisition, all the miscellaneous confounding factors involved in the data process, the data analysis, etc. Once an article has been written and published (i.e., polished and edited) only the most glaring deficiencies will be obvious to the reader. If the results are only evident upon statistical analysis then these associated uncertainties, that I mentioned, should cause the reader to be highly skeptical since there is no way to compensate (statistically) for problems or errors in the whole experimental process. It seems that in science, the closer one examines any particular area or study, the murkier things become (i.e., what seems simple and concrete is actually quite complicated) so (as our moderator often says) "moderation" is the best approach.Tom