I'd just like to point out that, statistically speaking, something has to be insanely strong to come up as effective on a double blind test.
Conduct thought experiment: We have a group of people. These people are of all ages, genders, weights, physical, mental and hormonal predispositions. This is a statistically representative sample of the rich people who can afford our new proposed medicine.
Oh, yeah, did I mention that we have a medicine to test. Okay, we do. In case A we will call it little blue pill.
We give little blue pill to half of the aforementioned group and a placebo to the other half, like good scientists. Since the little blue pill is so insanely strong, even the old guys with clogged arteries and dodgy tickers get a hard-on with the stuff. Hell, this stuff is so potent the women get a hardon. With the placebo - well, some things are just not possible without some hardcore vasodilation.
That's good medical science, that is. Medicine A approved. Notice how the entire spectrum of people were affected equally (or at least all were affected strongly) by the blue pill.
Now in test B we have a little white pill. When given to the incredible diversity of human forms it works on some, not so much on others, not at all on still others. Medical science decrees that it is bogus. Did it have an effect on some of the people? Undoubtedly. Was the effect statistically significant across the population. No. Is the medicine bogus? Beware lies, damn lies and misused statistics, even when eminating from 'experts'.
I agree that if a given preparation has a strong enough effect, it should show up obviously in any good clinical trial. Viagra, for example, has a very powerful effect and does work in the vast majority of cases, so it's statistically easy to demonstrate its effects by looking at the statistical data.
And yes, it's conceivable that some preparations have a more subtle effect that only apply in specific cases under specific conditions.
However, even a subtle effect can be measured and tested. With 6 billions humans on the planet, any given set of biological conditions should be replicated enough to derive a statistically meaningful sample of any given condition, and to do tests on that.
If, for some reason, it is impossible to derive such a sample (if the condition in question is just that rare) then no test on that sample can be considered very meaningful or valuable, as doing so would perpetrate the logical fallacy of the hasty generalisation.
But what we're talking about here is not something so unbelievably obscure that no scientist can test it. We're not talking about a genetic disorder that affects only one in a hundred million. We're talking about insomnia, that affects just about every human from time to time. Although its causes may vary, enough people suffer from it to be able to derive meaningful statistics on it.
We're also not talking about a customised homeopathic preparation, designed to cater to the specific needs of an individual (although I don't think it makes any difference). We're talking about an off-the-shelf remedy sold by the thousand. If it can be packaged and mass-produced like that, then there should be a demonstrable effect that would work on a large percentage of possible customers - me included.
I intend to test that. And I have little doubt that the effect will be zero.
I'd just like to point out that, statistically speaking, something has to be insanely strong to come up as effective on a double blind test.
ReplyDeleteConduct thought experiment: We have a group of people. These people are of all ages, genders, weights, physical, mental and hormonal predispositions. This is a statistically representative sample of the rich people who can afford our new proposed medicine.
Oh, yeah, did I mention that we have a medicine to test. Okay, we do. In case A we will call it little blue pill.
We give little blue pill to half of the aforementioned group and a placebo to the other half, like good scientists. Since the little blue pill is so insanely strong, even the old guys with clogged arteries and dodgy tickers get a hard-on with the stuff. Hell, this stuff is so potent the women get a hardon. With the placebo - well, some things are just not possible without some hardcore vasodilation.
That's good medical science, that is. Medicine A approved. Notice how the entire spectrum of people were affected equally (or at least all were affected strongly) by the blue pill.
Now in test B we have a little white pill. When given to the incredible diversity of human forms it works on some, not so much on others, not at all on still others. Medical science decrees that it is bogus. Did it have an effect on some of the people? Undoubtedly. Was the effect statistically significant across the population. No. Is the medicine bogus? Beware lies, damn lies and misused statistics, even when eminating from 'experts'.
I agree that if a given preparation has a strong enough effect, it should show up obviously in any good clinical trial. Viagra, for example, has a very powerful effect and does work in the vast majority of cases, so it's statistically easy to demonstrate its effects by looking at the statistical data.
ReplyDeleteAnd yes, it's conceivable that some preparations have a more subtle effect that only apply in specific cases under specific conditions.
However, even a subtle effect can be measured and tested. With 6 billions humans on the planet, any given set of biological conditions should be replicated enough to derive a statistically meaningful sample of any given condition, and to do tests on that.
If, for some reason, it is impossible to derive such a sample (if the condition in question is just that rare) then no test on that sample can be considered very meaningful or valuable, as doing so would perpetrate the logical fallacy of the hasty generalisation.
But what we're talking about here is not something so unbelievably obscure that no scientist can test it. We're not talking about a genetic disorder that affects only one in a hundred million. We're talking about insomnia, that affects just about every human from time to time. Although its causes may vary, enough people suffer from it to be able to derive meaningful statistics on it.
We're also not talking about a customised homeopathic preparation, designed to cater to the specific needs of an individual (although I don't think it makes any difference). We're talking about an off-the-shelf remedy sold by the thousand. If it can be packaged and mass-produced like that, then there should be a demonstrable effect that would work on a large percentage of possible customers - me included.
I intend to test that. And I have little doubt that the effect will be zero.