Placebo effect

[blutoski]

So then homeopathic medicine does work on animals as those where single blinded experiments(the animals did not know if they where getting the homeopathic treatment or not) so there was no observer bias to account for their behavior.

Again, I'm confused. Is this a question or a statement?

There have been homeopathic experiments done with animals. It's actually an issue within veterinary medicine right now. The results are mixed, arguments ensue, &c&c. My opinion is that the successful ones have not shown appropriate double-blinding. It's possible to blind the animal, but it's also important to blind the experimenter, since many animals (such as dogs) are ultrasensitive to non-verbal cues.

 

[ponderingturtle]

Whoa, only if there is a statisticaly significant variation vs. untreated animals in a large pool of trials. And the experimenter has to do everything the same for both groups of animals, except one gets the dose and the others don't.

All the homeopathy 'results' is usualy based upon single trial anecdotes and is therefore not a controlled study.

That is not the only way they can be flawed, the standard flaw is that they are not double blind but single blind, so the guy administering the remedy knows which animal got what, and can see what he wants in their behavior.

The thing is that the experimenter can not know which animal is in which group or he might well unconsiously cherry pick his data.

That is the point behind double blind experiments.

 

[ponderingturtle]

Again, I'm confused. Is this a question or a statement?

There have been homeopathic experiments done with animals. It's actually an issue within veterinary medicine right now. The results are mixed, arguments ensue, &c&c. My opinion is that the successful ones have not shown appropriate double-blinding. It's possible to blind the animal, but it's also important to blind the experimenter, since many animals (such as dogs) are ultrasensitive to non-verbal cues.

And you would have those same or similar problems with any no treatment vs placebo group. So you can not trust most observations made between those groups to be unbaised.

 

[blutoski]

And you would have those same or similar problems with any no treatment vs placebo group. So you can not trust most observations made between those groups to be unbaised.

I still don't understand why you say this, and you're still being vague. Do you mean unbiased observers or unbiased participants?

We're making an effort to manipulate the participants' bias - that's the point - so I'm unsure this would be a 'flaw': it's the purpose of such an experiment. We want participant bias when seeking placebo effect.

OTOH: if you're saying that there will always be an experimenter bias distinction between nontreatment and placebo groups, I disagree, and would like you to explain why you think so. So far, you've just been repeating the claim without explaining why you believe this. Two threat participants have provided counterexamples that show this is not the case and you have not addressed them.

You're explaining double-blind experiments to me as if I don't know. I designed medical experiments as an occupation for several years and it was the centerpiece of my doctorate, so I do understand the basics. Let's move to the next level.

 

[Eos of the Eons]

Aha,

good place to post a link on a new show I hope to see soon!

Medical Mythbusters

Placebos should come up a lot I would think.

Hey, any of you want to apply for the job? It will require double blind experimentations to prove why something doesn't work!

 

[ponderingturtle]

I still don't understand why you say this, and you're still being vague. Do you mean unbiased observers or unbiased participants?

We're making an effort to manipulate the participants' bias - that's the point - so I'm unsure this would be a 'flaw': it's the purpose of such an experiment. We want participant bias when seeking placebo effect.

You are also makeing an effort to not manipulate some participants bais. That makes measurements between groups difficult, as you want your observer to have no way to determine between the three groups. So the no treatment group might unconsiously by the way they answer questions and respond to the observer corrupt the observation.

OTOH: if you're saying that there will always be an experimenter bias distinction between nontreatment and placebo groups, I disagree, and would like you to explain why you think so. So far, you've just been repeating the claim without explaining why you believe this. Two threat participants have provided counterexamples that show this is not the case and you have not addressed them.

Because just like a big woo a doctor who beleives something strongly can convince his patients that something is changeing when nothing is, well you have reversed that effect here. My point is that the doctor will not be sufficiently blinded because the patient knows if they are in a treatment group or not.

You're explaining double-blind experiments to me as if I don't know. I designed medical experiments as an occupation for several years and it was the centerpiece of my doctorate, so I do understand the basics. Let's move to the next level.

And you have not seemed to sufficiently issolate just the placebo effect, which is strongest in highly subjective assessments anyway.

 

[blutoski]

You are also makeing an effort to not manipulate some participants bais. That makes measurements between groups difficult, as you want your observer to have no way to determine between the three groups. So the no treatment group might unconsiously by the way they answer questions and respond to the observer corrupt the observation.

Yesssss.... so you disengage this observer from the responses, as in survey format discussed above.

Because just like a big woo a doctor who beleives something strongly can convince his patients that something is changeing when nothing is, well you have reversed that effect here. My point is that the doctor will not be sufficiently blinded because the patient knows if they are in a treatment group or not.

So you make a protocol where they don't meet face-to-face, as in the above examples, to which you have not responded. You are merely repeating your claim. Are there setups where you think this format will not work?

I've been trying to think of counterexamples, and none come to mind. There are examples where it's hard to design an actual placebo, such as accupuncture, but testing the other side of the equation - accupuncture versus nontreatment - is pretty straightforward. Make sure the needlemarks are gone before the assessment and manage the chit-chat, and the observer is pretty well insulated from contamination from the patient.

And you have not seemed to sufficiently issolate just the placebo effect, which is strongest in highly subjective assessments anyway.

See above. I think I have sufficiently isolated the placebo effect in my examples. Why do you disagree? Enlighten me. I'm listening. Provide a counterexample or something.

 

[ponderingturtle]

With the changes in protocol that sounds like it might work then.