What if your life (seems) to go better with prayer?

[SpuriusMaximus]

What was my positive claim?

As I have already said before, if we're talking about medicine and the efficiency of treatments, then the medical standard applies. Namely, the positive claim is that it actually does anything better than placebo, while the null hypothesis is just that: null. "Nope, it was just coincidence or placebo effect."

Which is really just a sub-case of its being about statistics. What is being claimed is a correlation. If I want to claim any correlation between variables X and Y -- be it between talking to someone and depression, or between masturbation and asthma (yep, it's actually a thing), or just between how late it is in the evening and my blood alcohol level -- then the existence of that correlation is the positive claim. The null hypothesis is just that: "Nope, whatever dependency you see in your data is just coincidence." I need the data, and enough of it at that, to disprove beyond a certain required threshold that it can all be just a big series of coincidences.

Well, in medicine it's just that plus comparing it to placebo. The positive claim is any claim that something helps (or harms) more than placebo effect. I.e., the correlation must be stronger than that for placebo effect, and you need the sample size to be sure it is so.

E.g., if I wanted to claim that my homeopathic essence of snake oil actually helps with a cold, the positive claim is always just that: it actually helps. The null hypothesis is "nope, it doesn't." The onus is on the one making the former, not on the one falling back on the latter until evidence is presented.

I'm sure you're a smart person and can figure it out from there.

 

[SpuriusMaximus]

ETA: I'd still be interested in your answer to several questions I posed.

I think I did answer what seemed to be the important points, and several that didn't. But if you can write a list of the questions you still have, FSM knows I never discouraged anyone from seeking answers.

 

[whoanellie]

extracting the relevant portion where Minoosh pointed out that HM did not address the meta-analysis posted above.

...
Which is why I said that OTHER people may have tried to meet that burden of proof, but YOU didn't. So, sorry, YOU're not in a position to say that I'm having eyesight problems if I don't just accept whatever irrelevant stuff your googling turned up.

That said, I did ask them how that was tested against placebo effect, far as I noticed, I got no answer. I'm willing to be educated, but it means just that. I'm not under any obligation to accept an answer that wasn't actually given.
...

I posted the link to the meta-analysis. No where in this thread did you ask me about how the placebo effect was dealt with in that meta-analysis. You did post saying that it was not blinded and I responded by saying that that was incorrect (my post #74 above). That is where the exchange between you and I stopped. The meta-analysis looked at both published and unpublished results from dozens of different studies. Each study would have had a different design but is wrong to say that none of those studies was blinded or none of them had a no-treatment control. You would have to read the PLOS ONE article I linked to and the references therein for additional details. I think it accurate to say that you made an assertion about the nature of the work I linked to without actually reading or understanding what was in the paper.

The meta-analysis I posted addressed your assertion that there are no "studies about results" "medically speaking" that demonstrate the benefit of therapy, an assertion which is patently untrue.

It's just the fact that medically, speaking as in studies about results, it doesn't really do anything long term, if you actually had a problem to solve other than needing someone to talk to (and your friends didn't fit the bill for some reason.)

 

[SpuriusMaximus]

You probably missed the question in my message #71. Not only I pretty much start with that, but there's this bit verbatim in it: "Here how do you give a control group placebo therapy to compare it to that, and how do you make it double-blind?" No matter, I'm still interested in an answer if you're willing to provide one.

There was no question (from me) of whether there is a control group per se, although I do have to wonder what some of those groups are even supposed to be a control for. E.g., in one of Minoosh's links, apparently the control group were healthy people.

 

[Agatha]

Dial down the bickering and insults, please.

Replying to this modbox in thread will be off topic  Posted By: Agatha

 

[whoanellie]

You probably missed the question in my message #71. Not only I pretty much start with that, but there's this bit verbatim in it: "Here how do you give a control group placebo therapy to compare it to that, and how do you make it double-blind?" No matter, I'm still interested in an answer if you're willing to provide one.

There was no question (from me) of whether there is a control group per se, although I do have to wonder what some of those groups are even supposed to be a control for. E.g., in one of Minoosh's links, apparently the control group were healthy people.

Again, the report I posted a link to was a meta-analysis. It started with a pool of 55 funded grants. Each one of those studies would have had it's own design. You can read the meta-analysis and search for "blind" or "placebo" or "control" to get some information. For full details you'll need to look at the original research reports (if they have been published). For example, reference 51 of the meta-analysis is entitled "Treatment of atypical depression with cognitive therapy or phenelzine: a double-blind, placebo-controlled trial". So, if you really want to know how researchers would conduct a "double-blind, placebo-controlled" study of the effectiveness of therapy for the treatment of depression that might be a good place to start.

 

[SpuriusMaximus]

Trust me, I did try, but it wasn't easy to get to the information. Well, for someone with my skills, anyway. That's why I'm asking someone who is presumably better skilled in that.