The Global Consciousness Project

Linda, I notice that you ignored the most significant part of my post: the explanation based on unsynchronised XOR masks as to why we seem to be able to rule out a non-anomalous hypothesis. You are of course free to do that, I just want to note it.

You quoted and directed that part to MaxB. It didn't occur to me to jump in to that discussion.

That doesn't answer my question.

It was meant to. Commenting on the distribution of events within their threshold for "extreme" events would be relevant. Looking at how a shift in the curve changes the number of events within that threshold would not.  

Perhaps I can give you a better idea of what I mean by "exactly". It might look something like this:

You list, in your response in this thread, the experiments in which you claim that Bem, Radin, and others have been caught changing their hypotheses after the fact. For each of your claims, you state the prespecified hypothesis, quoting it exactly with a checkable reference, and then you state the changed, after-the-fact hypothesis, quoting it exactly with a checkable reference. Then you quote directly, with a checkable reference, the person who "caught" this change. If there was a response to the accusation by the researcher(s) in question, then you directly quote that response with a checkable reference.

That is, of course, exactly what was offered to you in the Feeling the Future thread I referenced. And to a less detailed degree on the Radin blog - you have to dig a little on your own to see that the experiment was based on a prior experiment which reported on treated vs. untreated groups, and that his power calculations were based on this comparison (not on the comparison he eventually reported - treated vs. untreated groups only amongst the small subset of believers). I think there would be difficulty finding another example which fits your fortuitously capricious requirements, given that the practice is common (and therefore gives the appearance of acceptability) and I have yet to see a proponent parapsychologist call out their colleagues for doing so. It seems to depend upon whether or not a non-proponent takes the time to look at a particular paper, for that problem to be mentioned. And of course, as has been mentioned numerous times here and elsewhere, it's hard to get non-proponents to take the time to look at research which is assumed to be bollocks a priori (unfortunately).

There are two questions here:

1. Does changing a prespecified hypothesis after the fact amount to deliberate fraud?
   
2. Do (para)psychologists regard such a thing as deliberate fraud?
   

Can you please confirm that you accept that the answer to the first is "Yes" even though you continue to answer "No" to the second?

The answer to the first is "no". Changing a prespecified hypothesis after the fact is excused as "following the evidence", "looking at those groups/conditions in which the experiment worked", "it was a pre-planned condition" based on recollection, etc. Unless you have some sort of documentation as to the hypothesis, recorded prior to the results being obtained/known, anything can be claimed as the hypothesis of interest. And the researchers may even sincerely recall that that was their idea all along - we all know the problem of how our recollection is overwritten once we are given feedback.

If you want to know what the pre-specified hypothesis was, without having to depend upon a researcher's self-interest, you can look at the research which the study is based on, you can look at what hypothesis the power calculation was based on, and you can look at study registries.

So, there are basically three parts to your claim:

1. Parapsychologists change prespecified hypotheses after the fact.
   
2. Scientists in other fields are aware of this.
   
3. It is because of this awareness that scientists in other fields don't take parapsychology seriously.
   

We might add the implied qualification to the first part: that this is done enough that it generally invalidates the results of the field.

That paper doesn't even begin to justify the first part of your claim: reading the abstract, it deals with the field of psychology, not of parapsychology - the word "parapsychology" does not even appear in the paper.

I said, "because they are aware that this is what researchers do, even if they are not supposed to." I wasn't referring to Paraspsychology, otherwise I would have said, "because they are aware that this is what parapsychologists do, even if they are not supposed to." I was referring to researchers in general. Because researchers in any field (particularly the social sciences) know that the use of QRP's are fairly ubiquitous (even sometimes using them themselves), if given the chance, they are skeptical of research where the use of QRP's could create the supposed effect.

I doubt it. I've tried it in three scenarios: logged in under my ordinary administrative account, logged in under a test, non-admin account, and not logged in. It takes me to the same (correct) post each time.

I wasn't criticizing you. It didn't work for me, and I was just explaining why I didn't know which claim you wanted me to address. I don't think it matters, because I don't recall that you ever addressed the fortuitous selection problem that Bancel found.

If in that context by "fortuitous" you mean something like "serendipitously guided by some anomalous force, entity, or other phenomenon" - which I think is the sense in which Peter Bancel intends his explanation - then I can accept the possibility of "fortuitous selection" after all.

"Fortuitous" means "event selection which worked out well for the researchers".

The sense of "fortuitous selection" which I think has been ruled out in this experiment is something like "occurring by a random process of blind luck". The post to which I linked explained one way by which it has been ruled out: a resampling analysis in 2008 demonstrated that results with the same level of significance could be obtained only once in 100,000 attempts (2008 was well before the experiment ended, so the likelihood is that the figure would have been even higher by the end).

To this, you responded: "I don't think anyone disagrees that the event samples are improbable under random sampling. The question is whether they would also be improbable if samples were drawn for other goals". The problem with this, as I discussed at length in the post to which you were responding, is that it implies that there is some causal mechanism at work which relates the goals (selection criteria) to an effect - but we seem to be able to rule out non-anomalous causal mechanisms, and you agree that we can rule out blind luck, so we continue to be left then with the one hypothesis that we can't rule out: that the results were obtained anomalously (including the possibility of "fortuitous selection" in the sense of "serendipitously guided by some anomalous force, entity, or other phenomenon").

I don't know why you think non-anomalous causes have been ruled out. The tests you have mentioned previously have been inadequate to rule out all non-anomalous causes (including the most obvious ones), and have even, in some cases, been inadequate to rule out the cause they purport to be testing.

And, again, this is all based on the conditional: if everybody involved is being basically honest, then how can we explain these results? Any argument for dishonesty is a separate issue.

If everybody involved is as honest as everyone else, then the results can be explained as defensible flexibility.

Linda