The Global Consciousness Project

So I was justified in asking the question then?

Malf: "The rabbits are just events right?"

Laird: "The hats are the events..."

[Laird]

The rabbits are just events right?

No - see my reply to Kam above.

The magic appears to be in aligning the correlations in a fortuitous way.

The real magic is why there happen to be so many rabbits that their overall significance is at the type of odds cited by Kam.

I’ve lost track of what you’re arguing for

I'm simply arguing that these data can't be explained in any conventional way (ETA: other than experimenter fraud), including by "fortuitous selection" (and I'm still waiting for somebody in this thread to lay out the case for such an explanation in detail). I'm not committed to any particular anomalous explanation.

[Laird]

So I was justified in asking the question then?

Malf: "The rabbits are just events right?"

Laird: "The hats are the events..."

Wait... you're malf?

ETA: Never mind that poor attempt at humour: yes, you were justified.

No - see my reply to Kam above.

I thought your analogy could be improved. Mine is better at showing where the magic is happening (IMO).

The real magic is why there happen to be so many rabbits that their overall significance is at the type of odds cited by Kam.

That is not what Bancel is saying. Kam is quoting the GCP's own calculations, I guess question is about how they are arriving at their calculations, given Bancel's conclusions.

I'm simply arguing that these data can't be explained in any conventional way, including by "fortuitous selection" (and I'm still waiting for somebody in this thread to lay out the case for such an explanation in detail). I'm not committed to any particular anomalous explanation.

I thought I was the one being vague?

[Laird]

That is not what Bancel is saying.

Well, for a start, Peter Bancel accepts that the results are anomalous and can't be explained in conventional terms, so your appealing to him for support is not necessarily the wisest of choices. To provide you with a citation for that, he asserts in his 2017 paper to have "shown that the observed result is consistent with goal-oriented psi" (emphasis mine).

But secondly, his hypothesis of a "goal-oriented" effect (GO) leaves somewhat open the exact (anomalous) nature of the mechanism - in that same 2017 paper, he writes of GO that (footnotes elided) "Proposals include precognitive intuitions that might preferentially select events at times when natural statistical deviations happen to fluctuate toward hypothesis confirmation, and psychokinesis, a mind-matter interaction that acts to perturb the network."

Note in particular that he is not suggesting anyway that any preferential selection would/could be "fortuitous" in the sense of "a happy and lucky random coincidence", but that it would be based on some sort of psi functioning ("precognitive intuitions").

Well, for a start, Peter Bancel accepts that the results are anomalous and can't be explained in conventional terms, so your appealing to him for support is not necessarily the wisest of choices. To provide you with a citation for that, he asserts in his 2017 paper to have "shown that the observed result is consistent with goal-oriented psi" (emphasis mine).

But secondly, his hypothesis of a "goal-oriented" effect (GO) leaves somewhat open the exact (anomalous) nature of the mechanism - in that same 2017 paper, he writes of GO that (footnotes elided) "Proposals include precognitive intuitions that might preferentially select events at times when natural statistical deviations happen to fluctuate toward hypothesis confirmation, and psychokinesis, a mind-matter interaction that acts to perturb the network."

Note in particular that he is not suggesting anyway that any preferential selection would/could be "fortuitous" in the sense of "a happy and lucky random coincidence", but that it would be based on some sort of psi functioning ("precognitive intuitions").

We live in a world where conscious/unconscious bias (normally for positive outcomes) in such scenarios is well documented. 

I like Radin. He seems like a good guy. He is very invested in finding ‘love’ and ‘magic’ in nature and that’s sweet. He invariably and unnervingly finds it too.

There’s an imbalance here. You want me to explain in detail some ‘conventional’ explanation, whilst remaining cagey over unconventional options. Fair enough I guess, nobody knows what’s going on here*. Hopefully you understood my reference to Hennacy-Powell’s study. The point of Bancel’s Work is that he has looked at the data and has completely altered the scope of the exercise. He’s rejigged a new hypothesis to fit the data because the original one (love affecting electronic devices) doesn’t appear to work. Surely, it’s time to pull the plug on this and redesign a study that zeroes in on what is going on? 



*I’m surprised by the traditional proponent support in some quarters for this fiasco. They should be demanding much better.

[Laird]

We live in a world where conscious/unconscious bias (normally for positive outcomes) in such scenarios is well documented.

So, basically, you're unable to justify your contention that selection bias can explain the results, but that's not going to stop you from repeating it ad nauseam. It's probably best to leave it at that then.

So, basically, you're unable to justify your contention that selection bias can explain the results, but that's not going to stop you from repeating it ad nauseam. It's probably best to leave it at that then.

I’m open to unconscious bias in a motivated experimenter. The evidence that such phenomena exist is very high. (I have made some justifications, but they’re normally in the bits you decide not to quote in your responses.)

I’m open to the possibility of some experimenter psi, but that wasn’t the hypothesis that was put forward to be falsified here. The evidence that such a phenomenon exists is much harder to determine.

Selected samples are those which are not representative of the underlying population from which they are drawn - in this case, the population of global events capturing our global consciousness. Regardless of whether you favor anomalous or non-anomalous causes, it appears that the events listed as formal hypotheses represent a selected sample.

Selection bias is the effect which is introduced into that sample by the selection process, for example the bias of a motivated experimenter.

The wikipedia article on this is pretty good at explaining the different kinds of selection and their effects/bias, although it takes more of a medical perspective.

This is what Bancel said about selection wrt the GCP in the 2014 paper:

"Prominent among these [alternate hypotheses] are selection hypotheses, which posit that, for each replication, the experimenter's choice of events and subsequent designation of event time-periods may lead to small biases in the event Z-scores. In these scenarios, the fortuitous selection of naturally occurring random deviations is responsible for the measured effect, rather than a physical influence which alters the RNGs' behavior. The ambiguity is aggravated somewhat by the conception of the GCP experiment. The formal experiment tests only the event Z-scores and does not propose or test for any underlying mechanism or signature of the effect. Furthermore, the latitude allowed in event choices is consistent with selection bias, as this freedom is necessary to generate biased Z-scores."

In his 2016 paper, he goes on to demonstrate that (at least some of) the registered events represent a selected sample.

So even amongst proponents there is general recognition that the conditions are present for selection bias to generate the observed effect, with a selected sample. The remaining question is whether or not the details of the selection bias are anomalous or non-anomalous, given that non-anomalous causes are not prevented (or made even a little bit onerous), and that researchers' use of flexibility often goes unrecognized by themselves (or if recognized is seen as defensible). Statements made about pre-specified statistical tests may be given sincerely, even in the setting of considerable flexibility (as per Gelman and Loken).

[Laird]

This is what Bancel said about selection wrt the GCP in the 2014 paper

And here's how he continued a couple of paragraphs down (emboldening mine):

One can distinguish between conventional and psi models of selection. In the former case, the selection of events and their time-periods is guided by direct information about the data before the replication specifications are determined. Conventional selection can arise from either poor methodology or data-peeking which allows the experimenter to gain prior information about the replication outcome.

You and malf have still not explained what it is about this experiment's methodology - or, in other words, its blinding procedure - that can account for the results in terms of selection bias. The experimenters are very clear that they were blind: that they pre-specified each event's start and end time, and the method of statistical analysis to be used, before looking at the data relating to it.

So, either you guys are maintaining that the experimenters are lying, and that they actually peeked at the data, or you are maintaining that this means of blinding is insufficient, or you are maintaining both. In the latter two cases, could either or both of you please finally explain what is insufficient about this blinding procedure?