So 6 answers are valid when they agree with your conclusions, but 3 of those are invalid when they do not. How non-scientific of you. ...
In this test, I got answers from
6 people (you [Hokulele], Ladewig, stanfr, Kid Eager, Femke and gabeygoat) who posted answers with "xx", and sent an unambiguous and valid numerical answer to the assistant Agatha (see
post #127 and
post #149). I think that one may reasonably adopt the point of view that these 6 people gave valid answers (not providing a MD5 hash is not very important if an answer is sent to the assistant, as Ladewig said:
...
.
I am not encrypting it, I will just send it to Agatha
).
Two of those were numerically correct ( the answers by Ladewig and Femke, this is what Agatha found in her analysis), this gives us a correct answer rate (CAR) equal to CAR = 2/6 = 33%, which a little higher than the CAR expected from chance alone, which is about 25%.
However, a closer look reveals that three of these answerers (stanfr, Kid Eager and gabeygoat) gravely violated the recommended protocol explained in the
opening post, because they had sent to Agatha a text which was
different from the one they had posted in the thread. For exemple, gabeygoat posted:
"
I'm gonna guess xx
not sure what im doing",
whereas he sent "
guessed 4 de4f022d0deba6911bedf9a0350256aa". These two texts "sound" rather differently, and the credibility rating (CR = 0) I gave to gabeygoat's first text, using the "blind" procedure, does not necessarily apply (and, even, probably does not apply) to his second text. In my
opening post, I "recommended" that participants send their texts (one, and only one for each participant in the test) to the assistant(s). I didn't say:"Hey guys, feel free to invent another text (please, show great imagination), and to send that one to the assistant(s)". I said:
It may also be useful (I recommend it) that you send your (full) answer, in the form of a private message, to either Agatha, or Femke, or to both.... I said "
your (full) answer". I didn't say "some text, with your chosen number". But now comes the interesting part.
None of the three participants who violated the recommended protocol (stanfr, Kid Eager and gabeygoat, as said above) provided correct numerical answers. In other words
the CAR of this group was equal to 0%. This is apparently an interesting confirmation of the ideas I have explained and developed in my previous threads, on the correlation between credibility (here, credibility stemming from carefully abiding by the recommended protocol) and numerical correctness. The average CR I gave (in a blind way) to people who followed the recommended protocol was (0-5+0)/3 = -1.7, the average CR I gave to people who did not, was (-5-5+0)/3 = -3.3 (a little less).
The correct answer rate for the people who followed the recommended protocol in its most important aspects (Hokulele, Ladewig and Femke) was equal to CAR = 2/3 = 67%. This CAR is the most important and significant hit rate, I believe, the hit rate of the most careful, serious, credible and motivated people.
...
No, I don't rate for credibility in this way (or, at least, I try not to)
...
But you did. Go read your analysis of my response again.
No, that's not true. This what I said about your answer in post #127:
...
Her answer: "
The first number that came to my attention is xx, so that is my choice for this test." does not make her answer credible, in my opinion, in a telepathy test, where the thing that matters is what you (telepathically)
perceive, not the first number which "comes to your attention". Hokulele's answer is written using a rather formal style, which is odd here (and which I even find slightly arrogant
, which is not favorable for credibility). Telepathic perception is probably mostly an involuntary process, something you mostly cannot control, it's not like "giving (visual) attention" to an image.
But she also said later 'I "predict" that his credibility ratings
will no longer correlate with correctness'. This does suggest that she wrote a (numerically) correct answer, with a not credible answer (she most probably realizes her answer is not credible), so that the usual correlation is lost (her answer does not seem credible; if the frequent correlation must be lost, one should expect it to be numerically correct). ...
So, instead of giving a negative credibility, I finally choose: CR = 0 (I note also she gave no MD5 hash). I prefer to not give a positive credibility to an answer which explains a number choice by "a first number coming to attention" (as if she had seen it).
...
In other words, I revised my CR of your answer from negative to zero, because of what you said later.
