It is remarkably difficult to track down the studies on Ganzfeld, and the one’s I did track down were mostly meta. It seems that the early Ganzfeld experiments have been abandoned by the literature itself in favor of an “autoganzfeld”. Conditions are basically the same except that a computer randomly generates the image, the experimenter is not in the room, and the stream of consciousness of the receiver is both recorded electronically and transcribed by the experimenter. (It controls for problem variables which include when an experimenter knows the image/interacts with the participants/images are physical(finger prints and warmth tip off to the receiver which was in the hands of the sender) ect.) Further, the experimenter does not know the image being shown to the Sender. Interestingly the initial Ganzfeld with the methodological problems generated a significantly higher correlation that autoganfeld.
There is a sort of end all be all study for AutoGanzfeld methodology that all the studies I found referenced as the basis for their experiment. It is Honorton 1983 or 1990, but I cannot determine which study it is, nor can I find the name of either. It may be this, as per Honorton’s pax vitae: Psi communication in the ganzfeld: Experiments with an automated testing system and a comparison with a meta-analysis of earlier studies, but it is not available online.
I am at a dead end. All that I can conclude at this point, without seeing the agreed upon conditions, and keeping in mind that there is a serious documented problem where negative results are not published, is that the evidence demands further experiments. They need to test the repeatability of autganzfeld and generate the requisite amount of data. Further, the literature itself is undecided as to proof and proper controls.
As for me, besides the above, I have two serious problems with Ganzfeld based on my current knowledge of the methodology, or one problem that is two pronged. The problem is with the stream of consciousness that occurs while the receiver is in a relaxed state. The experiment lasts 30 minutes which gives the receiver enough time to pretty much say every damn thing that pops into their head, which can and does cover a whole lot of material. In studies that use a judge, would they not have a tendency to key in on the bits of the 30 minutes that correlate to the proper image? It seems to me that they obviously would.
The judge takes notes of everything the person says, then finds out the image the receiver saw, then tries to connect the notes to the image. This is quite flawed. The problem is, there are not enough studies available for me to determine how often this technique is used in the literature.
Also, why only 4 images? To me that seems like they are asking for statistical anomalies in the 200 or so experiment. Humans are cued based on sequence images are shown, and with so few images and so few tests it cannot be ruled out that this occured. I also cannot figure out why the hell the put ping-pong goggles and white noise on with a red light background. How the hell would they know if blanking out your senses would help psi when they are conducting a study trying to figure out if it even exists.
Recommendations: The conditions are replicated as per the literature, only receivers are offered far more images to choose from. Instead of 3 false and one true, do 9 false and 1 true. Also, test it under different conditions and see how the results vary. Never tell the experimenter on site or the receiver what the image was, rather leave that to someone offsite. Why risk this sort of taint when it’s so easy to get rid of it.