It would seem that DAT mostly applies to things like RNG's and perhaps medical treatment effects, but perhaps Linda can provide insight about other possible applications of such a theory (I seem to remember her mentioning that she was a fan of it in some other thread). I confess to not fully understanding it, but I think that it has some traction from people on both sides of the debate so I thought it could be a fruitful avenue. Perhaps someone more learned than I can help to summarize it for us hobbyists.
RPKP: How did Decision Augmentation Theory come about, and who else was involved in its development?
May: DAT resulted from a carefully conducted RNG study we did for the Army in 1979. We spent over $250,000 on that single experiment so that if we saw an effect we would be certain from an engineering point of view to learn how it happened. Well we did see an effect and could prove that the hardware was not changed. The only thing left was that subjects were being statistical opportunists to capture deviant subsequences from otherwise unperturbed sequences. Beverly Humphrey and Jessica Utts are my colleagues from the beginning. James Spottiswoode joined much later but has also contributed.
RPKP: Could you elaborate on the basic hypothesis of DAT?
May: An RNG experiment is like an electronic coin flipper. Imagine that someone flipped a fair coin 10,000 times and wrote on a very long piece of paper the results. Might look something like this: htthhthttthhththhhhh.... Suppose I gave you a red pen and asked to walk along this long piece of paper and mark a spot where the next 10 coin flips had far too many heads in a row. You would have no problem making such a "decision." In the DAT model, subjects use psi to make a similar decision.
DAT is the idea that a selected sample is formed from an ordinary distribution. In the case of RNG's, what is selected out is shorter strings which are non-uniform, from an overall distribution which is uniform. In the case of medical studies, it would be selecting the subjects who are going to do well anyways, and assigning them to the 'intervention' group. In ganzfeld studies, it would mean that subjects who were going to mentate about worms regardless happened to be assigned the worm target.
What it means is that the outcomes are unchanged in the presence of psi. The output of the RNG is still completely random, the ganzfeld mentation is unchanged regardless of which target was presented to the sender, the subjects recovered from their heart attacks at the same rate they always do. What changes is that "what was going to happen anyway" manages to get matched up with the right outcomes to make it look like there is an effect - the RNG output is changed, the subject guessed the target, the subject recovered after they took aspirin.
It's still regarded as "psi". The idea is that the selection is made non-consciously and sort of before the fact, making it still anomalous.
Linda, I guess I did not understand your description. They say in the first paper:
Since the case for AC-mediated information transfer is now well
established (Bernm & Honorton, 1994), it would be exceptional if we did
not integrate this form of information gathering into the decision
process. For example, we routinely use real-time data gathering and
historical information to assist in the decision process. Why then should
we not include AC in the decision process? DAT holds that AC
information is included along with the usual inputs that result in a final
human decision that favors a "desired" outcome. In statistical parlance,
DAT says that a slight, systematic bias is introduced into the decision
process by AC
The "decision-making process" includes making sure that the targets are assigned in a way that makes the experiment look like it was successful. The idea is that anomalous cognition on the part of the researcher is used to let them know in advance which assignments will be advantageous. This is different from subjects altering the outcome based on anomalous cognition. In statistical parlance, there is a systematic bias introduced into 'selection from a normal distribution' vs. unbiased 'selection from a perturbed distribution'.