Vitor,
So far as I can tell, even if this "double-blind" experiment is accepted as valid, it doesn't meet Randi's criteria. What you guys did, so far as I can tell, is got a bunch of bottles of water, then Emoto just said, "this has been affected" and "this hasn't been affected". And he managed to get what you consider to be a statistically high success rate (although it is not clear to me how many different samples were used, which has a significant impact on interpretation of any results).
So let me present to you an example of what I believe Randi would look for as actual 'proof'; and then ask if you guys have actually done this.
Step 1: Water is put into several hundred identical bottles of water. The bottles are all supplied by an independent source, to ensure no secret marking is done; and all water comes from exactly the same source. Each bottle is labeled with a number, and nothing else.
Step 2: Various individuals -- all of whom are not part of Emoto's project, and who are not informed of the purpose of this study -- are given perhaps 20 bottles of water each, and told to express a particular emotion towards that water (hate, love, anger, happiness, etc.). An independent juror keeps track of the labeled bottles, and which emotions they were exposed to. 1/3 of the bottles of water are set aside as controls, and not exposed to any emotions.
Step 3: The water is frozen.
Step 4: Emoto is allowed to examine each bottle of water (but all bottles are handled by people who have no connection to him), in an entirely random order; but he must do it entirely alone, with nobody else there who might provide clues (and of course he is not allowed to observe or participate in any of the previous steps).
Step 5: Emoto must identify three things: 1) Has the water been exposed to emotions, or is it part of the control group? 2) If it has been exposed to emotions, were those emotions positive or negative? 3) What specific emotions were they exposed to?
RESULTS:
If Emoto fails to be statistically significant results to all three questions, his claims are demonstrated to be entirely false.
If Emoto gets statistically significant results to the first question, but not the latter two (which, so far as I can see, seems to be the question that was addressed in your "double-blind" study), it could be held as evidence that emotions have some kind of impact; but since it is impossible to distinguish positive from negative results, it still renders Emoto's claims to beneficial effects entirely pointless. Such a result would certainly raise enough questions to deserve more study (personally, I don't believe at a proper double-blind study would have any such results), but is nowhere near to deserving the million dollars.
If Emoto gets statistically significant results to the first two questions, that would be an indicator that there actually is a demonstrable difference in the properties of water exposed to positive or negative emotions, and would likely be worthy of a million dollar prize. However, from what you've told us, I don't see anything that indicates Emoto has accomplished this.
And if Emoto gets statistically significant results to all three questions, that would not only deserve a million dollars, he should get a Nobel prize for standing the world of physics on its head and revealing an entirely new principle to science.
I have strong doubts that the study provided enough scientific rigor to consider results to the first question reliable; and I've seen nothing to indicate answers to the second and third ones. Nor do I expect, if subjected to a proper double-blinded experiment, that Emoto ever will get statistically significant results to any of these questions.
