Art, the test protocol you describe is the same as what I've used with dowsers. 10 locations, 10 trials, 1 target. With guessing, the probability of success in a single trial is 10%. In 2 trials, the probabilities for successful guessing are
0 correct = 81%
1 correct = 18%
2 correct = 1%
Using coefficients from Pascal's Triangle, you can calculate the odds for 10 trials:
0 correct = 34.868%
1 correct = 38.742%
2 correct = 19.371%
3 correct = 5.740%
4 correct = 1.116%
5 correct = 0.149% = 1-in-672
6 correct = 1-in-7258
7 correct = 1-in-114312
8 correct = 1-in-2743484
9 correct = 1-in-111111111
10 correct = 1-in-10,000,000,000
So a 60% success rate has a 1-in-7258 chance with random guessing. Assuming guessing, if you were to test 10 people it's unlikely any of them would hit 60%, so maybe that's a fair threshold. If you were to test 10,000 people then it's fairly likely one of them could guess 60% and you might need to increase the threshold. When I've tested dowsers for my $25,000 prize I've set 70% as the threshold for success. The best result I've ever seen from a dowser is 2 correct out of 10.
Before the dowser attempts the double-blind trials I have him do a few "full-view" trials. In a full-view trial the location is randomly selected as usual, the target is concealed as usual, but the dowser gets to watch it all happen. That is, the dowser knows where the target is located. He then attempts to dowse for it. Typically I ask for 5 full-view tests; I've never seen a dowser perform less than 100% for this phase, and the dowser is always quick and confident. But when the double-blind trials start, the dowser suddenly struggles to locate the target and is never confident. Sometimes the look on their face is priceless; they go from super-confident to completely bewildered, not understanding why their dowsing skills have suddenly vanished.
There are several reasons for the full-view trials:
- To get the dowser familiar with the protocol and to see if he wants any last-minute adjustments.
- To see if there are any anomalies that cause interference.
- To provide a set of baseline results to compare with the double-blind results.
- To minimize the post-test alibis
#3 is pretty important. You say you will only test dowsers who meet the 60% threshold in a trial run; is this full-view or double-blind? If it's DB, then you are asking dowsers to pass the test in order to take the test, which doesn't make sense. If it's FV, then that makes sense; keep a record of those results.
#4 is also important. I've never seen a dowser succeed in a reasonable test and I've allowed for some loose protocols. But what's pretty consistent is that, after failing, the alibis and excuses start rolling out. The FV trials help minimize the alibis. The dowser was perfect and confident when he knew where the target was, but could not do better than guessing when he did not. If the results of the DB trials are held until the very end, it's also very useful to follow the DB trials with another short series (maybe 5) of FV trials. Most likely, the dowser will, again, score 100% which shows that dowsing conditions (temperature, solar winds, etc etc) did not degrade. All of this is to make the DB results as transparently obvious as possible.
What you are attempting to do, I've done a lot. Keep me in the loop on your progress, I'll be happy to help.