Differences between preliminary and formal testing

[Timothy]

I've heard several people say in various threads that the formal JREF Million Dollar Challenge test would be the same protocol as for the preliminary test.

The rules neither explicitly state nor refute this.

I had always assumed that the preliminary test was simply there to be a simplistic test that could be done relatively simply, so as not to go to the full-fledged expense, time, and trouble necessary for the formal test. The *essence* of the testing method would be the same, but the measures to prevent fraud, and the statistical needs of repeatibility would increase.

I know that if it was my money, I'd want to assure that someone wouldn't win due to chance or fraud.

Any comments from JREF, or someone who can reference JREF's stand? (I'm more interested in what the actual position is, rather than members speculations and opinions on what it ought to be.)

- Timothy

[Gr8wight]

Originally posted by Timothy
I've heard several people say in various threads that the formal JREF Million Dollar Challenge test would be the same protocol as for the preliminary test.

The rules neither explicitly state nor refute this.

I had always assumed that the preliminary test was simply there to be a simplistic test that could be done relatively simply, so as not to go to the full-fledged expense, time, and trouble necessary for the formal test. The *essence* of the testing method would be the same, but the measures to prevent fraud, and the statistical needs of repeatibility would increase.

I know that if it was my money, I'd want to assure that someone wouldn't win due to chance or fraud.

Any comments from JREF, or someone who can reference JREF's stand? (I'm more interested in what the actual position is, rather than members speculations and opinions on what it ought to be.)

- Timothy

Try PM'ing KRAMER on this.

[NoZed Avenger]

My limited understanding is that the test protocals are the same between tests, but that the "chance" element has to be removed.

While 8 of 10 positives (in a dowsing test, for example) might be a "1 in 1000" shot and be enough for the prelim, there would have to be many more trials to prove it to a statistically significant level -- sometimes called 1 in a million or whatever the level is set as.

The procedure remains the same, but something that would win by chance alone 1 in 1000 times would not generally be enough -- you'd need more like 49 of 50, for example, or perhaps even more.

Your mileage may vary.

N/A