Dave Rogers, the correct answer of this test is "fire" (the fifth word in the list of possible choices), and I find no clear evidence that you knew this, from reading your post.
However, I just completed a closely related test (to this one) on the French Yahoo Answers site:
https://fr.answers.yahoo.com/question/index?qid=20170513192840AARBM6p , where the correct answer was "5".
Four credible answers were given, two of which (50%) were correct (50% is of course much larger than the approximately 20% of pure random chance).
The person who gave the correct answer was SAINTMARC (you can probably easily figure out what such a name might mean regarding credibility). This Yahoo member commented, right after I chose his answer as the best: "Je le savais !! ..." (this means "I knew it!! ...")
Error correction: the highlighted phrase should be replaced by "The person who gave the best correct answer was SAINTMARC ..."
(feel free to point out any error).
Actually, I just checked that thread.
You scored 20% as expected.
Ten people answered (
here, it's in French!) and 8 people gave valid answers (equal to one of the five numbers: 1, 2, 3, 4, 5). So, even without the important credibility filter, the hit rate was 2/8 = 25%, not 20%.
With the credibility filter on, the hit rate rises to 50% (not 100% !). I have explained in post #664 why lejofrapit's answer is not credible.
SAINTMARC, the person who got the best answer in my test, has a best answer rate on Yahoo of 25%. The (known) best answer rate of other members who have participated in the test are: lejofrapit: 5%, Yô: 7%, Blondin: 10%, Cousine Itt: 11%, rene: 1%, Alexandre: 14%, and Nanard: 7%.
The fact that good member SAINTMARC said "I knew it!!" right after the end of the test is an important and valuable testimony, which is reminiscent of other testimonies, obtained even on this forum, for example:
... I do indeed have ESP, and know for a fact that he wrote 2!
I am hearing Michel H's thoughts. ...
Some members have offered suggestions for "better tests" (according to them). I read these suggestions very carefully, of course. An obvious problem I have on this forum is that nobody seems to want to participate in my tests any more. So, if one of your suggestions could fix that problem, it could perhaps be useful. However, I believe that a good test should be:
- simple, not unnecessarily hard and complicated
- protect the claimant's privacy (and also the privacy of participants)
- if possible, somewhat interesting, not boring
If you want to design a test which meets all of these requirements, there are not that many possibilities left, in my opinion.