APPENDIX 1

Statistical Tests

Except for comparing average ages of medallists, I have used the standard Chi Squared test throughout. Where there was an a priori reason for thinking that a sign would have a positive effect on the numbers of medallists for that sign, I used a one‑tailed test to obtain a probability value. For example, it seemed reasonable to expect that Pisces medallists would have significantly more rather than significantly less medals in Swimming and Water Polo. In cases where there was no obvious reason to expect a significant difference before doing the test, I used a two-tailed test.

It can be argued that the probabilities stated in the test are not “real” probabilities because of the numbers of tests needed to obtain them. This may be a valid argument in some two-tailed tests but cannot possibly be used to explain away the extremely high probabilities obtained in so many cases.

And it certainly cannot be used to dismiss the Pisces effect.

The test on Pisces against Swimming was the first one I tried, since it is the most obvious association to make. Pisces against Water Polo is probably the second most obvious choice of association

There is another aquatic discipline called Synchronised Swimming that may appear to be a candidate for association with Pisces but closer inspection reveals that this is not the case. Synchronised Swimming is actually a misnomer as the athletes scull rather than swim and are judged more on their artistic interpretation than any other factor.