Virgos play basketball more than Taurus

[Ray Murphy]

Hi All,

If people who are born under the sign of Taurus and Virgo have equal
access to the sport of Basketball, then those with Sun in Virgo elect
to play that sport at a much higher rate than the Sun in Taurus
people.

See the attached progressive-graph in date order for the 502 people
listed in Astrodatabank Version 3 who were born in either of those
signs. There are 285 Virgo and 217 Taurus people born during a 66 year
period from 1913 to 1979.

The observation is statistically significant at about the 0.03 level
when a comparison is made between those 2 signs in isolation.

Noon times have been used for all births.

http://tropical-astrology-research.googlegroups.com/web/Basketballers.gif?gda=U2MV2kUAAAAR4o_U4Mb1ar5u1N08e_lzoV3BSWdITZ_Q3wqNGgyUGOG2sVwGLF-i21hNBTL4y6OkNTLimnrPyTN8pl8xhyRsrnbirdOKW0UubKy7Gu2GBQ&gsc=1rI-yAsAAABS4Ee1kLUl647TiZkjb1-K


Ray

[Graham Douglas]

Hi Ray,

              We all find Geoffrey Dean a bit abrasive but he does have it right on stats usually, have you got a copy of Recent Advances ? Have alook at pages 105 - 110 or I can xerox them if you like.

           As you say 0.03 on only one test, did you use 11 degrees of freedom ?

           Unless you have a reason to claim a priori a certain preference then you are testing the null hypothesis and you have to consider all 12 signs.

          With sports there is a special problem which comes from birth date in relation to time of entry to school. If a child is born in late september as in the UK) they will have a birthday shortly after starting school and will be a year older then many of their class mates. This will give them a strength advantage in sports.

           Regards,

              Graham.

[Graham Douglas]

[Ray Murphy]

RM: G'day Graham, until this morning I would have agreed with you
100% about the need to use 11 degrees of freedom for this sort of
thing, but after a few days of developing a small program that shows
me how much data is required to reach the .05 level (or better) under
various circumstances, I realised that we are not going to get very
far in the near future unless we change the way we have been working
for decades and either:
(a) Start gathering enormous amounts of data or
(b) Get rid of degrees of freedom that kill virtually every project.

I found that the latter was far easier to do, and all it requires is
for us to change what we are saying with useful some projects.

In this series of examples which I've started to post here, I'm
deliberately not saying anything about the 12 signs of the zodiac in
any given study, so they don't need to be accounted for with degrees
of freedom. Each observation that I am reporting is for only 2
factors, which *happen* to be signs of the zodiac, but I'm not
comparing any scores with any of them except the one mentioned.

I could easily have substituted "born in the sign of Aries" with "born
in the country of Australia" but if I had done the latter I wouldn't
be asked by any stats expert to use 194 degrees of freedom for the
other 194 countries in the world.

Now if we imagine that I *had* used Australia instead of Aries, and
I'd used New Zealand instead of the sign of Taurus for the observation
about the tendency to gravitate towards basketball, then the
observation stands alone.

I suppose it would stand alone just as much as if I had given a bunch
of subjects in an experiment "A Virgo pill" or "Taurus pill" and we
compared their attitudes towards basketball afterwards.

[By way of example]
The difference here is that I'm not saying that the "Virgo pill" works
better than the "Aries pill" or the "Gemini pill" or any other pills
because I didn't make any such pills. I only made the two - Virgo +
Taurus pills.

Usually of course when we are conducting surveys or experiments we are
saying or implying that a high-scoring sign is better than ALL the
others, but we don't need to do that :-)

Ray
[P.S. I got your snail mail - thanks]

[Ray Murphy]

On Apr 19, 8:24 pm, Graham Douglas <ondastropic...@yahoo.co.uk> wrote:

> Hi Ray,
> We all find Geoffrey Dean a bit abrasive but he does have it right on stats usually, have you got a copy of Recent Advances ? Have alook at pages 105 - 110 or I can xerox them if you like.

RM: I missed a few points in my previous reply. Yes, I've got Recent
Advances, thanks.

> As you say 0.03 on only one test, did you use 11 degrees of freedom ?
> Unless you have a reason to claim a priori a certain preference then you are testing the null hypothesis and you have to consider all 12 signs.

RM: I'm not familiar with all that formal language you've used but I
presume that's what I've done, so only one Degree of Freedom was used.

> With sports there is a special problem which comes from birth date in relation to time of entry to school. If a child is born in late september as in the UK) they will have a birthday shortly after starting school and will be a year older then many of their class mates. This will give them a strength advantage in sports.

RM: That's something to keep in mind, however I'm inclined to think
that it's partly height and energy thing. I've seen so many tall
Virgos along the way that I think there must be more of them that tall
Taureans.

Ray

[Graham Douglas]

Hi Ray,

            No, you need to have the counts for all 12 signs and use Df = 11, which means Chi-squared for p < 0.05 needs to be at least 19.6.

            Regards,

                Graham.

[Graham Douglas]

Hi Ray,

           Sorry but you can't just make things disappear when they don't fit, if you want to be taken seriously.

            I suggest you read the section in RA called " Significance of 1 result among many" which is on p.110 or nearby.

           Regards,

             Graham.

[Ray Murphy]

RM: I know what you mean by trying to make things disappear, but in
this case it's not that, but rather narrowing the survey to accomodate
the amount of data being used - but only for suitable cases.

For example in astrological research if I was comparing the
manifestation of one 'hard aspect' with another one, I wouldn't expect
to find much variation in the (broad) type of manifestation that had
been recorded for each of them, but I *would* expect (based on
experience in readings) to find a reasonable difference between
particular hard aspects and particular soft ones; so searches like
that should be done anyway, and when that happens they wouldn't be in
the realm of "one result amongst many".

I followed your advice and read page 110 again and it reminded me of
the common belief that "if you look in 1000 places you'll find
something that can occur 1 in 1000 times just by chance". I've been
meaning to question this for a few years, because in reality it
doesn't happen very often at all in astrological research.

Dean indicates on page 110 that it is easy to find 0.05 but I've found
with perhaps a million various graphs (with clickable stats) that's
not the case at all unless one is using a LOT of data - but if one is
doing that, then 0.05 becomes pretty useless anyway and .001 or better
should be the target.

Thanks for the advice anyway. I'll keep looking to see if I can see it
applying to any of the things that I process, but even if it *is* the
case sometimes or even all the time, my main aim is to find anything
that appears to be working consistently above the expected level,
rather than the 0.05 level, so we can drive the wedge in for some more
refined searches.

Ray