The most striking thing about the comments is that everyone seems to
have accepted that a grouping in 12 bins of continuous varible is OK.
Had the grouping been in months we might not have had the discussion.
To the extent that an outcome is believed to vary with the time of year
one is born, it would seem more appropriate to fit some continuous
cyclical model. The notion that risk changes abrubtly when passing from
Cancer to Lion is at the same level of credibility as the one that
asserts that it changes when passing from July to August. Yet the latter
is common in the litterature.
Bendix Carstensen
______________________________________________
Bendix Carstensen
Senior Statistician
Steno Diabetes Center
Niels Steensens Vej 2-4
DK-2820 Gentofte
Denmark
+45 44 43 87 38 (direct)
+45 30 75 87 38 (mobile)
+45 44 43 73 13 (fax)
[log in to unmask] http://www.biostat.ku.dk/~bxc
> -----Original Message-----
> From: A UK-based worldwide e-mail broadcast system mailing
> list [mailto:[log in to unmask]] On Behalf Of Bankart, Dr M.J.
> Sent: Tuesday, October 16, 2007 10:32 AM
> To: [log in to unmask]
> Subject: Re: Astrological star signs.......
>
> Any significant effect which was found might reflect one of the
> following:-
>
> 1. People's BELIEF in the veracity of an astrological system,
> rather than the putative (and surely spurious) effect of the
> planets on the individuals themselves, which latter process
> is also exceptionally poorly documented in terms of its
> (fictional) causal mechanisms. Such an analysis might be
> useful in determining the extent to which the astrological
> beliefs of different populations results in different
> outcomes (if this is indeed found to be the case). Religious
> and astrological beliefs were after all responsible to some
> extent for the demise of the Aztecs (together with smallpox,
> etc.). You are what you think to some extent.
>
> 2. A type I error (most likely).
>
> 3. Time of year effects.
>
> 4. Some other explanation.
>
>
> John.
>
>
> Dr. John Bankart
> Medical Statistician
> University of Leicester
> Dept. of Health Sciences
> 22-28 Princess Road West
> Leicester
> LE1 6TP
>
> e-mail: [log in to unmask]
> tel: 0116 - 252 - 5430
> -----Original Message-----
> From: A UK-based worldwide e-mail broadcast system mailing
> list [mailto:[log in to unmask]] On Behalf Of Adaikalavan Ramasamy
> Sent: 15 October 2007 18:20
> To: [log in to unmask]
> Subject: Re: Astrological star signs.......
>
> Dear Dr. Sorkin,
>
> I agree with you that the use of astrological signs may not
> be scientific and it is hard validating any such findings.
> However, the original poster was Richard Szydlo, who did
> indicate that it was a far fetched idea, not me.
>
> I merely provided one possible solution to convert the dates
> to astrological signs as I found the technical challenge
> interesting. The solution provided here is useful for valid
> purposes as well. For example, one can create an anniversary
> or birthday reminder list using only Steps 1-3 below.
>
> Yes, I failed to to comment on the inappropriateness of
> Richard's proposed approach and like to thank you for
> pointing that out.
>
> Regards, Adai
>
>
>
> John Sorkin wrote:
> > Adaikalavan,
> > Let us assume that you find an association, what will you say about
> > it? Any significant result you might find is undoubtedly a type I
> > error. It is an abuse of statistics to test a clearly incorrect and
> impossible hypothesis. By doing so, you give a patina of
> truth to pseudoscience, and if you report a significant
> finding diminish the trust the public will have in legitimate
> statistical analyses. You might have a legitimate reason to
> perform the test you describe (such as Prof.
> Peto's test of astrological sign and outcome that was
> performed to demonstrate to journal editors, and journal
> readers the hazards of ill-conceived post-hoc subgroup
> analyses) but you have not given one in your posting. I urge
> you to look (and think) before you leap . . .
> > John
> >
> >
> > John Sorkin M.D., Ph.D.
> > Chief, Biostatistics and Informatics
> > University of Maryland School of Medicine Division of Gerontology
> > Baltimore VA Medical Center 10 North Greene Street GRECC (BT/18/GR)
> > Baltimore, MD 21201-1524
> > (Phone) 410-605-7119
> > (Fax) 410-605-7913 (Please call phone number above prior to faxing)
> >
> >>>> Adaikalavan Ramasamy <[log in to unmask]> 10/15/2007 8:31 AM
> >>>> >>>
> > You don't have to do 600 cases by hand. If this is a one
> off, then it
> > would probably be faster to do by hand than a macro. Here is how:
> >
> >
> > 1. Make sure your dates are stored as Date type not text
> etc in Excel.
>
> > (highlight the column and check format cells).
> >
> >
> > 2. Find out the birthdays (i.e. the year is now irrelevant) for all
> > cases. You can do this in many ways. For example, by creating an
> > adjacent column with the formula "=DATE(2400,MONTH(A1),DAY(A1))"
> > assuming that you stored your date of births in column A1.
> It is fine
> > to store this column as either general, number or date.
> >
> > Note: I choose year 2400 above as it is a leap year
> allowing for 29th
> > February. And why 2400 and not 2000 or any other leap year?
> Just as a
> > precaution in case someone accidentally uses this as the date of
> births.
> >
> >
> > 3. Sort the entire sheet using the birthdays column.
> >
> >
> > 4. Cross reference these dates with your astrological star
> dates and
> > fill in the star sign. For example, all birthdays that fall between
> > September 24th and October 23rd is a Libra.
> >
> >
> > BTW, I am assuming the zodiac dates do not change from year to year.
> >
> > Hope this helps.
> >
> > Regards, Adai
> >
> >
> >
> > Szydlo, Richard M wrote:
> >>
> >> I know this is far fetched, but I would like to see if a patient's
> >> astrological star sign is in anyway associated with survival......
> >>
> >> Does anyone have a macro for SPSS / Excel / SAS / Anything else -
> >> that will convert a date of birth to the appropriate astrological
> >> star sign.... I could of course do it by hand, but 600
> date of births
>
> >> will take me a long time!
> >>
> >> Thanks,
> >>
> >> Richard Szydlo
> >> Imperial College School of Medicine
> >> London, UK
> >>
> >>
> >>
> >
> > Confidentiality Statement:
> > This email message, including any attachments, is for the
> sole use of
> the intended recipient(s) and may contain confidential and
> privileged information. Any unauthorized use, disclosure or
> distribution is prohibited. If you are not the intended
> recipient, please contact the sender by reply email and
> destroy all copies of the original message.
> >
> >
> >
>
|