And the moon sinks low in Scorpio

[venta]

This is something I've been meaning to experiment with for ages, but never got round to. I've just been reminded of it, because someone posted on a thread elsewhere with the assumption that "birthdays are randomly distributed" - which I've always assumed they're not. For a start, there's always a bit of a clump of birthdays round February/March, which I've always blamed on people having early summer holidays in June.

Now, I'm assuming that unlike, say, musical taste, incidence of birthdays among my friends list will at least be representative of the population as a whole. If not, then we're into the suggestion that being born in a particular month predisposes you to particular behaviours and that smacks of astrology, so I don't like it.

(Yes, I know there are other non-astrological factors which might be relevant, but let's give it a whirl, shall we?)



You're getting no tickyboxes for this, dammit, it's radio or nothing. I refuse to allow people to be born in more than one month.

[Poll #424559]

Comments

Re: Some more data points...

[lnr]

Yeah, I was surprised to note how many of the people on my friends list don't have their birthdays listed. Maybe they just don't like giving out personal details on the internet.

I wonder too how many people read because other people have mentioned something, or they read it on a friendsfriends page, or whatever, and hence might fill in a poll like this.

Re: Some more data points...

[venta.livejournal.com]

I was rather hoping extra non-regulars might happen past and fill it in, yes.

Also, people on my friends list and people who read this journal regularly are non-identical subsets of LJ users :)

Re: Some more data points...

[wimble.livejournal.com]

From our database:

Month          Count   Bar Graph
-----          -----   ---------
January        19129   1234567890123456789
February       17797   12345678901234567
March          20142   12345678901234567890
April          19694   1234567890123456789
May            20888   12345678901234567890
June           19346   1234567890123456789
July           19857   1234567890123456789
August         19670   1234567890123456789
September      19377   1234567890123456789
October        19208   1234567890123456789
November       17994   12345678901234567
December       17951   12345678901234567
NULL            8086   12345678

Total:        239139

Which isn't scientific, but I would think is good enough to be indicative.

Re: Some more data points...

[venta.livejournal.com]

Thanks, that may actually be an even larger sample size that one can grab from an LJ poll :)

Now I suppose I have to dig out the A-level stats, and attempt to remember how to do proper significance testing. Ouch.

Re: Some more data points...

[lanfykins.livejournal.com]

I have a copy of Statistics Without Tears somewhere if that helps... ;)

Re: Some more data points...

[venta.livejournal.com]

Thanks. I'll get back to you if I start feeling particularly tearful :)

Re: Some more data points...

[wimble.livejournal.com]

Well, I just looked at the flatness of that righthand margin, and thought "nah..."

The sample isn't, of course, perfectly random: it includes all our staff, which probably is random, but also all our students for the past N years. For a single year, there are probably peculiarities introduced by the period of the academic year, simililar to the school year (eg. hypothetically, we could get fewer students who's birthday is late in the school year, as they decide to defer). I'm hoping that by simply hoovering many years, those boundary effects will cancel out.