Wednesday, July 04, 2007

 

Birthday Distribution

How do you think are birthdays distributed over the days of the year? Do you think certain months (days) will have significantly more (less) births than others?

Here is the distribution in the United States. (Source Hallmark)

Note that the expected value for months with 31 days is 8.49% and for months with 30 days is 8.21% and 7.67% for february.

Rank Month Percent
1. August 9.07 (+0.58)
2. July 8.80 (+0.59)
3. September 8.62
4. October 8.60
5. March 8.51
6. May 8.30
7. January 8.25
8. June 8.15
9. April 8.12
10. December 8.07 (-0.42)
11. November 7.96 (-0.25)
12. February 7.55 (-0.12)

Given the data, Do you think this is a random distribution, or is there a bias?

Ok here are two more facts:

October 5 is supposedly the most popular birthday and May 22 is the least popular. (source wikipedia)

I seem to remember this as a joke, but it is worth noting that October 5 - 9 months = approximately New years eve!

Looking at the table again, the birthdays probably are not randomly distributed - seems like there is a bias caused by more conceptions happening in colder months??

Comments:
imo, i think < 1% is statistically insignificant here(not to mention the err in data collection), unless u have these figures in each yr(as opposed to one final avg) and see the same pattern over years.
i think for india or esp. tamil nadu, there mite be some uneven distribution as few months of year have most auspicious days(marriage days i mean) and many indian couples give birth in the same year.so seeing the distribution for marriages(shud be very uneven) and birthdays mite be intersting to see.
 
There are several statistical tests to check if a distribution is random -- does student's t test ring a bell?

Intuitively, looking at the table I do believe there is a bias. Note that there is some locality in the rankings Eg.(July-August-September-Oct) and (Nov, Dec). This kind of locality usually indicates bias.

As you mention, it will be really interesting to see the distribution in places like TN, because of (in)auspicious days. Although, in tropical countries like India, I would expect weather to play a smaller part.
--
Vijay
 
i didnt think abt the weather. although,1 thing to consider when talking abt weather is, in US most couples(I guess) are conscious of when to have babies dep. on personal life, than on the fact that sex=>baby(i.e if cold weather causes more sex, need not necc give more births in 9 months later, although from the locality u mentioned it looks like :)).
in india, most couples tend to have babies in 1 yr of marriage. so def sex=>babies
 
I see a bias and especially during the winter season. Except for few states in US, the rest have real snowy winter. Its hard to drive and to get any emergency medical assistance during the winter owing to the heavy snow.

I think this is common sense ;-)
 
madhu,

So you are saying commmon dicates the proceedings at the time of conception...I doubt it :)

Just kidding...what you say is quite possible.
 
hehe, october fits in TN too. Thai masam + 9 = october.
 
Hey dude. Just run a chi-square test to see if it's random or not.

http://en.wikipedia.org/wiki/Pearson's_chi-square_test
 
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?