So last night I was reading another thread on here about the Chinese Gender Chart. I decided to see how accurate it was, so I called on my wonderful Facebook friends to volunteer as many birth dates as possible to come up with a percentage of right and wrong.
After 3 hours of finding volunteers, I ended up with 50 birthdays, which is a pretty good control group for a statistical experiment.
Out of 50 birthdays, the test predicted 43 of them correctly, 7 incorrectly. 43 divided by 50 is .83, making the test 83% effective according to my experiment.
Of course, as we know this number could have changed based on how many people I used, but i noticed a pattern as well. i would get a long string of correct predictions and then a block of about 2 or 3 incorrect all at once.
Another interesting note, the times it was incorrect was always for siblings, for example:
I asked my friend for his and his siblings birthdays, it was correct for him, incorrect for both of his siblings. Same thing with my friend Kayla, correct for her, incorrect for both siblings. It was also incorrect for my aunt and both of her kids, but correct for her brother. I doubt that has anything to do with the test but i thought it was interesting. It was never wrong for people who provided me a single birthday.