A very unusual coincidence occurred in the Purkey families, Mr. and Mrs. C. F. Purkey announce the birth of a son, born June 17, 1947.This is the third son born to these parents on this date. The first, Jimmie Marvin, was born June 17, 1941; the second, Chester Floyd Jr., June 17, 1943, and the third, Harold Edward, June 17, 1947.
Mrs. Purkey is a daughter of Marvin Downing of Wilmore and Mr. Purkey is a grandson of Mrs. Carsten Nickelson, also of Wilmore.
The above news article was featured under the title of "Stranger Than Fiction" in The RootsWeb Review: RootsWeb's Weekly E-zine 11 January 2006, Vol. 9, No. 2. Myra Vanderpool Gormley, Editor, quoted a comment I made when I sent the story to her: "Imagine how skeptical a family researcher would be upon seeing a family group sheet giving the same day of different years as the birth dates for all three sons in a family!"
A few days later, RBeda Coffey signed the Comanche County, Kansas, site guestbook: "The story of the sons is interesting. I have two sons born on the same day 7 years apart."
I also heard from someone else with a story that must have seemed like an amazing coincidence of birthdays to the children of one family: "I just read your post in the Jan 4th Rootsweb e-letter. It reminded me of a story a co-worker told me. Her grandparents were immigrants and did not speak English well. They celebrated all their kids birthdays on the same day of the month, the day her grandfather got paid. None of the kids knew when their real birthdays were."
Just today, the following comment by a reader was published in RootsWeb Review: 8 February 2006, Vol. 9, No. 6, in response to the story about the Purkey Brothers:
Rare or Common Happening? By Madeline De Long
"The letter "Stranger Than Fiction" (RootsWeb Review, 4 January) in which three sons within the same family were born on the same date on three different years, could have easily been referring to my father's first cousins in Texas.
The first three boys were born on the same date, every other year. The fourth broke the pattern and was born 15 months following the third.
I'm curious about the statistical likelihood of such occurrences. The reality is that wonders such as this do actually occur."
Numerically Speaking By Neal F. Jordan
Madeline De Long inquired about the odds of the first three children in a family being born on the same date every other year.
To simplify the calculation, let's not worry about the odds of their being born every other year but just assume that is so and then ask about the odds of it being the same month and day.
The first child has to be born on some day, so the probability of that is 1, a certainty, as long as we don't specify the date in advance, like insisting that it be the fourth of July.
Assuming that any day in the year two years later is equally likely, means that the probability of the same particular day is 1/365 (one out of 365). The odds are 365 to 1 against it.
The probability of the third child being born in the year four years later on the day in question is again 1/365. To get the probability of both the second and third children being born on the day in question, multiply the two previous probabilities to get 1/133225.
Stated another way, the odds against this happening are 133,225 to 1 against it. If you want to specify the date before any of the children are even thought of, multiply by another 365 to get odds of 48,627,125 to 1 against your prediction coming true.
Since Ms. De Long insists it happened, it is likely it did and that other factors are at work, such as the parents habitually having a really good time on New Year's Eve or anniversary of their own choosing.
Looking at it another way, though 133,225 to 1 against seem like long odds, they suggest that this ought to happen on average in 1 in every 133,225 families that have three or more children, with the first three having birth years two years apart. Dividing 133,225 into the number of such families in the U.S. suggests that there ought to be several cases in each of our more populous states.
One further observation is that had the children been born at only one year intervals the odds against coincident birthdays would have dropped dramatically, since ruling out the first nine months after the first child's birth reduces the number of available days in the next year to the extent that the nine months laps into the next year. By the above reasoning each probability is then 1/N with N likely smaller than 365. The overall odds against are (N times N) to 1, a situation more favorable than 133,225 to 1 against.
Previously published in RootsWeb Review: 1 March 2006, Vol. 9, No. 9.
You can subscribe to RootsWeb Review, a free online newsletter, by visiting the RootsWeb Newsletter Management Center: http://newsletters.rootsweb.com/.
-- Jerry Ferrin
Thanks to Shirley Brier for finding, transcribing and contributing the above news article about the Purkey Brothers to this web site!
This RootsWeb website is being created by Jerry Ferrin with the able assistance of many Contributors. Your comments, suggestions and contributions of historical information and photographs to this site are welcome. Please sign the Guest Book. This page was last updated 08 February 2006.