Even before Barack Obama’s long form birth certificate was released this week, stories were being told about the certificate number, claims that it was bogus because it was out of strict time sequence with another birth. Now that we have the long form, fresh from Hawaii, that possibility of a fake number can be ruled out; however, we have a new bit of information that might help explain the evidence.
We know these certificate numbers from August 4th and 5th:
- Nordyke, Susan – 10637 – 8/5/1961
- Nordyke, Gretchen – 10638- 8/5/1961
- Obama, Barack – 10641- 8/4/1961
- Waidelich, Stig – 10920- 8/5/1961
The list is sorted numerically by birth certificate number; however, you might also note that they are alphabetically sorted by surname and that they are not in order by date of birth. My hypothesis is a batch of certificates was alphabetized by surname and then fed into the numbering machine and that hypothesis is borne out with the new information provided by CNN of the Waidelich certificate number. Note that NORDYKE and OBAMA are close both alphabetically and numerically (just 4 off) and that WAIDELICH is not close numerically nor alphabetically. Armed with these examples, is it possible to estimate the size of such a sorted batch?
It might be possible to get a better sample, but I have a readily available list of 88,799 surnames of those in the United States from the 1990 census. Names between NORDYKE and WAIDELICH represent 27%1 of all individuals. The certificates between and including NORDYKE and WAIDELICH number 284, leading to an estimate of 1050 certificates in the entire batch (284 / 27%)2. Hawaii reported 17,616 births in 1961, or an average of 339 per week3. So the batch size, based on an alphabetization assumption, is almost exactly 3 weeks worth of births.
1The surname frequency table lists all surnames, but only the frequency of those names exceeding .001% of the population (about 18,000 names), representing 80% of all persons. The name frequencies added up between Nordyke and Waidelich were adjusted by dividing by 80%.
2The calculations were done to a higher precision and so the rounded math here will be a little off.