There have been numerous discussions here regarding the certificate number on President Obama’s birth certificate and how that number relates to the numbers assigned to the Nordyke twins who were born about the same time. I’ve written articles before including:
- Date Filed v Date Accepted: APPEALED!
- Birther math (part 5)
- Birther math (part 6)
- Where is the Nordyke birth announcement? (update)
- Obama Certificate Number Proves It’s a Fake – Busted
I have used my background in public health and vital statistics information processing to attempt to describe what seems to me the most likely workflow for the certificate process based on my experience, sample certificates, information from state spokesperson Janice Okubo, and logical deduction. I begin by noting some facts and assumptions. (If you don’t want to read a long and ponderous article, you could skip down to the last paragraph and get the main point.)
I assume that the Kapi’olani Maternity & Gynecological Hospital had someone on staff designated as a local registrar by the state health director. We know from legislation of the period that the Board of Health could appoint local registrars. We also know that one certificate from 1963 for the Tripler army base in Honolulu was signed by an Army officer as local registrar. It stands to reason that if the Army base had a local registrar, then the main birthing hospital in the state would have one too.
We know from state’s spokesperson, Janice Okubo, that file numbers were applied at the central office, and not by the local registrar, or pre-numbered on the form. I have placed the numbering towards the end of the process to facilitate handling, because once the number is assigned, the certificates have to be kept in order for filing.
I am assuming that Kapi’olani hospital sent certificates to the Bureau of Vital Statistics on a daily basis. This assumption is plausible given the volume of births involved and the single example of the Nordyke twins certificate where the local registrar date is the same as the registrar general date (and the same date the doctor signed the form). Further, the fact that Obama and Nordyke announcements appeared at different times in Honolulu newspapers speaks against them both being sent to the Bureau of Vital Statistics on the same day.
Following is my draft processing flow:
- Hospital admission
- Child Delivered
- Mother (typically) completes worksheet
- Hospital clerical staff types original Certificate of Live Birth based on information from the worksheet and the medical record.
- Mother (or whoever completed the worksheet) signs Certificate
- Physician signs Certificate
- Local registrar (probably a Hospital employee) signs a batch of certificates and stamps Date Accepted by the Local Reg.
- Certificates are batched (daily in Honolulu) and sent to the Bureau of Vital Statistics.
- The Bureau of Vital Statistics verifies that the Certificate is in order
- The Certificate is entered for transmission to the Hawaii Newspaper Agency (daily? batch)
- The Date Accepted by the Reg. General is stamped on the form and the form is numbered by a numbering machine (batch)
- The Certificate is entered into a log for indexing
- The Certificate is filed by number into a volume and shelved
Now comes some educated guesswork on how Barack Obama’s certificate fits into all of this. There are three things we know for sure that constrain the possibilities. Barack Obama was born the evening of Friday, August 4, 1961; some registrar signed his certificate on August 8, 1961, and a newspaper published the announcement on August 13. Given that Obama was born on a Friday night, it seems reasonable that the typed certificate was not ready for the doctor’s signature until Monday the 7th at the earliest, and a signature date of August 8 by the physician is reasonable. This suggests that the “Date Filed” on President Obama Certification of Live Birth (August 8 ) is actually the Date accepted by the local reg. (there is no “Date Filed” on the actual birth certificate form). Obama’s certificate would then be ready to go to the Bureau of Vital Statistics on August 8 (Tuesday) and most likely completed by them August 11 (Friday) for inclusion in the Sunday newspaper (August 13). This puts a maximum processing time at the Bureau of Vital Statistics at 4 days. Obama’s certificate was numbered 10641.
Now let’s apply the same process to the Nordyke certificate. In this case, we have more dates. The Nordyke twins were born on Saturday, August 5, in the afternoon. Mrs. Nordyke signed the form August 7, the doctor signed the form Friday, August 11 and both . The newspaper announcement appeared on August 16. Susan Nordyke received certificate 10637.
Based on the fact that all the Obama dates are before the Nordyke dates, one would think that Obama’s certificate should have a lower number. It doesn’t, but not by much. There is a difference of only 4 between the two. One then asks: how is it that two certificates that presumably arrived and processed on different days (based on the fact that they appeared in the newspaper on different days) got numbers that close?
Here I have to speculate, but in all honesty I feel pretty good about this suggestion (and frankly, this is the first time I really feel comfortable about a conclusion on this topic). One important aspect of the processing is making the records available to the public, and this means indexing them by name and date of birth. Another factor is the need to file records efficiently by certificate number. Filing by number, means assigning the number at the last possible moment in the process so that it is easy keep the records in order to put the records into volumes by number without having to remove pages and insert pages in the book. Indexing the records for public access means maintaining an alphabetic index. Records have to be insertable into the alphabetic index because of delays in receiving certificates from the other islands and because of late registrations. The easiest way, however, to merge records into a sorted index is by sorting the batch of records to be inserted first. So here comes the good part. If, as I believe, a batch of records was numbered on or after August 11 (a batch containing both Obama and Nordyke), then it would have simplified the task if those certificates had been alphabetized before numbering. An alphabetized stack would be Nordyke…Obama and Nordyke would have a slightly lower number, which is what we find is the case. The two certificates cannot have been numbered on separate days or else they would be more than 4 apart (on average, there were 33 babies born per day in 1961).
This then leads us to a revision to the workflow, the addition of step (11a) “Certificates in batch are alphabetized by last name of child.”
This was a very long article to arrive at some simple conclusions. The fact that the Obama and Nordyke certificate numbers differ only by 4 implies that they were numbered on the same day. If they were numbered on the same day then most any processing variation could have influenced the order. The fact that the numbers are very close, and that “Nordyke” and “Obama” are alphabetically very close, suggests that the forms were alphabetized before numbering.
[Update: later information including additional examples suggests that the certificates were alphabetized in monthly batches before numbering, and possibly that they were first grouped by birthing facility.’