That’s not what the abstract said. It said “the first of the month and days divisible by five”. So 1, 5, 10, 15, 20, 25, 30.
Also, human birthdays aren’t evenly distributed. People tend to have sex around holidays and cluster somewhat. Late August through Early October are popular, and April and May less so.
So the statistical significance of the “round number” no doubt exists, but may not be as significant as it sounds to normal modern people in Western society. My grandfather, for example, was reported to the authorities 8 weeks as his country of birth had a poll tax… in his country baptism records are more accurate than civil birth records.
If that's the cause, then you'd expect to see the same patterns in all birth records of people born in 1923 who died, for example at the age of 50 in 1973, as you would in those from the same cohort who are still alive today.
So if you took samples of birth records for people born in 1923 who died in 1953, 1963, 1973... and so on, does the pattern hold? If these patterns of birth records are indeed evidence of fraud, you would expect to see a difference for people born in 1923 who are still "alive" today and those who died last century.
You see similar patterns in many populations with the convergence to even numbers.
Typically, it’s old people struggling with memory or folks who immigrated. In the United States, there are >10,000 jurisdictions issuing vital records, many of which are less than reliable.
Fraud is a strong word and inaccurate connotation. Are people lying about their age to scam pension or other benefits? Sure. But that is just one of many frauds in this area. Folks will fail to report deaths, steal the identities of dead children or any number of things.
What do you think that will tell you? For example it seems to me that people born in more rural area a century ago are both less likely to have accurate birth records, and less likely to have good access to health care. There are probably many other covariates that would need to be controlled for.
She only knew that she was born in the spring in "the year of the big flood" which neither she nor local records could identify.
When we brought her to city hall, the just asked us what birthday she wanted and she picked her favorite number in her favorite month in spring and we guessed a year which was probably +/- 2-3 years correct.
