Virgin Births

It is the time of year when someone wants to always bring up the idea of the virgin birth of Jesus as described in the Christian Scriptures. This time it is a study published in the Christmas edition of the British Medical Journal that claims reports of virgin births are also a modern day phenomenon. The study, titled US researchers ponder modern day virgin births goes into considerable detail about the sex-education of the virgins and how many took a chastity pledge.

The study was conducted by researchers at the University of North Carolina at Chapel Hill using data from the U.S. National Longitudinal Study of Adolescent Health (Add Health). They found a reported 45 such births reported over a 15 year time span. This was an under one percent characteristic.

I liked the Reuters head line piece that noted near the end that the mothers in question were more likely to have boys than girls and be pregnant during advent. They did say that these observations were not statistically significant.

In fairness to the authors they do admit such measures are subject to some degree of respondent bias.

My questions – why does this kind of thing show up in the literature at all? Rather than tout measures of virgin births they need to be discussion issue of data quality. A little Google search revealed a paper using the same data source that did just that titled: An Exploratory Study about Inaccuracy and Invalidity in Adolescent Self-Report Surveys. They discuss two groups of respondents – those who make mistakes in answering the questionnaire and the jokesters.

These virgin birth are likely nothing more than data errors. If respondents is able to answer the survey questions in such a way that they can report a birth prior to any sexual activity they will do so. Some will do it in error and some will just be having a bit of fun at the expense of the survey authors. A good survey, if automated, should have consistency checks in place while the survey is conducted that will flag and prevent the errors. The editing process should also catch the inconsistencies.

But to publish a paper discussing the detailed characteristics of the group is something I expect to see on April 1st, not in the week before Christmas.