Can I call you Father?

There is a lot of noise being made by the recent report that up to 4% of all fathers are raising another man's child. I'll quote before commenting:

"Professor Mark Bellis and his team said that the implications of so-called paternal discrepancy were huge and largely ignored, even though the incidence was increasing.

In the US, the number of paternity tests increased from 142,000 in 1991 to 310,490 in 2001.

Vital information is being delivered to people without very much thought about how it is going to affect them

Lead researcher Professor Bellis

Demand for testing has grown by a factor of 10 in the last decade in the UK, according to University Diagnostics, Teddington.

The current level in the UK is somewhere between 8,900 and 20,000 tests per year.

About 5,000 of these tests are instigated at the demand of the Child Support Agency to resolve who should be paying child maintenance.

Others are done to investigate inherited health disorders and others for social reasons.

The Liverpool team found that rates of cases where a father was not the biological father of his child ranged from 1% in some studies to as much as 30%.

Experts have generally agreed that the rate is below 10%, with a 4% rate meaning that about one in 25 could be affected.

However, increasing use of genetic testing is likely to boost the rates of paternal discrepancy, say the authors."

Medical statistics, especially those that make news headlines, are often the result of very limited studies. These statistics are often lacking the sort of controls we'd expect to see in modern linear regression within the social sciences. In fact, a simple correlation is often enough to make the front page of a newspaper.

It is worth noting that this number: 1 our of 25 fathers, cannot apply to the population of the UK, nor any other country. The sample that the fraction was derived from is the sample of all people who have either chosen to have paternity tests or have been ordered to by the Child Support Agency.

If we really wanted to know whether or not this statistic applied to the overall population, it would have to be a result of a study with random sampling. There is too much self selection with the current sample as parents who already concerned enough to get fertility testing are more likely to have been victims of infidelity. It is even arguable that people fighting over child support are more likely to have experienced an 'illegitimate' birth.

Rantings over statistical accuracy aside, I don't see many ramifications resulting from a 4% rate of false fatherhood. This is pure conjecture, but I'd imagine that this number is at a historical low, as the existence of modern contraception and the disincentive to infidelity of paternity tests have undoubtedly reduced this occurrence.

However, on a case by case basis, I can see how earth-shattering a discovery could be.