A few years ago I found myself in an ABC Radio National studio discussing misattributed paternity – those traumatic and sometimes humiliating cases where a child’s biological father is not who he or she thinks he is. I had been invited onto the program, Australia Talks Back, because I was researching the social implications of new technologies for testing blood and DNA, which were making it possible to uncover mistaken paternity. Most of my research had involved interviews with the entrepreneurs who had established the emerging paternity testing industry in Australia and the United States.
The first question I was asked concerned the extent of misattributed paternity. We don’t know the true figure, I said, for the simple reason that it’s impossible to test a genuinely random sample of the population for paternity. There are always some people who will refuse to participate in random studies. This isn’t a problem if those who refuse are no different from those who do participate, but it would certainly be a problem for a study of this kind, because women with doubts about who fathered their children would be the most likely to find reasons not to participate.
Almost immediately an irate listener called to ask where on earth the program obtained its “experts.” He referred confidently to a British study which showed that misattributed paternity affects more than 30 per cent of the population. By this reckoning, more than three-in-ten people are mistaken about the identity of their father. Later in the program an interviewee from a fathers’ rights group reiterated the claim that misattributed paternity is widespread in Australia.
I did many similar interviews around this time, and they all followed the same course. Interviewers wanted to know the extent of misattributed paternity and I invariably said that the answer was unknown. Yet my interviewers invariably found other “experts” who declared unequivocally that misattributed paternity is widespread. Most of them said that 10 per cent of the population was mistaken about the identity of their biological fathers, but sometimes they said the figure was as high as 20 per cent or even 30 per cent.
I was a little annoyed by this experience, so I set about reading every piece of available research on the subject, including the British study that allegedly showed a rate of 30 per cent. I also kept close track of new findings in the field, and did my own research. In the process I found out quite a bit about how an urban myth was born and transmitted around the world, and how it helped create a new industry.
The stubborn figure of 30 per cent comes from the published transcript of a symposium on the ethics of artificial insemination that was held nearly forty years ago, in 1972. “We blood-tested some patients in a town in south-east England,” Dr Elliot Philipp told the symposium, “and found that 30 per cent of the husbands could not have been the fathers of their children…”
At this point Dr Philipp was interrupted by a judge, who observed that “surely the figure of 30 per cent must be a minimum?” The judge clearly understood that while blood tests could definitively exclude paternity, they could not definitively establish it. This is why experts in paternity testing generally speak of an “exclusion rate” rather than a “non-paternity rate” or “misattributed paternity rate.”
Dr Philipp agreed that the figure was “a minimum.” He then explained how he came to be doing the tests. His team was “screening some female patients by testing their husbands for their blood groups” as part of a study about the formation of antibodies. The results surprisingly showed that “30 per cent of the children could not have been fathered by the men whose blood group we analysed.” Another participant asked about the number of people who were tested. Dr Philipp replied, “Not large – between 200 and 300 women – but large enough to give us a real shock.”
As we’ve seen, this brief conversation took on a life of its own, despite the fact that Dr Philipp never published the findings of his study. As a result, his precise tests and his population sample were never identified. One participant at the symposium described the sample as “highly biased,” but we can only guess what this means. It might mean that the sample consisted of unmarried mothers, who were an easy target for medical studies at the time. Later studies leave no doubt that misattributed paternity is much more likely for unmarried mothers than married mothers.
More generally, the fact that the findings were never published means that they were never independently evaluated by other experts. Ironically, during the 1940s, 1950s and 1960s there had been refereed studies about paternity in Britain and the United States, based on blood testing and published in reputable journals. None of them came close to 30 per cent.
Dr Philipp’s findings survived the test of time simply because they are shocking. Talkback radio gave them a new lease of life, and so has the internet.
Since Dr Philipp’s study, testing for paternity has become much more sophisticated. DNA techniques invented in the 1980s mean it is now both cheaper and more accurate – so much so that the accuracy of these tests is now more than 99.99 per cent, a pretty reliable basis for establishing paternity.
As a result, an industry dedicated to paternity testing has emerged, pulling in people from all kinds of related businesses. One of the entrepreneurs I interviewed had his start in a small bird-sexing business in the outer suburbs of Melbourne. One day, he recalled, “the phone rang and I picked it up.” When the person on the other end asked whether he tested for paternity, he replied, “No, we don’t.”
The next day there was another call: “Do you do testing of paternity?”
“No, we just do animals.”
The next day there was another call: “Do you do testing of paternity?”
“Can you hold on one minute?” the entrepreneur asked. He put down the phone and said to his business partner, “Can you believe this is the third phone call I’ve had this week? Somebody wants paternity testing.”
And his partner said, “We’ll do it!”
Most of the demand for paternity tests comes from mothers who want to establish paternity in order to secure child support. Some of it comes from fathers – often divorced or separated – who want to disprove paternity in order to avoid child support. In the United States more than 400,000 paternity tests are carried out every year. In Australia the figure is about 10,000, but only half of them involve Australian citizens; the balance are “export” cases, with offshore customers sending their specimens to Australian laboratories for testing.
These figures translate into more than five times as many tests per 1000 births in the United States than in Australia. One of the main reasons for the difference is that in Australian law the “marital presumption” that a husband is the father of his wife’s children also applies to de facto couples. (This was introduced without fanfare back in 1975.) In the United States the marital presumption doesn’t apply to de facto couples, which means that the paternity of children born outside marriage must be established legally. This routinely leads to paternity testing.
In 2008 the non-paternity rate reported by paternity testing laboratories in the United States was 25.9 per cent. My survey of a selection of Australian laboratories for the same year arrived at a non-paternity rate of 23.7 per cent.
The problem with these figures is obvious. The participants are not a random sample of the population. On the contrary, they are a group of people who have doubts about the paternity of a child or children. The main thing we can say on the basis of these figures is that about three-quarters of people who have some reason to doubt paternity will find that their doubts are unfounded.
Perhaps the most surprising thing about the figures is that the non-paternity rate in Australian laboratories appears to be lower than the rate in American laboratories. Given that Americans are five times more likely to have a paternity test at all, this suggests that the extent of misattributed paternity is much lower in Australia than the United States.
For the best available evidence about the extent of misattributed paternity we need to turn to medical research. This research is motivated by a medical condition or treatment, such as cystic fibrosis or bone marrow transplantation; the discovery of misattributed paternity is an unintended consequence. Occasionally researchers – like Dr Philipp – report their unintended findings; fortunately, most of them report their findings in more detail than he did.
Some scholars have aggregated all this evidence to produce an average “non-paternity rate.” The problem with doing this is that the quality of reports varies widely, as illustrated by Dr Philipp’s findings. It makes more sense to identify the best studies – those that fully explain how they came to their conclusions.
The best British study, published in 1991, suggests a non-paternity rate of about 1 per cent. A 1992 French study indicates a rate of 2.8 per cent. A 1994 Swiss study has a maximum rate of 0.78 per cent. A 1999 Mexican study comes in at 11.8 per cent. And the best North American study, published in 2009, proposes a rate between 1 and 3 per cent. There are no published Australian studies.
These figures could all be distorted by the problem of who is willing to participate, which I mentioned earlier. Even so, there are some striking differences between countries. The rate of non-paternity in Mexico is especially high and the rate in Switzerland is especially low. Almost certainly there are underlying cultural differences in how marriage, sexuality and parenting are organised in these countries, which shape these different rates.
Following this logic, it seems likely that the Australian rate is in the same ballpark as rates in Britain and the United States. Australia has more in common with these countries in terms of how family relationships are organised than with other countries, such as Mexico and Switzerland. By this reckoning, the rate of misattributed paternity in Australia would be somewhere between 1 and 3 per cent.
More precisely, the rate of misattributed paternity in Australia is probably closer to 1 per cent. By most measures family relationships in Australia have more in common with those in Britain than in the United States. That figure is also consistent with the evidence from the paternity testing industry that Australian rates are lower than those in the United States.
There is one other source of evidence on misattributed paternity. A succession of large-scale representative sex surveys were launched in rich countries in the wake of the HIV/AIDS crisis. They included questions about multiple sexual partners, which are a necessary condition of misattributed paternity, and so they might provide a vehicle for independent estimates of misattributed paternity.
I have done my own calculations based on British surveys in 1990 and 2000, for which raw data is available. They indicate an underlying non-paternity rate for children born in 1990 of somewhere between 0.7 per cent and 2 per cent. The estimated rate differs widely according to the marital status of the mother. For the offspring of married women, the rate is between 0.3 per cent and 0.6 per cent. For cohabiting women, it is between 1.1 per cent and 2.7 per cent. For other women – single, divorced, separated or widowed – it is between 2.3 per cent and 8.1 per cent.
My calculations for the survey in 2000 indicate an underlying non-paternity rate for children born in that year of somewhere between 1.3 per cent and 3.4 per cent. The non-paternity rate rose between 1990 and 2000 for two main reasons: first, a higher proportion of women (married, cohabiting or other) with two or more sexual partners, and second, a higher proportion of births outside of marriage.
Published data from the 2001 large-scale sex survey in Australia suggests that Australian rates are in the same range as those in Britain. Australian women are slightly more likely to have had two or more sexual partners in the previous year, which implies a higher non-paternity rate. They are slightly less likely to have children outside of marriage, which implies a lower non-paternity rate. In other words, the two patterns roughly cancel each other out.
At the very least, these surveys indicate that the extent of misattributed paternity is increasing in rich countries such as Australia, largely because of the weakening hold of marriage on sexual behaviours. Yet the increase is taking place from a low base. The evidence from sex surveys is pretty much the same as the evidence from medical research. It shows that estimates of 10 per cent, 20 per cent and 30 per cent non-paternity rates are massively inflated.
The really interesting question, then, is how the urban myth of rampant misattributed paternity persists, despite all the evidence to the contrary.
At least three parties have promoted the myth. The first of these are the fathers’ rights groups who have mobilised around DNA paternity testing as part of a broader campaign against the child support system. They believe that the system is stacked against them, and that paternity testing can correct this bias a little. They cite high rates of non-paternity to support their claims of widespread paternity fraud at the expense of fathers. They are especially active on the internet, which provides a medium for the rapid spread of such claims.
Second, there are the DNA paternity testing laboratories and their agents. The industry in the United States is especially large, and includes specialist laboratories with dedicated media units and call centres, and independent brokers who recruit customers on the internet and on-sell the tests to laboratories. The US industry can’t do much to increase demand for the tests among single mothers, but it can take active steps to increase demand in the other main market, alienated fathers who are often paying child support.
These steps include promoting the view that misattributed paternity is widespread. Brokers and laboratories promote this view through the internet (including links with fathers’ rights groups) and the media (including live television shows where paternity disputes are played out in front of a studio audience).
Finally, evolutionary psychologists provide intellectual credibility for inflated non-paternity rates. Evolutionary psychology explains human behaviour in terms of our genetic code formed in deep ancestral time. It prides itself on its scientific approach, in contrast to what it calls the “standard social science paradigm.” Specifically, evolutionary psychologists argue that while men’s short-term sexual strategy is based on obtaining large numbers of partners, women’s strategy involves obtaining men of “high genetic quality.” In close connection, men are “hard wired” to take care that they do not raise the progeny of other men.
Evolutionary psychologists believe that high non-paternity rates provide independent evidence of their theory of human behaviour. They are responsible for very badly designed studies that arrive at high estimates of misattributed paternity. They also coordinate meta-studies that treat all existing research as having equal merit. Their studies provide academic legitimacy to claims of extensive misattributed paternity.
After I did that research on the true extent of misattributed paternity, I had several articles on the topic published in top-ranking international journals. But my finding that the extent of misattributed paternity is tiny is neither shocking nor, it seems, newsworthy.
Bad news travels fast; good news more slowly. No wonder that fathers’ rights groups, the paternity testing industry and evolutionary psychologists find an audience for their inflated estimates. No wonder that a snippet of conversation lives on as an urban myth. •