With DNA tests now widely available, so-called paternity fraud has become a staple of talk shows and TV crime series. Aggrieved men accuse tearful wives who profess their fidelity, only to have their extramarital affairs brought to light. Billboards in Chicago and other cities provocatively ask, “Paternity questions?” and advise that the answers are for sale at your local pharmacy in the form of at-home DNA paternity tests. Some fathers’-rights groups in Australia have called for mandatory paternity testing of all children at birth, with or without the mother’s consent or even her knowledge.

And people are pretty well convinced there is a need for all this vigilance. When asked to estimate the frequency of misassigned paternity in the general population, most people hazard a guess of 10%, 20% or even 30%, with the last number coming from a class of biology undergraduates in a South Carolina university that I polled last year. I pointed out that this would mean that nearly 20 people in the class of 60-some students had lived their lives calling the wrong man Dad, at least biologically. They just nodded cynically, undaunted. Even scientists will quickly respond with the 10% figure, as a geneticist colleague of mine who studies the male sex chromosome found when he queried fellow biologists at conferences.

What are the real numbers? Zuck asserts that they’re more in the 1-5% range, with 3.7% being a high-bound figure for one study. This varies by culture and socioeconomic group, and the segment of the population being surveyed. Studies which rely on a data set consisting of men who have requested paternity tests are strongly sample biased toward those who have a reason to have suspicions. It’s somewhat like some of the critiques of the Kinsey Reports and deviant sexual behavior; if you survey sexual deviants (however you define that) to get a sense of the proportion of deviancy in the population you’re going to get a higher than representative figure. And yet even in the cases of men who have suspicions only a minority have misattributed paternity.

What is this telling us? As I said above my own interactions are with people of generally liberal inclinations and values, and the high paternity uncertainty numbers aren’t offered up as evidence of the lascivious nature of women or the degraded state of modern morals. Rather they’re presented as evidence of gender equality, sexual liberation, and a generally praiseworthy reflection on the weak emphasis of genetic ties as the root of the parenting bond. But what if both assessments (positive and negative) of perceived low paternity confidence emerge from the same evolutionary psychological bias: to weigh false positives much more highly than false negatives in terms of plausibility. In other words, it doesn’t hurt to be suspicious about something as important as paternity in fatherhood (from an evolutionary perspective; something which is going to be generalized across the kin group, not just for the male in question). More generally it doesn’t hurt to be suspicious or a little over-active in your imagination as a rule. This is one of the major models for why people see supernatural agents all around them, it’s evidence of an over-active agency detection module in your brain which incurs minimal cost for false positives (waste a lot of time propitiating ghosts and gods) in comparison to the deleterious consequences for false negatives (you ignore the threat of a dangerous animal or a hostile tribal band and get killed).

Since I’ve only presented assertion so far, let me point to one of the most thorough cross-cultural studies I’ve stumbled upon, How Well Does Paternity Confidence Match Actual Paternity?:

This survey of published estimates of nonpaternity suggests that for men with high paternity confidence, nonpaternity rates are typically 1.7% (if we exclude studies of unknown methodology) to 3.3% (if we include such studies). These figures are substantially lower than the “typical” nonpaternity rate of 10% or higher cited by many researchers, often without substantiation…or the median worldwide nonpaternity rate of 9% reported by Baker and Bellis…

Men who have low paternity confidence and have chosen to challenge their paternity through laboratory testing are much less likely than men with high paternity confidence to be the fathers of their putative children. Although these men presumably have lower paternity confidence than men who do not seek paternity tests, this group is heterogeneous; some men may be virtually certain that the putative child is not theirs, while others may simply have sufficient doubts to warrant testing. Most of these men are in fact the fathers of their putative genetic children; only 29.8% could be excluded as biological fathers of the children in question.

To me it’s striking that the majority of the men who have low paternity confidence and suspicion enough to submit to a paternity laboratory are still the biological fathers of their offspring! In fact, the rates of non-paternity among this set is closer to the urban myth proportions found among the lay public. Of course it is this set of men who show up on the Jerry Springer show, not men confident in their paternity, so the public may be getting a false cultural picture of frequencies which they’re projecting.

Since you want data, here are the relevant tables.

Obviously there’s some variation. The rule of thumb seems to be that males of higher socioeconomic status, and from more conventionally bourgeois societies, have greater warranted paternity confidence. Lower paternity confidence among those who are the principals for sensational media shouldn’t be surprising then. But if you’re a Bayesian you should “update” accordingly (if you know what Bayesian probability is, you are probably the type of person who shouldn’t worry*).

In fact, from what I know about contemporary genetic genealogy they’re reinforcing the the idea that when paternity confidence is high, it is high for a reason. Men who are interested in their patrilineage often get their Y chromosomes sequenced, and it turns out that in societies such as England where surnames have a relatively deep history in some families (generally high socioeconomic status ones) the vast majority of men share the same male ancestor hundreds of years in the past. The balance share many different other male ancestors (usually following a power law distribution) . What’s going on is that each generation some males will have misattributed paternity, so the original male with a particular Y chromosome and surname will pass on the surname, but not the Y chromosome. The disjunction between these two in a given generation from t = 0 can be used to infer approximate rates of cuckoldry. If for example only 1% of males in each generation are from genetic lineages outside of the primary one assumed in their family, then after 100 generations only 37% of males will be from the original lineage to which the surname was associated. In contrast if you assume a 10% figure then it will only take 10 generations to reach an equivalent proportion (10 generations is about 250 years, which is a short enough time period so that many males from Western nations could track their male ancestor and their distant relatives relatively easily through parish records, so these back-of-the-envelopes are easily checkable now).

With the spread of genetic sequencing for recreational and health reasons arguments by some groups for mandatory paternity testing will seem quaint as the information will be available as a matter of course. Nevertheless there is going to be the conventional hand-wring from bioethicists as to whether to share the data, in part to preserve the family unit. Here’s a somewhat dated survey result from 1998:

….On the issue of misattributed paternity, two thirds of US medical geneticists would not disclose this information to a woman’s partner, even if he asks, because the disclosure might jeopardize the marriage or endanger the woman. In contrast, three quarters of potential counselees believe that the physician should give this information to a partner who requests it, after informing the woman of the intent to do so…Possibly, the counselees presume that the unit of confidentiality is the family, not the individual, or are less aware of potential adverse consequences of disclosure.

In an interesting coda to this story, barring government action to make this information unavailable without professional assistance (I’m skeptical that this would happen), the ubiquity of genetic testing no matter its utility in the near future will make these disclosures by professionals only selectively relevant. If you’re intelligent it shouldn’t be that hard to analyze the data, skim the results, or find someone willing to do so for a nominal sum (I’m assuming there will be software which will do a lot of the analysis of your raw sequence which you can input from a file yourself). But it is the less intelligent and unsure who are the ones who have to worry most about paternity uncertainty, and the ones least likely or unable to understand the clear confirmation of misattributed paternity in the data (I’ve seen some papers which claim that some fathers who know their child has a recessive disease, and also find out they’re not a carrier of the recessively expressed allele, still can’t connect the dots!).**

* Yes, I’m making a normative assumption here that if you’re male you should be displeased if you find out that children whom you assumed were your biological offspring turn out not to be. If, on the other hand, you think it’s fun and adds more zest to your life, you’re just kind of weird. Sorry if I sound prejudiced, but I know that the cuckold community is going to link to this post, so I’m hoping you guys don’t start leaving angry comments for disabusing you of your fantasies, as has occurred before when I post on this.

** By the way, if misattributed paternity was very common in the past because men didn’t care one would assume that evolutionary pressures would have selected against it until it was rare. So if the urban myth figure was correct, it would be relatively new, or, there’s no heritable trait associated with getting cuckolded and it happened in a scattershot fashion across the population.

Citation: Anderson, K. (2006). How Well Does Paternity Confidence Match Actual Paternity? Evidence from Worldwide Nonpaternity Rates Current Anthropology, 47 (3), 513-520 DOI: 10.1086/504167