I think this is an excellent example to serve as a focus of arguments for and against evolutionary psychology. The idea here is not to praise or critique this particular piece of research, but rather to discuss what kinds of things good scientific researchers might do with this subject matter vs the things that sloppy researchers might do. The cited paper is not available online, but a later survey paper by the same author gives the rough outline.

As I understand it, DeKay is suggesting that differential care for grandchildren is a (genetic?) fact of human nature. And that the explanation for this differential can be found in the fact that some "fathers" are actually cuckolds, together with the standard evolutionary theory of inclusive fitness (aka kin selection). The empirical part of the research examines a sample of grandparents and finds that the predicted differential grandchild-care behavior is actually observed.

knb finds this research to be "clever", "surprising", and "impressive". I, on the other hand, am more skeptical. Since we don't have the actual research paper, we can't criticise the actual research, but we can ask the kinds of questions that an evolutionary biologist might ask if it had been submitted for peer review to a biology journal.

First a scientist would want to know what other hypotheses had been considered and then empirically rejected. For example, might the results tell us something about culture (social norms) rather than about human nature. The scientist would want to make sure that the empirical sample of grandparents was taken from a broad diversity of human cultures and hence that a "human nature" label is not mistakenly being attached to a parochial cultural trait.

Besides culture and genetics, another possible explanatory factor for human behavior is rational self interest. I would want to know whether any of the sampled grandparents had any reason to rationally suspect that they were not true ancestors.

Another alternate hypothesis would be that the modern western cultural practice of divorce with maternal custody might have something to do with this - particularly when one party moves to another town. I would want to know how the research compensated for this. Otherwise, an attractive alternative hypothesis is that grandparent care happens because/when such care is solicited by the primary care-giver, and that primary caregiver usually solicits assistance from her own primary caregiver.

Since the proposed explanation is that the differential evolved by natural selection, driven by the probability that the nominal parent is not the genetic parent, I would want to have measurements of the differential taken in two populations with different values of the probability. For example, compare a population in which, for many generations, it has been the case that a nominal father has only a 15% chance of not being the genetic father, with a different population in which the probability of cuckoldry is more like 30%.

I would also like to see some results comparing adoptive vs genetic parentage. This kind of data might tease out which hypothesis regarding the causation behind the behavior (genetic, cultural, or rational) is more likely. In particular, this kind of data might show whether a propensity to care for grandchildren is a genetic trait, (passed on to genetic children), or a cultural trait (passed on to adopted children as well).

Ok, those are some of the issues that an evolutionary scientist would expect to see addressed in a paper purporting to be about evolutionary science. Does the cited research deal with these issues? I don't know, but if anyone here has read the paper, I would be curious.