Welcome to “The Sniff Test,” a semi-monthly column in which we examine extraordinary claims made by science journalists and compare these claims to the evidence actually presented in the relevant publication. We will attempt to be as educational as possible, but excuse us as we work out the kinks in our first few posts.
First up: are fifth-graders cheating off ants in math class? If not, should they be? Setting aside the ethical dilemmas, this seems like a logical follow-up question to a recent story put out on April 11 by Discovery News, “Are ants smarter than fifth-graders at math?” I guess, on the one hand, this title may be a humorous reference to the FOX show that probably inspired it. On the other, it seems like it could be a bit of a stretch, considering the content standards for fifth grade math. The article itself describes a recent review published1 by Z. Reznikova and B. Ryabko in the journal Behaviour (and a related report published2 in Electronic Transactions on Artificial Intelligence). The research shows apparent counting and arithmetic abilities in a wood ant species, Formica polyctena. Scouts were placed in a food-containing trough and navigated back to their group. Here, they had to transmit information about the location of the food, and then they were removed from their group, so that the foragers had to find the correct trough using only the information transmitted by the scouts. The researchers attempted to remove any means of information transfer except for communicating numerical quantities (e.g. they replaced the experimental apparatus to remove odors and odor-trails). At the very least, the authors showed fairly convincingly at least a type of counting ability in the ants. Whether they performed actual arithmetic is debatable, but their evidence is suggestive. All in all, pretty cool research. Not only does this species seem to possess some numerical faculty, but the way in which individuals communicate the location of food to each other seems to be a remarkable ability.
The Discovery News article does a decent job of distilling some of the main concepts of the research. However, some of the language used seems to oversell the work. First, the title itself seems clearly overblown, though admittedly, it seems to have been inspired by a press-release put out by the publisher of the Behaviour paper itself; we don’t need scientific publishers contributing to sensationalism. Next, the first sentence says that the study suggests “ants are smarter than many grade-school kids, at least when it comes to math….” However, the researchers say no such thing in either of their papers. Instead, they suggest that some ants’ numerical abilities may be on par with a very restricted subset of human abilities, and even this clearly applies to a narrow subset of human cultures and developmental stages. In the last paragraph of the article, the author discusses some of the other species reviewed in the scientists’ work (birds, mammals, etc.). She mentions that high numerical aptitude in pigeons, crows, and parrots must carry over to other birds, because these three species simply may be relatively easy to study. This could be a reasonable suggestion (though crows and parrots, for example, are known to be exceptional rather than ordinary as far as avian intelligence goes3, 4), but it seems more appropriate for an opinion piece, rather than a report on research results—the research papers themselves never try to extend these results to all birds. Finally, while the research reports make clear that their results on Formica polyctena likely only extend to other highly social ant species, the Discovery News article makes it seem as if the research can be applied broadly across ants. A small point, but one not to be taken lightly; the question of how generalizeable research should be is one that crops up repeatedly.
In sum, based on the title and some of the sensationalist language, this article does not quite pass “the sniff test.” The results of the exciting research being covered should be allowed to stand on their own; the article instead seems to detract from the thought-provoking science being done.
References
1. Reznikova, Z. and Ryabko, B. 2011. Numerical competence in animals, with an insight from ants. Behaviour 148:405-434.
2. Reznikova, Z. and Ryabko, B. 2001. A study of ants’ numerical competence. — Electronic Transactions on Artificial Intelligence. 5: 111-126.
3. Emery, NJ. and Clayton, N.S. 2004. The mentality of crows: convergent evolution of intelligence in corvids and apes. Science 306:1903-1907.
4. Huber, L. and Gajdon, G.K. 2006. Technical intelligence in animals: the kea model. Animal Cognition 9:295-305.
-Tom Eiting
No comments:
Post a Comment