Causal Inference

Once again, just because there are two things that coincide with each other doesn't mean they have any effect on one another. The only way to actually determine if an infant sleeping with a fan provided any sort of benefit or not would be to randomly assign fans to a test group. However, like the debate over smoking's deleterious effects, this is unethical to actually force subjects to do something which might be dangerous to their health. But, there is no validity to this kind of study in the article, since it's based on a hypothesis and the results of the study do not prove or disprove the hypothesis. After all, the infants of families that have fans simply might be less likely to suffer from SIDS.
I could also argue that for infants in households that have dogs or cats, the infants are less likely to suffer from SIDS. There might be some actual valid reason for why this might be, but more likely than not, having dogs or cats would not directly affect an infant's susceptibility to SIDS, and it would just be a random chance that families more susceptible to SIDS would have fewer pets.

On another note, this is about as good as CNN reporting that scientists are worried that redheads might disappear, as if heterozygosity can't occur. Sheesh.