Shannon was kind enough to forward me a link to Jezebel’s post highlighting a recent study done in Scotland that purports to show a benefit for kids who face their parents in their strollers. I’m not one to care a whole hell of a lot about things like this, but for some reason, this piqued my interest enough to go looking for the original study, bu a Dr. Suzanne Zeedyk, and I have to say, I managed to stumble upon one huge, steaming pile of meconium (that’s newborn baby shit for those of you who haven’t had the pleasure).

(It’s here I’ll reveal that my alarm bells first went off when I was only able to find the research on the National Literacy Trust website and not in a peer-reviewed journal; that sort of thing is usually the initial sign that the research couldn’t pass peer-review muster.)

One of the biggest reported findings of this research is that kids who are facing their parents are far more likely to be sleeping (which Zeedyk reports has “tentatively been interpreted as an indicator of stress levels”). The problem is that Zeedyk doesn’t look at that number in light of the ages of the kids she observed, and realities about behaviors of children at those ages. So let’s do that, starting with the numbers.

In her second research results table, Zeedyk reported 386 total children who were observed to be sleeping. In the same table, she reported 828 total children who they estimated to be under a year of age, 35% of whom were sleeping, for a total of 290 children under the age of one and sleeping. That means that 75% of the sleeping children were under the age of one. A reasonable researcher might stop here and ask herself: is there anything about children less than one year of age that might explain increased sleep other than stroller choice? Apparently, Zeedyk didn’t think so and plowed onward, not considering the simple fact that infants sleep more than toddlers no matter what the variable. (Seriously, it’s one of the cardinal three behaviors of an infant: eat, poop, and sleep — I say this both as a parent and a pediatrician. A quick persual of tables of sleep duration by age, in anything from Ferber’s book to childhood neuroscience texts, shows that infants sleep on the order of a few hours more a day than toddlers, who sleep more than school-age kids.)

Now let’s go onward with Zeedyk, looking at her eighth research results table, the one in which she “shows” that stroller direction affected rates of sleeping in the youngest age group. I literally cannot explain her results in this table; the numbers just don’t make any sense. In it, she claims to have 392 total observations of whether or not a child was sleeping (287 sleeping, 195 awake) — but she has 766 observations of stroller direction (492 facing away from the parent, 274 towards). How the hell does this make any sense? There’s no possible way for those numbers to be unequal without you throwing away those 374 missing observations of whether the kid was sleeping; this is literally crap science, something that would be laughed out of the most basic clinical research class. (Nevermind that in her prior table 2, Zeedyk recorded 828 observations of children under 1, out of whom roughly 538 were not sleeping — but in this table, she only shows 105 children in this age group not sleeping. Again, I can make absolutely no sense out of this.)

Another thing: Zeedyk openly admits that she has no clue if the adult in an observed child/adult paring was a parent (versus a grandparent, a nanny, an aunt, or whomever), but then dismisses this with, “Because there is no way of making such distinctions, for the purposes of this study we have treated all adults in the sample as parents. The conclusions that will be drawn from this study, regarding the extent to which infants are experiencing interactions with adults during outdoor journeys, are in no way compromised by this decision.” Ummm, huh?!? Does she try to justify this immense logic leap at all? Nope.

I could go on and on with this — it’s actually sort of fun, a way to use all my critical-research-reading skills for parenting purposes — but I’ll just leave it at that. If you’re a parent who’s considering turning your dear kiddo around to face you just because of Dr. Zeedyk’s “study,” I’d suggest that there’s probably more evidence of a cellphone in Africa causing cancer in your child than the direction of his or her stroller affecting jack shit.

And now, because I can’t resist, a few little things:

• There are all kinds of places where even Zeedyk’s simple numbers don’t match up. For example, her tables show that she had more observations for what the one to two year-old kids were doing (980) than she had observations of how the one to two year-olds were being transported (978) — why is that? Are we to believe that there were any cases in which she was able to figure out if the kid was awake or asleep but unable to tell whether the kid was in a forward-facing or rearward-facing stroller? This is just weird. Similarly, in the table comparing stroller direction to child activity, she shows 1,659 observations of the stroller facing away; in the table comparing stroller direction to age, she only has 1,650 observations of the stroller facing away. What happened to those other nine kids? Including them in one evaluation while excluding them from the other doesn’t make much sense.
• Zeedyk is very loose with her rationalizations that she uses to rebut potential weaknesses in the study. For example, she posits that the incredibly brief time in which she observed any parent/child grouping might be a weakness in her conclusions about whether or not the parents and children were talking to each other, and then dismisses it with this: “The fact that the observation was made at a random point in each pair’s journey means that we have adequately sampled trends for the sample as a whole. Statistical logic allows us to predict that, if talking behaviour was random, we should have observed parents talking as often as we observed them not talking.” Not quite, Dr. Zeedyk — that’s only true if the talking behavior were both random and equally likely to be occurring as not occurring, or in other words, if the chance of the parents and kids talking at any point in time were 50%. If instead we were to posit that that chance is 10%, then she’d have to observe 10 parent/child groupings to see one of them having a conversation, which would be reflected very different in her findings.