Useful research, problematic headline

Past research has shown that print leisure reading (e.g., reading comics) has a positive relationship with reading comprehension performance. Given young people’s tendency to do more digital leisure reading (i.e., on phones, on tablets, etc.) than print, it’s fair to wonder if digital leisure reading would have the same positive relationship with reading comprehension. If so, cool, let kids continue to read in whatever medium they wish. But if not, um, that would be a problem given digital is replacing print reading. Well, Altamura et al. (in press) did a meta-analysis on the relationship between digital leisure reading and reading comprehension, and the results weren’t so encouraging. The first author sums it up nicely in this video:

These findings suggest the benefits of print leisure reading do not extend to digital leisure reading. Now, what I like about the video, and the article, is that the authors acknowledge the many limitations of the meta-analysis (no research is perfect!), including that the included studies were correlational and thus causal inferences cannot be made.

What’s a bit troubling is that a press release by the American Educational Research Association uses the title “Study: Digital Leisure Reading Does Little to Improve Reading Comprehension for Students.” The phrase “does little” implies causality, to me (maybe I’m being picky or sensitive, but I don’t think so). We don’t know if digital leisure reading does little - we just know the correlation between digital leisure reading and reading comprehension is very small. As the authors state in the video and in print, more research is needed and they aren’t necessarily “against” digital leisure reading. Thus, I think the authors are more measured than the headline, but I worry the headline is going to be what people remember and infer from (i.e., a causal relationship). I wish it was different.

Of course there are other limitations. For example, on social media, Dan Willingham pointed out there were numerous types of digital leisure reading used in the studies. So, it’s tough to know what to infer about a meta-analysis that included social reading with linear/informative reading in the same analysis. The authors tried to do a moderator analysis to look for differences in type of digital reading, but the sample sizes were pretty small. There was some indication linear/informative reading was more strongly related to reading comprehension than other types. Yet another direction for future research! Let’s hope people hold off on pronunciations of the “evils” of digital leisure reading until that research is done.