It’s sad to see HN take this at face value and parrot the “screens bad” view without understanding it.
I dug deep into this a while ago, starting with the “how legit is the science” question because I wondered if the studies had looked at any tradeoffs (e.g. did laptop use improve programming skills in ways paper books do not?)
It’s a rabbit hole. I encourage folks to read up and form more nuanced opinions.
This being HN I need to assure you that my learned skepticism regarding harms from screens in schools does NOT mean I want to ban all books in schools, strap toddlers into VR for their entire childhood, or put Peter Thiel in charge of all curriculums. Intuitively I think paper allows greater focus. But the data is not nearly as clear as politics-driven advocates claim.
Some info:
- The move back to books was a centerpiece of election policy by the center-right government, and is at least as much about conservatism as it is about science.
- Actual studies in this area are mixed.
- A lot is made of PISA scores, which dropped from the 2010’s to early 2020’a (when this policy became popular). But: the scores started dropping before 1:1 computers were rolled out, and also correlated with teacher shortages and education policy changes, and of course COVID. I could not find any studies that controlled for these other factors, and the naive “test scores can be entirely attributed to computers” view really doesn’t hold.
- There was a major change in pedagogy in Swedish schools that predates introduction of computers and seems like a better explanation for lower scores [1]
- One meta analysis does show a very small but stat sig decrease in reading comprehension for non-fiction when read from screens rather than books [2]
- Another meta analysis found zero difference between screens and books for reading comprehension [3]
- A third meta analysis found a tiny and decreasing negative impact from screen use, and some evidence that the effect is transitional as teachers and students adapt [4]
- The vast majority of studies in this area use no children at all, only adults. There are good ethical reasons for this, but it is a mistake to assume that a 25 year old’s reading comprehension from screens in 1995 is predictive of an 8 year old’s in 2026. [5]
- One of the few studies that did look specifically at children found that paper outperformed screens… but only in traditional schools. Homeschooling and lab testing did not show any difference between mediums [6]
1. https://www.edchoice.org/is-swedish-school-choice-disastrous...
2. https://www.edweek.org/leadership/screen-reading-worse-for-c...
3. https://www.tandfonline.com/doi/full/10.1080/15213269.2022.2...
4. https://oej.scholasticahq.com/article/125437-turning-the-pag...
5. https://www.sciencedirect.com/science/article/abs/pii/S03601...
6. https://journals.sagepub.com/doi/10.3102/0034654321998074
Education research is really low quality. Like so many other fields in social sciences, the results rarely generalize beyond the direct findings, and only support the hypotheses in the mildest way. It cannot robustly guide decision making.
The fact that studies on screens vs books cannot get a consistent answer says enough. I checked #3 of your links, and the amount of bullshit is astonishing. The cited articles offer vague, unresearched explanations for contradictory findings, or point at differences in the stimuli, something which should obviously never have happened. After some cherry picking, article #3 treats the remaining studies as equal and reliable enough to throw in a big bag, as if that solves the problem.
Think of it like this: the replication crisis in cognitive psychology was found trying to replicate some of the better studies. The average education research study is several levels below that. It'll have a replicability of 0.1 or worse.