back

by EndXA·7y ago·view on hn ↗
The original article is available at: https://psyarxiv.com/xynwg/ (note that it's a preprint, and hasn't yet been subjected to peer review)

Abstract:

> Based on the analysis of 190 studies (17,887 participants), we estimate that the average silent reading rate for adults in English is 238 word per minute (wpm) for non-fiction and 260 wpm for fiction. The difference can be predicted by the length of the words, with longer words in non-fiction than in fiction. The estimates are lower than the numbers often cited in scientific and popular writings. The reasons for the overestimates are reviewed. Reading rates are lower for children, old adults, and readers with English as second language. The reading rates are in line with maximum listening speed and do not require the assumption of reading-specific language processing. The average oral reading rate (based on 77 studies and 5,965 participants) is 183 wpm. Within each group/task there are reliable individual differences, which are not yet fully understood. For silent reading of English fiction most adults fall in the range of 175 to 300 wpm; for fiction the range is 200 to 320 wpm. Reading rates in other languages can be predicted reasonably well be taking into account the number of words these languages require to convey the same message as in English.

1 comments
I think words per minutes seems like a highly problematic metric, for a variety of statistical logics reasons.

Instead, I think we should try to measure multi-word expressions (MWEs) per second, and consider special characters (including spaces!) as words.

Why? Well, to start with, due to us already knowing about the Auerbach-Estoup-Yule-Zipf-Pareto-Mandelbrot-Simon-Price Law(s) only holding for multi-word expressions, not for phonemes, words, characters or syllables. Can't dig up the papers on that right now, tho.

(Personally however, I suspect that MWEs per second ain't the be all, end all, of reading speed, either. It ignores concepts like Paronomasia & Polysemy, as well as even more obscure things, like Polyphonemes. I just so far lack sufficient knowing of any possibly plausible hyperparameterizations which'd generally let one compile some metric for further complexity reductions.)

I'd elaborate on /how/ that (partially) answers the "why", but I lack the time to do so at the moment.