Interesting, but the sample size is very small and they only processed the most talkative few hours a day from each 16 hour daily recording. I'm a little surprised they didn't seem to incorporate how often children were spoken to, rather just what happened in the most talkative moments, which were around 1/8 to 1/4ish of the day each day.

Cool preliminary findings, but this needs way more intensive studies to determine anything with confidence.