Language development: it is not words heard that matter most, but conversational turns
The "30 million word gap" is the most repeated figure in early education. It comes from a study of 42 families — and the current evidence points somewhere else.
In their study published in 1995, Betty Hart and Todd Risley observed 42 families: 13 professional/high-SES, 10 middle-SES, 13 lower-income and 6 on welfare. Observation was one hour a month for about 2.5 years, from age 7–9 months to 36 months, for 1,318 observations in total. Words heard per hour: 2,153 in professional families, 1,251 in working-class families, 616 in welfare families. The "30 million word gap" is an extrapolation of those hourly rates to age four — roughly 45 million, 26 million and 13 million words. It is a calculated projection, not a measured cumulative count.
In 2019, in Child Development, Sperry, Sperry and Miller published a study of 42 children (18–48 months) across five American communities that did not support the original conclusion. Their central argument is methodological: measuring the verbal environment as the speech of one primary caregiver directed at the child systematically underestimates what children hear in settings with multiple caregivers and a great deal of overheard talk. In the same journal, Golinkoff, Hoff, Rowe, Tamis-LeMonda and Hirsh-Pasek published a rebuttal: the Sperry team sampled only middle- and low-income families, with no high-income comparison group, so it is not a direct replication; and the mechanism that matters is child-directed speech, not adults talking to adults. The question remains genuinely contested.
Much firmer evidence came from another direction. Romeo and colleagues, publishing in Psychological Science in 2018, measured separately in children aged 4–6 the number of words overheard, the number of words directed at the child, and the number of conversational turns — cycles of back-and-forth between adult and child — while imaging the brain with functional MRI. The result: the measure associated with language skill and with activation in Broca’s area was the number of conversational turns, not the total number of words heard. More importantly still, once conversational turns were accounted for, the association between socioeconomic status and language skill disappeared.
What does this change for a teacher? Above all, the target. "Talk more to children" is hard to measure and weakly evidenced. "Take more conversational turns with children" is measurable and evidenced. Practical steps: build at least one conversation of five or more turns with an individual child each day, and record that you did; replace closed questions ("What colour is this?") with open ones ("Why did you do it that way?"); after a child answers, do not move straight to a new question — expand and reflect back what they said; shorten whole-group lesson time and increase small-group and one-to-one talk; and plan mealtimes, dressing and outdoor time not as transitions but as the primary opportunity for dialogue.
