The number of words that children hear from adults throughout the early years of life is thought to permanently affect their intelligence and cognitive development. However, previous studies in this area relied exclusively on extrapolations from just one hour of recording to estimate how many words children hear from adults across days and years. Here, we obtained for the first time 3 daylong recordings (mean length in hours = 15.07, SD = 1.88, range: 5.81 to 18.08 hours) from 107 monolingual British children (i.e. total of 321 recordings) aged 24 to 48 months. Recordings were completed on 3 separate days in the families' homes using digital language recorders that children wore in designated chest pockets to record all sounds within a 6-foot radius.
In addition, all families completed the Parent Report of Children’s Abilities (PARCA), which was developed to assess non-verbal intelligence in early childhood, including parent ratings of children's ability to perform 28 activities (e.g. "Can your child stack seven small blocks on top of each other by him/ herself?") and a testing booklet that the child completed at home. The booklet included 9 drawing tasks, 7 copying tasks and 10 matching tasks.
We made three important discoveries. First, the number of adult words that children heard per day differed dramatically across families, ranging from 3,177 to 45,478 with a mean of 18,022 words (SD = 7,336). By comparison, previous studies had estimated that the number of words that children of comparable socioeconomic status as our sample heard on average 34,328 words, which could not be confirmed here.
Second, the number of words that children heard per day differed substantially within families across study days. Absolute differences in adult word counts ranged from 350 to 14,433 words within families across days (M = 3,477, SD = 2,443). Accordingly, the number of adult words that children heard was inconsistent across days (intra-class correlation (ICC) = .47) suggesting that about half of the variance in adult word counts occurred within and half between the families. We also found that 86% of the variance in the number of words that children hear across hours occurs within families (ICC = .14), implying that children hear vastly different quantities of adult words from one hour to the next.
Finally, we found that adult word counts across days were positively associated with the PARCA testing booklet scores, accounting together for 10.1% of the variance in children's intelligence scores. This result confirms previous claims that children who show greater cognitive ability in early life tend to grow up in households where a greater number of adult words are heard.
Our study highlights that the language environment that children experience varies dramatically across time between and within families. Our data also substantiated that children, who are exposed to more adult language, show greater intelligence. The mechanism that underlies this association remains to be uncovered.
