Sunday, May 13, 2018

Sometimes, all you can do is believe

Hmmm. From Reexamining the Verbal Environments of Children From Different Socioeconomic Backgrounds by Douglas E. Sperry, Linda L. Sperry, and Peggy J. Miller. It is gated so I cannot see anything but the abstract.
Amid growing controversy about the oft‐cited “30‐million‐word gap,” this investigation uses language data from five American communities across the socioeconomic spectrum to test, for the first time, Hart and Risley's (1995) claim that poor children hear 30 million fewer words than their middle‐class counterparts during the early years of life. The five studies combined ethnographic fieldwork with longitudinal home observations of 42 children (18–48 months) interacting with family members in everyday life contexts. Results do not support Hart and Risley's claim, reveal substantial variation in vocabulary environments within each socioeconomic stratum, and suggest that definitions of verbal environments that exclude multiple caregivers and bystander talk disproportionately underestimate the number of words to which low‐income children are exposed.
That is disappointing, though not especially surprising.

Hart and Risley's was an heroic effort to ground important ideas in empirical research and their findings held out hope for a couple of ideas I would really like to be true.

A summary of the book from the original study.
The authors conducted this study to look for the cause of the disparity in linguistic/academic progress among children from different socioeconomic backgrounds. For 2.5 years, 7 month old children in 42 families were observed for one hour per week until the child turned 3 years old. Everything said to the baby, all talk the baby overheard, and everything the baby did or said during an hour of daily life was recorded and analyzed. Charts of monthly growth in vocabulary, utterances, and use of grammatical structures, as well as a list of 2,000 vocabulary words used during early language development are provided. The authors found the average welfare child had 1/2 as much experience per hour (616 words per hr.) as the average working class child (1,251 words per hr.), and less than 1/3 as much experience as the average professional class child (2,153 per hour). For example, in one year, extrapolating to a 100 hr. week (given a 14 hr. waking day), 11 million words for professional, 6 million words for working class, and 3 million words for welfare class were recorded. Parent talkativeness or “sociableness” to their infants accounted for a correlation between SES and the children’s later linguistic/academic development. For example, in low SES families, parent-child interaction tended to involve directives being given to the child whereas in higher SES background families the parent-child interaction tended to be more conversational. These are likely reasons why children from lower SES perform worse on standardized vocabulary tests than children from middle SES backgrounds. The study also demonstrated that the size of a child’s vocabulary could be based on experience and not necessarily attributed to an inherent language learning difficulty.
You can see the problem right there in the second sentence.

42 families. That is a horribly small sample size. It is understandable because this sort of labor intensive study is astonishingly expensive. But 42 only gets you to suggestive.

Ironically, Sperry et al are disputing Hart and Ridley's original study with five studies based on a sample size of . . . wait for it . . . 42.

I am not arguing that one or another is accurate. Both are too small to be anything more than suggestive.

But the Hart & Ridley study was so appealing because it suggested a simple set of solutions to what is obviously a complex set of causes. Read more to your children. Talk more to your children. It was also appealing because these are things which can easily be accommodated in a fashion which most education recommendations cannot.

Instinctively I believe reading to children and talking with children to be good things and likely to be materially beneficial in academic outcomes. I have believed that and that is what I have practiced with my own children. But the evidentiary base for such recommendations is thin at best. Hart & Ridley were one of the more credible elements in that support. All along it has had to be caveated with that wretchedly small number 42. Now it cannot even be used much at all. Another 42 family study has come in on the other side of the ledger.

I will continue to believe in the value of reading and talking but I sure would like some compelling empirical evidence to support that belief.

2 comments:

  1. The Lena device people wrote a journal article in 2017 about their computer analysis of the kids wearing the device:

    https://pubs.asha.org/doi/10.1044/2016_AJSLP-15-0169


    After "Reexamining..." talked about in this post, a rebuttal was written:
    https://pubmed.ncbi.nlm.nih.gov/30102419/

    With a response:
    https://pubmed.ncbi.nlm.nih.gov/30102424/

    ReplyDelete
  2. Jason A. - Thanks for the update.

    ReplyDelete