Paul Morgan is an associate professor in the Department of Education Policy Studies at Pennsylvania State University. Along with a team at Penn State and colleagues at the University of California, Irvine, and Columbia University, Morgan examined thousands of data points to determine the effect of vocabulary size at age 2 on academic performance and behavior at age 5. The results, while maybe intuitive to early educators, are fascinating. We spoke with him to discuss his study, which was published in Child Development in August, for our Ask A Scientist series.
The conversation below has been edited for length and clarity.
Can you summarize this study on early-childhood vocabulary that you did recently?
Sure. We examined population-level data and factors that seem to be associated with or predictive of a very early onset of vocabulary gaps. We’re looking at differences that are evident between groups of 2-year-olds in the U.S.
We analyzed data collected on a longitudinal, nationally representative sample. These were children that were born in the U.S. in 2001, and then followed as they aged until they reached kindergarten. The children were surveyed in terms of their vocabulary size [as reported by their parents] when they were 2 years old. In kindergarten they were administered independently, untimed, individualized assessments of their reading achievement and mathematic achievement.
There were also kindergarten teachers who completed ratings of their classroom behavior. The first thing we were interested in was the factors that were associated with or predictive of the onset of vocabulary gaps at age 2. Second, we were interested in whether vocabulary was uniquely predictive of children’s academic and behavioral preparation for kindergarten.
You were trying to figure out if how many words kids spoke at age 2 predicted how they were able to behave both academically and socially in kindergarten?
That’s right. [There’s one body of research that] is trying to understand vocabulary gaps and when they emerge and who’s at risk for being “off track.” There’s been quite a bit of work in that area, but most of it has used pretty small sample sizes.
The most famous study in this area was conducted by [University of Kansas researchers Betty] Hart and [Todd R.] Risley in the mid-90s. They did a very intensive study of vocabulary learning and differences between higher-socioeconomic-status families and lower-socioeconomic-status families. But their sample size was very small. It was like 42 families.
And in other work, [researchers] have observed that gaps start to be evident at 18 months of age. But they’ve been based on, again, pretty small samples.
So our unique contribution is that we’re analyzing a very large sample of about 10,000 children when they’re first assessed. [Note: There is always some loss in longitudinal studies so there weren’t 10,000 still in the study at age 5.] There’s some advantages that come with that. There’s a lot that’s known about the differences in the background characteristics of the children, so you can examine factors that have been hypothesized to be related to vocabulary.
The second contribution of this study is that the dataset [allows us to] look at this relation between vocabulary at 2 years old and reading achievement in kindergarten. We can look at it not only in terms of reading achievement but also math achievement, self-regulation, attention span, and task persistence, as well as indicators of problem behavior like acting out, anxiousness, and withdrawal.
What we find is that vocabulary [size] is predictive of both greater reading achievement and math achievement as well as greater self-regulation and lower frequency of problem behaviors [in kindergarten].
Your saying children at lower socioeconomic levels tend to have smaller vocabularies at age 2 and therefore have smaller vocabularies in kindergarten as well as problem behaviors. Is the vocabulary size necessarily causal, or might there be other contributing factors that change behavior at the kindergarten level?
That’s a good question. This not experimental data so we didn’t randomly assign an intervention to the children. Ours is an observational study, but there are ways to provide preliminary or suggestive evidence of causality. Because there’s extensive measurement about the children’s background characteristics, we can do two things that relate to a causal relation. That just means that vocabulary’s predictive of later achievement and behavior.
And the other thing we can show is that that relation holds after controls for many different factors. We’re controlling for even things like children’s general cognitive function at 2-years-old and we’re still finding that this relation holds.
The other thing is that there are a couple things that make this finding a bit more robust than the evidence to date. One of which is the measurement of children’s vocabulary in the dataset is very simplistic. Parents are just responding to a series of 50 words and asked when their child says the word. So the words are things like “meow” and “thank you.” And the parents just respond “yes” or “no.” So it’s a very coarse, quick measure of vocabulary. What we’re showing is even that simple measure are predictive three years later on measures [of reading and math performance, as well as behavior] that were done completely independently.
And did you find there were children from lower socioeconomic backgrounds who had high levels of vocabulary? And then did those kids behave better?
We didn’t conduct exactly that kind of analysis.
From a teacher’s perspective there’s this question always of “OK, I have this low SES kid.” Would it have made a difference if their parents had read to them more, or had spoken to them more, so that that kid had that higher level of vocabulary as the higher SES kid? You know, that’s the question kindergarten teachers are asking. Do we know that?
I can’t form that from my research. But not all kids who are coming from low socioeconomic backgrounds are displaying low levels of achievement. It’s not a fixed kind of deterministic relation. It’s true that it makes it more likely, but go back to our study, we find that these differences start to merge prior to entering school. I was a teacher, and I understand how teachers approach this. A lot of the action in terms of children’s cognitive development is taking place prior to school entry. But there’s always more we can do to help.
From a practitioner’s standpoint, our results provide additional support for the idea that if we were to do a better job as a country to help put children on equal footing during the early-childhood period, and as they’re getting ready to transition to kindergarten; that if we helped parents who are stressed or overburdened or may not be talking to their children or providing lots of exposure to storybooks; if we did that, I think it would help put children on a better footing as they make the transition in kindergarten.
So these interventions—trying to get parents to read, talk, and sing with their very young children—are worthwhile?
Yes. There are lots of things that are difficult for practitioners, especially in the education field, to change. It’s hard for a Head Start teacher to change the family’s employment status, right? And it’s hard for a practitioner to somehow change a child’s low birth weight. But something that is malleable and can be affected by practitioners or by parents, is children’s vocabulary size. And it really does seem to matter in terms of preparing the children both academically and behaviorally for kindergarten.
A version of this news article first appeared in the Early Years blog.