Nearly 20 Years Later, What Have We Learned From Hart and Risley?

The 30-Million-Word Gap: How Vocabulary Impacts the Achievement Gap

Dr. Steven L. Miller is a research neuropsychologist with expertise in the assessment and treatment of developmental language and reading impairments, and a co-founder of Scientific Learning. He has extensive experience in organizing clinical studies and conducting longitudinal studies of children and adults with such disabilities.

I chatted with Dr. Miller recently by phone about the famous 1995 language study by Hart and Risley to find out what we’ve learned since then, and why so many learners in our schools are still struggling.

SL: In 1995, Betty Hart and Todd R. Risley published the results of a study finding that 4-year-olds from working-class families and families on welfare had considerably smaller vocabularies than their agemates from professional families. This difference has been called the “30-Million-Word Gap” and “The Great Catastrophe.” Why has this study been so influential in education circles?

SM: The primary reason why the work has been so impactful is that it helped to quantify the challenge that educational systems face when children enter school with vast differences in educational readiness. This work demonstrates so clearly the enormous variation that kids present when they come to our public education system, and that traditional intervention approaches in many cases do not provide adequate advancement for students to “catch-up.”

So many people think children come to school at 4 ½ - 5 ½ years of age as blank slates. This work really points out the fact that there’s a lot of learning and experience that has gone on in the preschool years. And in the case of the Hart and Risley study, some students are millions and millions of words behind in experience when they enter the classroom.

SL: Why does this matter? What are the implications?

SM: Our national Reading First campaign has really dealt with the idea that there are critical subcomponents of reading instruction - that the research has proved are necessary, but not sufficient – for students to become good readers. And one of the best pre-school predictors of who’s going to be a good third-grade reader is their vocabulary development.

Adding to the Hart and Risley findings is research from NICHD funded research programs showing that kindergarten assessments can accurately predict greater than 90% of struggling 3^rd grade readers using their kindergarten assessment performance.  As such we can now accurately predict who’s at risk for becoming a struggling 3^rd grade reader.  This work led to the tiered intervention models that are positively changing student outcomes in schools across the country

So the question was, do we need different instructional paths for them? Do we expect them to develop academic skills at the same rate despite these early differences?  These aren’t biological restrictions in the students. This isn’t about their ability or potential to learn. This isn’t about whether they can learn to read, or become doctors, lawyers, engineers, or scientists. The answer to that is, yes, they can, but at the same time, if you’re not a good reader by third grade, there’s a high likelihood that you will not graduate high school on time, or possibly at all.  This is not a biological limitation but a consequence of the fact that a lot of academic learning occurs during independent reading after 3^rd grade.  Reading is critical because if you can’t read you fall further and further behind.  As we try to maximize the learning that occurs each day in school the consequences of falling behind can be devastating.  We see this is one of the simplest yet strongest findings in public education.  Student attendance predicts learning, missing school or falling behind at school makes catching up difficult because each day we have new learning goals being achieved.  This is why early learning and preschool program effectiveness is so critical to later academic success.

SL: It’s not just the gap in vocabulary size at age four that puts the children from lower-income families at risk, according to Hart and Risley. The gap actually increases over time even after the children are in school. Is the learning trajectory set in early childhood, or do differences in experience continue to widen the gap year after year?

SM: Great question. I believe the evidence is clear that the older students get, the smaller the impact the parents’ vocabulary plays in language development compared to other experiences (e.g., reading). However, children with lower vocabulary skills are often poor readers so they continue to fall further and further behind in academic language and cognitive skills. 

Language, and later reading experience, are two of the largest contributors to brain plasticity in the developing brain.  This is also a large contributor to the way we build our cognitive skills as well as our brain’s reward systems that contribute to our communication patterns. The most interesting part for me, if I’m a teacher, isn’t that I have to simplify my language use to reach students in my class, which is true for every teacher that works with first graders – they don’t talk like a 30-year-old, they talk in a way that uses simpler language structures.  Teachers present and provide language to students as a means of reinforcement and social support within the classroom.  For many impoverished students language was used more often to communicate negation, disapproval or punishment.

The data from the Hart and Risley study was that the average child from a professional family hears six times the number of encouragements for using language for every one discouragement or prohibition. This is critical. If there’s one thing, it’s this idea about using language to encourage vs. discourage further interaction and language use by the student. Saying “no,” saying “be quiet,” saying “knock it off,” those are ways to discourage and we want to use them for very important things as parents – don’t kick, don’t bite, don’t throw.

But when a child asks a question like, "What do I use this for?", professional families are using words of encouragement. "What would we use that for? Why do you want that?" Asking the child things that encourage more talking. Whereas in the middle class families, it’s two encouragements to every one discouragement. In families in economic distress (e.g., receiving welfare), we have almost a one encouragement to every two discouragements ratio. You really have to say to people, think about that for a moment. For that child, language is a way to be punished twice as often as it is to receive positive reinforcement or praise.

It’s not just that there’s a 30-million-word difference between us, but that I had language used twice as often to put me down, to make me feel like less, to make me not want to talk to you, and now I’m going to enter a classroom where the teacher primarily uses language as a way for communication, for grading me, for teaching me.

Then this child comes into a classroom and you think to yourself as a teacher, I know, I’m going to work with this child more, I’m going to ask them to wait a few minutes after class. I want to talk to them. Well that’s like saying, come up here I want to talk to you, and the child averts their gaze, drops their head, maybe shrugs their shoulder, walks up and is completely trying to avoid the oncoming punishment because that’s what talking is to them.

The numbers, if we estimate the frequency over years, are quite significant.  Students received 57,000 – 60,000 discouragements before they came to school, delivered by language. So by 60,000 times, my experience is that talking is not a good experience. 

SL: Are there studies about how much encouragement is required after school starts to counteract this?

SM: Yes. Using language – establishing a safe communication zone, which is what so many great teachers do – has that child eventually saying, "Wow, when I come to this classroom I love it here. I get to talk, to share my ideas, and nobody puts me down."  We learn to disagree in respectful ways.  Many teachers are already sensitive to this and their classrooms reflect a positive language environment.

Many, many teachers will say, "That explains why these kids always seek me out two or three years later to talk to me." They don’t have a lot of other adults in their life where language is used in a positive way. How do they explore the world in a way that they might get a reward or somebody might say to them, "Well, tell me more about that, why do you want to do that?" Who’s going to explore that with them? They go find that first, second, or third grade teacher and they want to hang out after school because they’re looking for that great, positive feeling. Why? Because they may not have it in other places in their world. The challenge is, how does language as the means and the media for safe communication become school-wide vs. teacher specific?

It’s extra work for the teacher to say to the other students in the class, "There’s no bad idea, there’s no bad conversation, nobody in here asks a bad question" – and we all know the teachers that say that. But they also have to enforce it. They have to make sure that kids don’t make fun of other kids.

I’m sure you’re aware from speakers that talk about the emotional elements of not being a good reader that reading out loud for them is a horror in their life. They’ll do anything not to expose that they’re not doing that well in the classroom. It’s a really big task for a teacher. Many teachers look at the Hart and Risley data and think about the 30-million-word gap, but I have to point their attention to the fact that the thing that they’re so gifted at – the tool that they love, which is language – is not really the tool of kindness for most of these kids and establishing that is probably the hardest task that they have.

SL: What are some possible solutions? How can the gap be reduced in early childhood and beyond? And whose job is it?

SM: Whose job is it? It’s going to be everybody’s job because it really is about creating more language tolerance and good language models for kids. However, I don’t think we’re going to mathematically catch them up by trying to find a way to sneak in the three years they’re missing, so to speak. I think the real solution – and I’m biased here – is that this is an area where technology can help.

This is an area where technology can give you experience with language, let you manipulate language, play it back for you in a safe, positive environment. In developing software, we want to have a six affirmation per one discouragement ratio. We want kids to be right the vast majority of the time. We want to make the content adaptive and we want the technology to interact in language with that child as much as that student can gobble up and tolerate. With products like Reading Assistant and Fast ForWord, we had an internal obsession with trying to get more word experiences – and the right word experiences – to kids per unit time than a lot of other software developers.

I also remind people that even in their own work environments – and often school districts will laugh when I tell them – with caller ID you can see that affirmation phone call versus that discouragement phone call coming in. How often do you answer the discouragement line? You let it go to voice mail, right?  And you say, "I’ll take care of that later." That should help them understand that for these kids, that is their exposure before they come to school. Twice as many discouragements for an encouragement in a lower volume environment. And then I say to people, "How many of you have a work environment where you’re encouraging to your colleagues two to six times more often than you’re discouraging?" And it’s the same brain mechanisms.

This really isn’t about genetics or about poverty in the formal sense. This is about a covariant element with poverty. For example, if I’m not making an income that allows me to sustain myself, I might have two or three jobs. I’m probably physically exhausted. I’m not eating right. I’m not sleeping right. These are all normal things that we expect people to do when they’re not generating adequate income to support their family. Do we really become reflective? Do we really ask those W, H questions [who, what, when, where, why, how] when somebody asks us? We’re going to be short and abrupt. We’re going to just make a determination and say, "Did you do something wrong today?" We deal with those basic safety issues and the language encouragement isn’t there.

I’m very encouraged by the research that Craig Ramey at Virginia Tech and others have done where they’ve shown that just educating young moms – even young pregnant moms – about what it means to communicate with their child results in an increase in vocabulary development for the mom, and in vocabulary development for their baby. And when they’ve actually taken brain scans of the babies, these babies when they’re older look more like babies from another income level. They didn’t change the income of the family. They just basically said, "This is how important language is." You want your baby to develop as much as they possibly can, and this is what language can do.

SL: Who should educate the pregnant mother? Should the doctor play a role?

SM: Great questions and we don’t have all the answers. If you remember, a couple of years ago in Georgia they used to give new moms a little pack of information from the governor. So people have taken the research, including the Hart and Risley data, and they have tried to make an impact. And the research from Craig and Sharon Ramey and their colleagues have shown that effective early childhood intervention studies can change the learning and economic outcomes for at-risk populations.  In some communities, the best conduit for educating the community will be a combination of the school district as well as talking to the pastors and the ministers who are going to be talking to their congregations.

We forget that some of the most effective social programs are the ones that support daycare for children. Again, not because we think moms and dads don’t know the language or don’t know the words. It’s really about the idea of providing a fun and supportive opportunity for communication using the best language you have.

You don’t have to know a lot of English. You can speak in whatever language you want to speak.  It’s about developing those cognitive skills. The more we talk, the longer the memory span becomes. The more we talk, the better our attention gets. So we have better memory, better attentional skills, and better processing skills with language. These are the skills, if you’re not practicing them, which are not so good when you show up at school. A teacher who has to work with 20+ students in kindergarten and first grade doesn’t always have an opportunity to go over and spend more time with the kids who, to be honest, are a year or more behind when they start.

SL: What’s the most important takeaway from the Hart and Risley study in your view?

SM: The most important thing is that this gap exists when kids show up at school. Number two is that, from a practical standpoint, we’re not going to fix it by supplementing education trying to recreate the missing vocabulary development. In other words, if the 30 million words equates to x number of hours of experience, we’re not going to catch the kids up because the higher oral language students won’t slow down. That’s what those rates of learning actually mean. Those kids that are ahead aren’t just ahead. They run faster in vocabulary and reading development.

SL: Are you saying that the achievement gap can’t be bridged?

SM: It can be bridged but you have to be more clever about how you’re going to present the information. Again, in my opinion, you’ve got to look at technology.  The kids aren’t just 30 million words behind. The rate at which they acquire and use words is also behind. So you have to look hard for ways to make them faster at acquiring vocabulary than normal. Otherwise they can’t catch up. The idea from Hart and Risley is there’s a hole in vocabulary development and usage, which has an impact on other cognitive and academic skills. But to fill that hole you have to be strategic.  You have to work harder and smarter.

SL: Do you think schools have absorbed that fact yet or are they still trying to solve the problem in the same way?

SM: They’re still trying to solve the problem in a very linear fashion. There are a lot of states that have recommendations. They’ll say, if the child is behind in reading, give them an extra 30 minutes a day. And there are two issues. First of all there’s an opportunity cost because something else is being pushed out of school. You’re already accepting lower performance in a different domain, so you’re going to read more at school if we just drop math or science, as an example. Something fell.

Number two is that you’re not changing the rate at which they acquire new information. So that 30 minutes is going to help – it’s better than nothing – but you’re probably not going to catch up under that model and that is why we have so few kids catch up. So again I’ll go back to my earlier statement: 92% of low-reading third-graders can be predicted in kindergarten. In the fall of kindergarten – 92% of them.  So are we only helping 8% or is that just error in our prediction?

What we’re doing is not a matter of effort. The incorrect answer is teachers need to work harder. The incorrect answer is districts need to work harder. The correct answer is we need to work smarter.

SL: Can you say more about how technology works smarter?

SM: Because we can create learning experiences, driven by the student, that exceed what can happen in a normal classroom or small group environment. In other words, think about five kids at different computers each receiving differentiated individualized instruction.

SL: It’s an individualized learning approach…

SM: …over a short duration that can work systematically on being faster learners, not just providing them with the experience. My comment about technology is really for public education systems.

The part that people misunderstand the most is that a lot of technology that’s presented in the first year to year-and-a-half of life actually predicts negative language growth, not positive, because you need those base skills. And so the best development for kids is really that interaction. You know after a year-and-a-half, two years of age, with adults, with other kids, then technology can play a role once they have those basic skills.

For parents, it’s not about putting a laptop in a crib and coming back two hours later. We want to make sure that people really understand that watching TV with your child and talking about what you’re watching, and letting them respond -- having that interaction the way you might use a book is exactly what you’re trying to do.

Further reading:

Meaningful Differences in the Everyday Experience of Young American Children (Hart and Risley)

The Social World of Children Learning to Talk (Hart and Risley)

The Read-Aloud Handbook (Jim Trelease)

Related reading:

As Classrooms Become More Diverse, How Do We Help All Students Grow?

Underachieving Students: Why They Struggle and How Educators Can Help