A psychologist’s 22-year project followed 26 children who spoke late at age two — until they were 17. Here is what she found.
Introduction: A Common Worry That Doesn’t Feel Common
Two years old, and your child uses just a handful of words. The neighbor’s kid the same age is already saying “Mom, give me water” — two words combined. Yours points at things, grunts, uses one word for the world.
Search the internet and you’ll see “50 words by age two” listed as a benchmark. The sleepless parent that night counts on their fingers what the child actually said today. 23. 27. Maybe 35.
Some say “they all catch up if you wait.” Others say “if you don’t act now, this lasts a lifetime.” Parents get lost between the two voices. So what really happens? Do toddlers who were late talkers actually catch up?
The most honest answer can come from someone who actually followed those children for 17 years.
A 22-Year Follow-Up: Rescorla’s Patient Cohort
Leslie Rescorla, a developmental psychologist at Bryn Mawr College, has done one stubborn thing since the late 1980s: recruit a group of toddlers whose speech was clearly behind their peers, and revisit them, decade after decade, with the same tests.
Ages 3, 5, 6/7/8, 9, 13, and 17. Following the same children with the same instruments — this kind of work is called a longitudinal study. It asks not about a snapshot, but about a person’s time itself.
Her 2009 paper is the final chapter of that 22-year follow-up. What did those late-talking 2-year-olds look like at 17, on the verge of college applications?
The Core Research Question
Plainly stated:
For children whose expressive vocabulary at around age 2 was markedly behind peers — by adolescence, are they at the same level as peers in language and reading? And if there is a gap, how big is it?
Translated for parents: was “they catch up if you wait” the truth, or was “if you don’t catch it now it lasts forever”? Rescorla wanted to put data, on one hand, against the question.
Study Design and Participants
The beauty of this study is in how the comparison group was built.
Late-talker group (LT): Children seen at 24–31 months whose expressive vocabulary was under 50 words or who couldn’t combine two words. Critically, nonverbal cognition was normal, and receptive language (understanding speech) was also normal. By age 17, 26 children were still being tracked.
Typically developing comparison group (TC): Same window, same socioeconomic status (SES), same nonverbal cognitive level — 23 children. At first contact, language was on track.
The important point: both groups grew up in socioeconomically similar families. This isn’t “wealthy vs. poor.” Same environment, same cognition — only the early speed of speech differed. Two cohorts compared across 17 years.
Adolescent measures used standardized instruments. Vocabulary/Grammar, Verbal Memory, Reading/Writing — combined into a factor score for age-17 language ability.
Result 1: All 26 Were in the Average Range
The warmest single line this study can offer parents:
All 26 children who had been late talkers at age 2 were, at 17, in the average range on every language and reading test.
“Average range” means falling within the typical band of the general population on standardized tests. None of these 26 fell into the territory where vocabulary couldn’t keep up with school, or reading struggled enough to disrupt friendships.
Those late-1980s 2-year-olds, 22 years later, preparing for college applications — were speaking, reading, and writing within normal ranges.
Just that one fact answers half of what a parent counting words at 24 months wants to know.
Result 2: Still, They Were Slightly Below Same-SES Peers
The study was honest. It didn’t stop there.
The 26 were in the average range, but compared to the same-SES comparison group (TC), they scored statistically significantly lower on vocabulary/grammar and verbal memory.
What does that mean? The TC group, raised in similar environments and on track from the start, sit somewhat above the general-population average — the effect of a good environment. The late talkers reached the average band but did not reach the level “expected for their environment.”
By analogy: starting later, they ran for 17 years and got onto the main track, but their same-environment peers had moved one step further during that time.
Result 3: Did the Age-2 Score Decide the Age-17 Child?
This is the most interesting part of the paper. The team asked: how much of the age-17 vocabulary/grammar score is predicted by the age-2 vocabulary measure (LDS)?
The age-2 vocabulary score explained 17% of the variance in age-17 vocabulary/grammar.
The number alone may not look small. Flipped, the meaning is clear: the remaining 83% was determined by something other than the age-2 score.
That “something else” is everything that came after — the environments, books, conversations, schools, friends, the child’s own developmental pace, the parent’s daily responses.
The vocabulary card at age 2 shows part of the future, but it does not by itself decide age 17. That is the most honest thing this study tells us.
The Warmest Finding: A Continuum, Not a Diagnosis
The paper’s subtitle — Support for a Dimensional Perspective on Language Delay — is the most accurate compression of its conclusion. A dimensional view: think “spectrum.”
The team writes:
Late talking is not a question of crossing a line between normal and abnormal. It is a point on a continuum that moves at that child’s pace.
In plain terms, an age-2 vocabulary count is not a diagnosis. It is one point on the child’s language curve, and that curve moves upward over time. Same-environment peers move too, so the late-talking child may not pass them — but the curve itself goes up.
This is neither “nothing to worry about” nor “trouble.” It’s the simpler truth that the age-2 word count is not the single variable that decides the rest of life.
Practical Implications
A late-talking 2-year-old isn’t a “late child”
For a child with normal cognition and normal receptive language whose expressive vocabulary is delayed, the chance of falling out of the average range at age 17 was, in this study, 0%. The age-2 word count is part of the child, not the child.
But “they all catch up” is also not the message
The small gap to peers persisted across 17 years. That doesn’t mean the good things you can offer in this period — reading, rich conversation, responsive interaction — don’t matter. Quite the opposite. The 83% that accumulated after age 2 is much bigger than the 17% set at age 2.
Look at the curve, not the label
Rather than the label “our child is a late talker,” it’s more accurate to ask “where on the language curve is this child right now, and which direction is it moving?” Diagnosis is for clinicians. For day-to-day life, curve-thinking is more useful.
Limitations
The team’s own caveats:
- Small sample: 26 late talkers, 23 comparisons. Generalization is limited.
- Upper-middle-class focus: most participants came from U.S. upper-middle-class families. SES-disadvantaged contexts may show different trajectories.
- “Pure” late talkers only: only children with normal nonverbal cognition and normal receptive language were included. Children with co-occurring receptive delays or cognitive concerns may follow different trajectories.
These limits don’t reduce the study’s value. They define which children the results apply to.
Closing
In one line:
A list of words written at age 2 does not decide who that child will be at 17.
The night a parent counts words on their fingers at age 2 is one day in the child’s 22 years. The weight of that day’s number is not nothing — but it doesn’t draw the entire curve.
This isn’t to say raising a late-talking child is easy. It is to say the weight a parent carries through that period might be slightly lighter than it feels — at least, that is what the data say.
Source: Rescorla, L. (2009). Age 17 language and reading outcomes in late-talking toddlers: Support for a dimensional perspective on language delay. Journal of Speech, Language, and Hearing Research, 52(1), 16–30. https://doi.org/10.1044/1092-4388(2008/07-0171)