by Emily Oster
These studies focus on delayed talkers. Within the normal range, we have less work, but there is at least one 2011 paper entitled “Size Matters” (I guess it’s funny?) that compares children who are earlier versus later talkers at the age of two.7 Their “later-talking” group had an average of 230 words at age two, versus 460 for the high talkers. These are different portions of the distribution, but in the normal range.
Studying kids through age eleven, this paper again found lasting differences across the groups, but there was a lot of overlap. To give a sense of this: On one measure of language ability (something called “word attack”) at grade 1, the later-talking group had an average score of 104, versus 110 in the early-talking group. Clearly, the early-talking group was doing better. But there was also a huge amount of variation within each group.
The following graph gives a sense of the range for the two groups.8 On one hand, we can see the (on average) higher scores in the early-talking group. On the other hand, there is a tremendous amount of overlap in the distributions. The individual variation completely swamps the difference in averages.
What about really exceptional language ability? Again, we see some small-scale evidence that being a precocious talker correlates with precocity later.9 But this correlation is not enormous in this or other studies, and being a very good talker before two is by no means a decisive determinant of early reading or other achievements.10
It is natural, probably unavoidable, for us as parents to want to compare our children to others. Language development is among the first cognitive processes that we really see in kids, so it is not surprising that it becomes a focus of comparison. And if you are really curious, it’s definitely possible to use the data here to do some more concrete comparisons. But it is crucial to keep in mind that the predictive ability of early language, while there, is really quite poor. Early talking doesn’t guarantee later success—even at four—and late talkers mostly look like everyone else within a few years.
The Bottom Line
There are some standard tools to determine child vocabulary size, which you can use on your own. There are also some metrics you can compare.
Girls develop language faster than boys, on average, although there is a lot of overlap across genders.
The timing of language development does have some link with later outcomes—test scores, reading—but the predictive power is weak for any individual child.
16
Potty Training: Stickers vs. M&M’s
My mother is fond of telling the story of how I was potty trained. “When you were twenty-two months old, one day you announced you would now be using big-girl underwear. That was a Friday, and on Monday, I brought you back to day care without diapers.”
This story is implausible on several levels (the announcement, the speed of training, and so on). When she first told me this, I also thought the age was impossible. Twenty-two months? I think not. And it is true that usually when we return to her written notes on topics like this (yes, I know that not everyone’s mother kept detailed written notes—it’s a family affair), she is often revealed to have exaggerated. However, in this case, she did not. Her notes from the time suggest, largely without comment, that I was wearing underwear by this age.
Not to be outdone, my mother-in-law insists Jesse was potty trained by eighteen months, and pooped in the toilet at thirteen months. She also suggests this was pretty typical.
But I distinctly remember that my younger brother (sorry, Steve) was not potty trained when he started preschool at age three. This was unusual at the time, and was a source of a tremendous amount of parental anxiety.
The question of when to potty train remains a source of stress for parents. Should you push your child to train early? If you do, will you stress them out? If you don’t, will they be somehow behind?
And the experience of our parents’ generation, and therefore of the grandparents who are speaking over our shoulders, doesn’t seem necessarily typical now. My brother, a late potty trainer by the standards of the time, seems like he would be quite typical now. Training at eighteen months—especially for a boy—sounds like the more unusual thing.
This is, however, only my casual impression, and I was curious if it lined up with any actual data. I decided to be a bit more systematic. In other words, rather than just asking my friends, I ran a survey. I sent it to my friends, their parents, their parents’ friends, people on Facebook and Twitter—basically anyone I could find. I asked a few simple questions: When was your child born? And when were they potty trained?
The first graph on this page shows the average age of potty training by birth timing in my survey.1 And, indeed, the average age has crept up over time—from thirty months for births before 1990 to more than thirty-two months in the most recent period. But perhaps even more notable is the second graph, which shows the share of children who are trained at or after thirty-six months (that’s three years). This was only about 25 percent of the children in the earliest birth years, but 35 to 40 percent in the most recent period.
Of course, this is not exactly a scientifically valid sample. It definitely would not pass peer review. But the casual impression—and the findings from this data—is supported by the literature. Studies from the 1960s and ’80s show an average age of twenty-five to twenty-nine months for daytime toilet-training completion, and virtually all the children were trained (for the daytime) by thirty-six months. In contrast, in more recent cohorts only 40 to 60 percent of children have trained by thirty-six months.2
This suggests toilet training is occurring later. Why?
The Journal of Pediatrics published a study in 2004 asking this very question.3 The study enrolled four hundred children when they were about eighteen months old and followed them through potty training. They found three factors significantly associated with later training. The first—and probably the one that explains variation over time—is later initiation of potty training. Children who start training later complete training later.
The other two factors relate to poop. Children who were frequently constipated, or who showed resistance to pooping in the potty (formally “stool toileting refusal”—more on this to follow) tended to train later. The authors argued that these factors could also increase over time, although they largely put the blame for later training on the later initiation of the process.
It is interesting to speculate why people have started training later in recent years. My mother insists it relates to diaper quality—diapers used to leak a lot, which made it much less fun to use them. The generation born in the late 1970s and early ’80s was the first to commonly use disposable diapers, perhaps due to innovations in the early 1980s that dramatically decreased the size of disposable diapers.4
Income may also play a role. People have become, on average, richer over time, and the inflation-adjusted price of diapers has gone down. This may make a prolonged period of diaper wearing more acceptable, although affording diapers is still a challenge for many families.
There is likely also some feedback loop. If everyone potty trains their child when they turn two, people may feel some social pressure in this direction. If everyone else waits until three, that becomes the norm. This may also affect when, for example, day-care centers push potty training.
Regardless of why this occurs, the fact that later initiation correlates with later completion suggests that it is, in fact, possible to potty train your kid at a younger age. Should you do so?
The main and probably only benefit to potty training a child earlier is that you do not have to change as many diapers. The main reason to wait is that the earlier you start, the longer it takes to complete. We can see this in the same data described above, with the 400 children starting at 18 months.
The first graph on this page, re-created from their study,5 shows the age of completion of potty training as a function of the age of initiation (
both of these are as reported by the parents). Here they define the age of initiation as the first age at which parents try to train their kid—as in, asking the child at least three times a day if they need to use the potty. And the age of completion is when the parents say the kid is fully trained in the daytime.
What we can see is that the age of potty-training completion is similar starting anytime between twenty-one and thirty months. The second graph shows the duration of potty training—the earlier you start, the longer it takes to complete. A somewhat depressing aspect of this graph is that the duration of training is about a year if you start young.
The authors suggest that if what you care about is when potty training is done, there is not much point to starting before twenty-seven months or so. But it is the case that after this, starting earlier does generally mean finishing earlier. If you start training at twenty-seven or twenty-eight months, you can expect to be done by around age three, but it will take ten months to do it. If you start at age three, you finish later, but it’ll likely take you less than six months to fully train.
As we think about the contrast between doing this at age two and doing it at three, it is worth thinking about the ways in which your two- and three-year-old are different that makes this harder or easier. A three-year-old has a lot more control over their bathroom functions (also maybe over you, but that is a different story). This is partly physical, and partly emotional. An eighteen-month-old is much less likely to simply decide they will not poop in the potty no matter what you say. They have less will to defy you. This may make littler kids easier.
On the other hand, a three-year-old can be reasoned with and—yes—bribed. They have more will to defy you, but you can also take advantage of their better ability to understand and control themselves. That may make them easier to deal with. The data on timing suggests that, on net, the latter probably matters more.
METHODS
Having chosen a time to start your potty-training adventure, there is then the question of how to do it. Broadly, potty-training approaches come in one of two forms.
First, there is the parent-led, “endpoint-oriented” potty training.6 These methods are discussed in books like Oh Crap! and 3-Day Potty Training. In general, the idea is to just take the diapers away and start putting your child on the potty a lot. Ideally, within a few days they are (mostly) trained. There are less intense (and more intense) versions of this as well, but they all share the same basic structure: parents decide when it is time to potty train and then push toward the end goal. Based on the timing data we saw earlier, either most people do not use these approaches or most people do not use them successfully.
(I have promised my children that I will not discuss their potty-training adventures in detail here, but I will say that we did use this method, and broadly I was happy with it. However, it worked better for one child than the other, and we definitely did not achieve full success in three days.)
On the other side is a more laissez-faire approach, where you more or less let the child lead with the timing that works for them. This approach involves looking for signs of readiness and encouraging toilet use when they become apparent. This is goal oriented in the sense that ultimately you would like the child to use the toilet, but it does not work on the same time frame.
There is a third approach, “Elimination Communication,” which tries to have kids use the potty from birth. More on this in a bit.
These approaches were developed many years ago—the original description of child-led potty training is from 1962. A major difference between them is the age at which they seem plausible—generally, using a child-led approach will involve starting later.
There is virtually no data on which of these works better or even how well any individual system works.7 To the extent there are studies on this, they are extremely difficult to interpret. Consider, for example, one study of twenty children (twenty!), which considered a potty-training intervention in a preschool classroom.8 The program contained three different interventions (using underwear, making the child sit on the potty frequently, and rewarding potty use). With a subset of the children, researchers had the preschool teachers do all three. With others, they did them sequentially.
Some children improved, some did not. There were virtually no consistent associations. The best the study authors could say was that many of the children who wore underwear seemed to improve. And, perhaps most important, that all the children were eventually toilet trained.
There are other small studies. One, of thirty-nine children in the UK, compared a wetting alarm method (where children are fitted with a special diaper that plays an alarm when they pee in the diaper) with a method of putting the child on the potty at regular intervals. They found evidence that the wetting alarm was more effective, but again, this was a small sample and not a comprehensive study of particular approach. Also, clearly the alarm approach won’t be for everyone.
If you are desperate for some evidence-based guidance, one randomized study of seventy-one children from 1977 compared a child-led approach to an intensive one.9 The study argues in support of the more intensive method, showing that accidents per day declined more in the intensive training group, and successes went up more. But this study is very old and small, and didn’t look at any other outcomes for the children (for example, were they stressed out about the training).
The main thrust of the literature on this is that we simply do not know much about the best methods, if there even is a single best method.
This last point is probably the most important: there may not be a single best method for every child or every family.
When my twin nephews were being potty trained, my mother made a book to read to them, entitled The Lion Gets Potty Trained. It featured a series of pictures of my niece (their older sister) and a stuffed lion. The theme is that she was attempting to train the lion to use the toilet using various rewards—M&M’s, Skittles, a kumquat, etc. Finally, she is successful by offering the lion a meatball.
I have read this book to Finn many times, and in many ways it really epitomizes this experience. You will try anything—literally anything!—to get your child to use the bathroom, but you cannot actually force them. And, probably most important, all kids are different. Some kids respond to stickers. Some respond to M&M’s. Maybe some respond to meatballs.
The bottom line is that potty training is really all about what works for your family and your kid. The evidence on changes over time suggests it is possible to train your child at a younger age than is now typical, if you want to. To do this, you’ll probably have to adopt a more goal-oriented approach (rather than a child-led approach). Or you can wait until your child decides they are ready, which will probably be when they are closer to three years old or even a bit older.
The child-led approach to training may take longer, but it also may be more pleasant for you. Or maybe this is your last child, you are totally over changing diapers, and you want your twenty-five-month-old to get with the program. If this is the case, your best bet is probably to try an intensive, goal-oriented regime and see if she takes to it.
There is no evidence linking age of potty training with any later outcomes like IQ or education.10 So if your child is trained early, that might be great (for you) but irrelevant in the long term. It may be hard to see through the haze of running your child to the bathroom every twenty minutes and cleaning up the poop in their underwear, but everyone does eventually use the toilet.
PROBLEMS AND EXTENSIONS
Stool Toileting Refusal
At some point, before Penelope was born, I had a conversation with a friend about her son. How are things going? I asked her. “He’s doing well, although of course we are dealing with an STR issue.”
“A what?”
“Oh, stool toileting refusal.”
This was my first exposure (but not the last) to the seemingly widespread “STR” problem, which
I continue to feel is a great name for your child not pooping in the toilet.
This problem is surprisingly common (or, rather, surprising for people who do not have children yet). Perhaps a quarter of kids will experience some degree of this during potty training.11 As weird as it might sound, a lot of kids like to poop in their diaper. Children who will successfully urinate in the toilet will nevertheless refuse to poop there, and unlike urine, bowel movements are something over which even young children do have some control.
When the refusal to have a bowel movement in the toilet continues well past when kids regularly urinate in the toilet, this rises to the level of a problem. The main issue is that withholding poop can cause constipation. This can lead to painful bowel movements when they finally arrive, which further exacerbates the problem. Now the child associates using the toilet with pain and really doesn’t want to do it. Chronic constipation can also lead to problems with urination.
There is some work studying how to address this issue in older children—stool withholding is also a common issue in school-age children—but virtually nothing systematic in younger ages.12 One study of four hundred children, published in 2003, showed that the length of refusal (i.e., the number of months this goes on) decreased with a child-oriented intervention where, among other things, parents made a big deal about the child pooping in the diaper before potty training started.13 This means saying things like, “Wow! You pooped! That’s so great!” and so on. The kids in this treatment were no less likely to have the problem at all, but it lasted for less time.