An Economist’s Guide to Potty Training

Children, it turns out, can work out how to do lots of things on their own. Crawling and walking are obvious candidates. Eating is also right up there. But some kids teach themselves to do much more complex things, like reading. This, of course, makes one wonder what all the parental effort is for. Usually, the best that can be said is that it speeds things up a little.

Nothing, however, taxes a parent’s energy like toilet training. If a child reads a year later, it usually doesn’t affect much. But taking a year longer to learn how to go to the toilet is very costly. Even aside from the obvious costs of diapers and various cleaning routines, there comes a time when a parent feels enough is enough (usually somewhere around the six thousandth diaper change). After attending to your child’s personal waste for years, you feel it is time for some independence.

It is a rare case, indeed, that a child works this out for themselves. The best example I heard of was a mother who noticed that her six-month-old (!) seemed to be doing a ‘‘number two’’ at the same time every day (because of the smelly ‘‘insult,’’ as they term it in the diaper business). So, to save herself some trouble, she’d hold the child over the toilet. This went on for months while the child still wore diapers. The diapers were never soiled, but the mother thought it was just luck.

Then one weekend they went away and left the routine at home. The child didn’t go all weekend (much to her parents’ alarm), which was how they worked out that, at least for the worst insults, she had been trained.

For normal children (that is, all the rest), there is no such luck. The pressures to act build up quickly. At two or so years, your child is starting to look like a real person. And, let’s face it, in a short while that diaper bottom will start to look obvious. Soon enough, they’ll need to attend kindergarten, before which they need to be certified as independent, toilet-trained individuals. Later on there are school, sleepovers, and marriage. If the cord isn’t cut, none of those things will be available to them. But here you have a child with no interest whatsoever in abandoning the cushy, full-service attention they have enjoyed their entire life. They just don’t understand the urgency here. The clock is ticking.

At this point someone pipes up, ‘‘I think we should start potty training junior this weekend.’’ Then the other one says, ‘‘Let me look at my schedule. Hmm, I have to travel for work, but it sounds good to me!’’

So begins a discussion of the plan, the timeline, and the incentives. Even if you somehow delude yourself that you don’t need incentives, your child generally has other ideas. It is all much easier said than done.

One issue I should raise right at the start is that this is all partly the fault of the comfortable disposable diaper. In the olden days — say, three decades ago — cloth diapers were the repository of choice. These had the feature that when they were wet, everyone knew it. Parents knew it, but the child also felt the discomfort.

As the child grew older, the discomfort grew with them. The causal link between what they were doing, in a bodily function sense, and the discomfort was felt early and often. Thus, not only did parents want their child out of these things quickly, the child was on board too.

The disposable diaper has changed that relationship. It is a remarkable piece of technology. Author Malcolm Gladwell believes that it is an achievement that ranks alongside the microchip:

We tend to credit those who create an idea, not those who perfect it, forgetting that it is often only in the perfection of an idea that true progress occurs. Putting sixty-four transistors on a chip allowed people to dream of the future. Putting four million transistors on a chip actually gave them the future. The diaper is no different. The paper diaper changed parenting. But a diaper that could hold four insults without leakage, keep a baby’s skin dry, clear an insult [which is actually what you think it is] in twenty seconds flat, and would nearly always be in stock, even if you arrived at the supermarket at eight o’clock in the evening — and that would keep getting better at all those things, year in and year out — was another thing altogether. This was more than a good idea. This was something like perfection.

It is like perfection because while the technology does its job, it also affects the parent and the child. The parent, not surprisingly, wants to employ a disposable diaper right away. Sometimes they buy the propaganda that not having a cloth diaper might lead to training issues later on. We did this for the first few days of Child No. 1, but we don’t talk about such foolishness anymore.

But for the child, it is comfort that can continue forever. When they are told it is time for the diaper to go, the expression on their face says it all. They appreciate the beauty and functionality of the design. Perhaps they also suspect they will be wearing one again in 70 or so years’ time. Why deny them in the interim?

The unintended consequence of the disposable diaper is that it changed the ‘‘toilet or not’’ cost equation for the child. Incentive issues arise from a misalignment of interests. The disposable diaper was something both parent and child wanted, but when it comes to taking it away, their relative benefits are quite different.

Now this issue has occurred to many a parent at the toilet-training stage of the modern child. I, for one, wondered whether an electrically fitted diaper could be constructed that would administer a short, sharp shock when wet. Others have suggested a powder that becomes itchy when wet. Somewhat less torturously, the latest training pants (that is, diapers a child can pull up and down by themselves) have a liner in them so that a child can feel the wetness immediately. That may help them associate the physical act with its result, but it’s still far from incentive alignment.

No, if you want to take the diaper away, these days there are few options but to manage the whole exercise. And that requires thinking about the incentives.

In our three experiences with all this, we have tried various things and each time been presented with new challenges. The notion that no two children are alike — despite genetic similarity — comes to the fore here.

Toilet training is an exercise in behavior modification: you try to convince an otherwise happy and contented child that they have to take responsibility for their own actions — namely toileting. It is a classic situation where the interests of one party (the parents) differ from those of the other (the child). If you want to align those interests, someone is going to have to pay up. The only question is, how much? What reward do you need to offer to get the behavior you want?

Enter Child No. 1. Despite my economics background, we decided that, initially, we would bid low and see if we could get away with it. We thought we would appeal to some broad, vague sense that it would be good to grow up and wouldn’t that be ‘‘exciting.’’ Not surprisingly, those claims had about the same impact as calls for whiter, brighter towels or peace on Earth, or what have you.

So it wasn’t long before the price started to rise. But how do you engage in shameless bribery with a two-year-old? There isn’t a lot to work with there. While money can appeal by about the age of eight, for the younger ones you have to appeal to a base motive. All we could think of was food; in particular, ‘‘special’’ food.

It was fortunate that we had decided at some earlier point — several weeks earlier — to make a distinction between general food (food that we would always allow our daughter to have) and special food (food that we would virtually never allow her to have). Into the former category fell rice cakes, fruit, yogurt, chicken, and all vegetables. Into the latter category fell just about everything else. So now we had lots to work with. But here’s a hint for the unprepared: engage in some pretraining deprival. It creates its own reward.

We decided to begin with jelly beans. The basic reward was to give out jelly beans for successful toileting behavior. You got one or two jelly beans depending upon what you did, with one for  .  .  .  and I’m sure you can guess the rest.

But the whole scheme went further. You see, the jelly bean may be the reward, but game theory teaches us that unless the ‘‘incentive contract’’ (that link between action and reward) is clearly communicated, it won’t work. So we didn’t simply have a bag of beans that we took out of the cupboard when required. Nor did we rely on a pretty chart to record success. Instead, we installed a whole publicly displayed apparatus. It stood atop a kitchen counter. When you wanted a jelly bean, you had to push a button. This started everything but a song and dance, following which a single jelly bean was dispensed.

Now this is all very well, but right at the beginning the child isn’t going to engage in any behavior that demonstrates how the whole deal works. Demonstration is the key. To overcome this, we decided to apply the reward universally. That is, anyone — including us and our guests — would be entitled to a jelly bean as they emerged from the bathroom.

For weeks, I would emerge from the bathroom to find my daughter standing outside, asking me if I was allowed one or two jelly beans. If we forgot to take one, we would be reminded. Outsiders could monitor the whole household’s performance as they came in: ‘‘Looking a bit empty there; better cut down on the coffee.’’

Did this work? It certainly put the whole issue on the table, and our daughter showed much interest in spending time on the potty. But, after a couple of weeks, we had little to show for it except a personal dislike of jelly beans, as their consumption came to us to be associated with unpleasant activities. We made a conscious effort to disguise our toileting activities to avoid having to eat any more!

At this point, we had to contemplate a ratcheting up of the reward — to chocolate frogs. This was a decision we did not take lightly, as we were well aware that at some point we would want to end our little incentive contract. But we knew that if anything was going to get the behavior we wanted, it would be this; we could work on our exit strategy later.

To a casual observer, it did the trick. While precious few jelly beans had been awarded to our daughter, she was now getting two or three chocolate frogs a day. Moreover, we started saving on diapers.

Alas, a more careful audit of our household performance would show dubious results. Our daughter realized that there was a very easy way she could acquire a frog. In her time-rich life, she would simply wait. Basically, she worked out that if you sit on the toilet long enough, something will happen. And so she did this, for hours and hours on end.

Now we let this go on for a while in the hope she would get a feeling for her own bodily functions. But it was starting to get worrying and to interfere with normal daily activities. So when she didn’t appear to tire of these activities, we changed the reward. She was only allowed to sit on the toilet for short (!) half-hour bursts.

This all worked and, after a week, she began to anticipate her needs. In a flurry of activity we could make it to the toilet, and there was much rejoicing. Although that gave rise to a new issue. Our daughter worked out — as many kids do — that this was a way to instantly gain attention.

I recall one such incident at a department store. I was buying some clothes, and my daughter announced that she had to go to the bathroom. Her mother scooped her up immediately and asked the sales clerk where the toilet was. She was told that she would have to go outside into the mall. ‘‘Well, that is just great. You are going to end up with wee all over your floor.’’ And then she fled, having pushed the hapless clerk out of the way in violent protest.

Upon their return, some 15 minutes later, it was discovered that all this was a false alarm. It was then that we developed a greater sense of calm about all of this, including how to cope with the same ‘‘I need to go’’ announcements in the car on the freeway. What can you do? It turns out that she always held it in.

Nonetheless, during all this training, no sooner had we fixed one problem than another began; she got too much control! Our daughter realized that by holding back, she could convert one trip to the toilet into two or three, and thus triple her frog consumption. The economist Tim Harford likens this effect to the way pole-vaulter Sergei Bubka, who was paid a cash bonus each time he broke the world record, chose to do it a centimeter at a time. That’s the risk you face when you set down clear, objective rules for rewards: you often get what you pay for.

There are two reactions to such gaming. If our daughter was an employee and we were trying to reward her for performance we wanted, we might promise a reward if she obtained a ‘‘favorable’’ evaluation. The problem here is that, to avoid gaming, the evaluation cannot be clearly defined. You actually want it to say something like: ‘‘You get a chocolate frog when you go to the toilet in an appropriate and sustainable manner.’’ Suffice it to say, employees don’t often buy this type of deal, let alone children. Discretion makes those being rewarded wary of your commitment to paying up.

So we opted for an alternative course. We decided to phase out the reward program. This was not done without some fear: what if removing the system eliminated all our good work? But we did it slowly, first by increasing her quota (the period of time over which she had to exclusively use the toilet before getting a frog), and then removing the reward entirely.Suffice it to say, there was no reversion, and all in all we were happy with the outcome. Eventually, the motive we initially hoped for — getting out of diapers and into more exciting underpants — took over. But the management process was painful, and I can’t prove whether this wouldn’t have all happened of its own accord anyway, without rewards. Such is the nature of the first child.

Now you might have thought that all of this experience with Child No. 1 might have meant that things would be a little easier with Child No. 2. One thing we planned on doing was waiting a little longer before we started toilet training, and checking for more signs of readiness. Then we would go for a ‘‘big bang’’ behavioral change.

Alas, training our son proved to be a very different ball game from training our first daughter — and not simply due to gender differences, which some believe are significant in these things. You see, our daughter possessed several qualities that helped us when applying incentives. First, she was highly strategic and easily understood what rewards meant. Second, she had some basic motives — most notably food — that made it easy to give her high-motive rewards. Put simply, she was her father’s child.

Our son possessed neither of these qualities. He was not strategic and hardly had a self-interested bone in his body. He didn’t care about much that was material. For him, toilet training was something that was going to please us, and for that reason he was interested. Read him a book and he would stay on the potty. The attention was enough. This all seemed to suggest that we might be able to get results for a song.

Nonetheless, there was very little progress. So we decided to implement the more explicit incentive system, which had worked (kind of) well with Child No. 1. But there was a twist: Child No. 1 was now around, and in many ways, we needed her help. She could easily detract from our efforts if not fully on board.

So the new plan became this: Child No. 2 would receive a sweet reward — one or two jelly babies, as the case may be. But also, whenever he was successful, Child No. 1 would receive the same. We viewed this as a team effort, and Child No. 1 was part of the team. To align her incentives we gave her a share of the pie.

That part worked swimmingly. She encouraged our son to sit on the potty and spent time showing him books. It seemed that we may have efficiently outsourced this activity: something valuable in our time-strapped lives.

Alas, it was not quite to be. We had to put a stop to it all when we discovered Child No. 1 feeding Child No. 2 copious amounts of water to help the process along! (By the way, in case you’re wondering, Child No. 1 was four years old at the time.)

Nothing seemed to work with Child No. 2, and a deadline was looming. He was about to turn three and could go to preschool so long as he was toilet trained. Three weeks away from his birthday, we were desperate (Not just for us but for him. Preschool was much more fun than day care.)

So we established toilet training boot camp. The diapers were confiscated. The rugs were removed from our wooden floors, and we would do this as long as it took. He understood the basics but just needed intensive experience — so that is exactly what we provided.

Now you may think this sounds somewhat cruel. Well, I did some research, and apparently this was nothing compared with the methods of Nathan Azrin, whose book is tantalizingly entitled “Toilet Training in Less Than a Day.” This book, published in 1989, ranks in the top one thousand books sold even today. But if you read the reviews, the results are definitely mixed; even fans say the method takes a lot more than one day. I didn’t buy the book, and we persisted with our methods; if we had to, we would come back to it as a last resort.

The first week was tough: he definitely wanted his diaper back, and we had to buy more floor cleaner. The second week was more successful — 100 percent results on number twos but still some problems with number ones. We were getting close. Then, in the final week, he got it.

We turned up to the preschool and announced our success. They then told us that training pants would have been just fine. So much for rules. Sigh.

In the end, it was the attention and intensive experience that did it, but all this indicates something else, too: it will happen when it happens. Sadly, we were not done yet. Incentives came back in force when we moved on to night training. I’ll take a breath and recount that can of worms next.

Night training is a particularly difficult business. Your child has the day down pat, but at night, they might be asleep or too drowsy to know the signs. That means a period of time when they wear diapers or training pants at night. For parents, success begins when these are dry in the morning, and the training is complete when they are dry for a week or more.

In relation to Child No. 1, our philosophy was: what doesn’t go in, doesn’t come out. Her bed time was 6:30, so for two hours prior to that she was mostly off fluids. This wasn’t a conscious plan, and we fell upon it by accident. But it worked, and we achieved official night training. (Actually, it turns out that we were just lucky. Whatever chemical processes needed to work did work, and the routine of before and after bed seemed to stick.)

We were not so lucky with Child No. 2. For starters, he got thirsty, and it is not advisable to deny a child water. But also, the concepts didn’t quite click. He didn’t seem too bothered about the whole thing, and we noticed that he would get up dry, claim he didn’t need to go to the toilet, and then go in his diaper; or later, his training pants. This smelled (literally) of the basis of an incentive problem: once again, his interests were not aligned with ours.

Training pants have little pictures that disappear if ‘‘accidents’’ occur. This gave us a visible and external monitoring device. It was viewable both to us and our son. So that is the first thing we would check in the morning, and a celebration would ensue if the pictures were still there.

As you can guess, celebrations only get you so far. So, as has been our pattern throughout all of this, we moved to more tangible rewards. He was old enough to understand a points system that would lead to rewards. So a dry night would get a point, and seven points would get you a reward — usually a book or a toy. This was sufficient motivation, and he had a clear focus: ‘‘Make sure the pictures don’t go out and you get a point.’’

Well we had some good nights and intermittent accidents. But then we had a full week of dry training pants. Much rejoicing ensued, including a bonus; no more training pants. Sadly, the next night there was an accident. ‘‘These things happen,’’ you might say. But it turns out, the problem was that these things hadn’t happened.

Our son had a small rubbish bin in his room. Upon inspection, we found five full training pants. Child No. 2 was getting up in the morning, noticing the pictures were gone, and getting himself a new pair of training pants! There was nothing malicious in this. He just understood the rule as: ‘‘Produce training pants with pictures.’’ And so he worked out how to do just that.

Once more the old adage ‘‘you get what you pay for’’ was raising its ugly head. We paid for dry training pants, so that is what we got.

Our response was to impose a new requirement: you must have the same training pants on in the morning as you did in the night. It was easy to monitor, and we did.

We got a couple of nights of success, and then one morning I went into his room and found his bed wet. His training pants were dry. I asked him about this and he said, ‘‘It just happened.’’

‘‘But how? It should have wet your training pants.’’

‘‘No it wouldn’t. I didn’t have them on. They were on the night stand.’’

It turned out that Child No. 2 had been removing his training pants to ensure they were dry in the morning. I guess that worked. And we didn’t notice because the relevant part of his body was concealed under the covers.

When it comes down to it, giving children incentives is a bit like programming a computer. Unless you get the instructions just right, problems can ensue. There is an episode of Star Trek: The Next Generation where Geordi programs the holodeck for a game ‘‘that could defeat Data’’ (the android) and ended up creating a sentient program that almost destroyed the Enterprise. Programming our son posed the same challenges.

We focused on the pictures on the training pants, so that is where our son placed his considerable creative energy. What we needed was a ‘‘program’’ that gave exactly what we wanted: no accidents. That is what we turned to after that night. It took some months, but eventually we were successful. We moved our focus away from the training pants (although getting rid of them became a common incentive as he grew older) and onto the behavior, which is what we really cared about.

So, just as we found with Child No. 1, incentives can work — but sometimes they work too well. So much care and management is required.

Finally, after eight-and-a-half years of this, we arrived at Child No. 3. And what had we learned? We’d realized that we were not equipped for the job — so we acted accordingly. We adopted a strategy, ruthlessly efficient in its application and very light in terms of taxing our own energy: we outsourced the whole deal.

Now, by outsourcing, I don’t mean that we just sent our daughter away to some service and then they delivered her back to us, ready to go. Apparently those do exist for dogs, and I won’t gloss over the fact that we wouldn’t have availed ourselves of a human service had it existed. But it does not. Instead, we relied on her day care providers to handle the entire exercise. They initiated toilet training, encouraged our daughter and eventually succeeded, well before we did much at all at home. All we were left to do was to set her straight at home which, suffice to say, was not too hard once she had revealed her abilities to the wider community.

Day care is the perfect place for all this. First of all, the providers have as much, if not more, incentive as we do to get children trained. They change more diapers and also have to potentially deal with the children for years to come. They have no desire for a ‘‘slow to train’’ child. Of course, our son had to leave their capable hands before he was done and move to a preschool that required a trained kid. That dampened incentives somewhat. But give day care providers a future with another thousand diaper changes and we have a tight alignment of interests.

Second, and this goes without saying, they have seen it all. They are simply more capable in terms of knowing the signs, assessing readiness, and doing all of the other crap (literally) that first-time parents think about but cannot do.

Finally, the children have peers at day care. Now the power of peer pressure is something that can lead to both good and evil. The evil usually becomes apparent when your child follows others to leap off a several-feet-high structure or starts sucking noodles up their nose. But the good can be equally powerful. If all the other kids are successfully going to the toilet, there is intense pressure to join in, and to do so in a meaningful way. Your child wants to get the same cheers their friends are getting for demonstrating this activity. And they don’t want to be tended to for a soiled diaper.

Even wearing a diaper can be socially difficult. A friend’s three-year-old son, who didn’t attend day care, shed his diaper himself when a random older kid made fun of him in a playground. Of course, that meant no nighttime diaper either and a few difficulties for his parents as a result of that.

But our daughter gave up diapers at day care. Indeed, in the early days, she would convince some of the more part-time workers that she didn’t wear a diaper; although apparently those earlier forays met with unfortunate results. But it continued later on, too. I remember being informed, after having collected her and driven halfway home, ‘‘I don’t need to wear a diaper!’’ Turned out she was diaper free. Being on the freeway, there was not much I could do. So I went with it, and it turned out, thankfully, that she was correct.

I won’t pretend that we were totally free of obligation. For a while, there was a distinct difference between her behavior at home and elsewhere. But once we got on the program and deployed a few incentives, we had complete success in a matter of days.

So the moral of our story is simple. When you come across (virtually) one-time activities that you have no competence to manage, you should outsource them to those who deal with these things regularly and who also have plenty of experience. The end result is mostly the same as if you did it yourself, but with less stress and much cleaner carpets.

Joshua Gans is Professor of Strategic Management and holds the Jeffrey S. Skoll Chair of Technical Innovation and Entrepreneurship at the University of Toronto’s Rotman School of Management. He is the author or co-author of, among other books, “Innovation + Equality,” “The Disruption Dilemma,” “Prediction Machines,” and “Parentonomics,” from which this article is excerpted.