Bryan Caplan  

Separating Twins as Economic Illiteracy

Subsidizing Mortgages... Holiday Reading...

Schools usually try to put twins in different classes.  In part, it's for the convenience of the teacher - identical twins can be hard to tell apart.  But the main rationale is that if you separate twins, they will make more new friends.  Isn't that great?

If you've had a week of intro economics, I hope your answer is "No."  Consider: What would you do if aliens abducted all of your friends?  In all likelihood, you'd wind up making a bunch of new friends, right?  But it would be absurd to claim that the aliens had done you a favor.  Before the body snatchers came along, you'd didn't need those new friends, because you were happy with the ones you already had.  To argue otherwise is just make-work bias.

The same logic applies to splitting up twins.  Sure, if you separate twins, they'll make more friends.  But that hardly means you're doing them a favor.  The reason why twins put less effort into making new friends is that they've already got a better friend than most of us will ever have.  For twins, the marginal benefit of trying to making new friends unusually small - and cliquishness is their optimal response.

You could argue, admittedly, that kids underestimate the benefit of making new friends.  But what reason is there is to believe that this is true?  And in any case, that's no reason to specifically target twins.  If you really believed that kids had an "anti-new-friends bias", you'd ask all parents to name their kids' best friends, and make a effort to separate as many children as possible from the kids they like the most.  "You'll thank me later," right?

A decade or two ago, pundits often said that the Japanese economy was doing well because all their factories were destroyed during World War II.  As a result, the story went, the Japanese built shiny new factories from scratch, and the victors with their musty old plants just couldn't compete.  The standard economists' response to this nonsense: "If that's true, all we need to do to catch up to the Japanese is just bomb our own factories!"

Folks who want to separate twins are making the same mistake.  We always have the option of "destroying in order to rebuild."  But it's usually a bad idea - especially if you're destroying over the objection of the people you're trying to help.

So what should we do with twins?  Simple: Ask them what they want to do.  If they want to stay together, keep them together.  If they want to separate and develop their individuality, split them up.  And if you're really sure that "Stay together" is the wrong answer, why not try to convince us with a little self-experimentation?  Cut the time you spend with your best friend in half, and tell us how you like it... 

Comments and Sharing

COMMENTS (25 to date)
Caliban Darklock writes:

Considering the incentives, how many people would join me in the hypothesis that this is complete garbage, and the real motivation is to prevent the teacher from getting them mixed up?

After all, the primary goal of everyone in any profession is "don't look stupid" and mixing up your students would impede that goal.

Thomas DeMeo writes:

The only place where make-work bias really makes good sense is in the context of a school or learning environment. I'm not sure you could have picked a poorer example to make this point.

David R. Henderson writes:

"So what should we do with twins? Simple: Ask them what they want to do. If they want to stay together, keep them together. If they want to separate and develop their individuality, split them up."

Bryan, shame on you. Next you'll be saying that people should be able to buy the kind of car they want or marry the person they want or save for their own retirement. How incredibly anti-social!

Sravana Reddy writes:

"If you really believed that kids had an "anti-new-friends bias", you'd ask all parents to name their kids' best friends, and make a effort to separate as many children as possible from the kids they like the most."

Actually, that's what used to happen in my school. (They didn't have to ask the parents, they just surmised who the friend-cliques were.) So we all ended up knowing lots of people and making fewer friendships.

bjk writes:

Kids also love ice cream. Only proto-Stalinists insist on feeding vegetables -- just ask the kids, they'll tell you!

Prakhar Goel writes:


(Most) kids also have a recorded bias for ice cream, a case specifically mentioned by the post. At least read the entry before commenting.

bjk writes:

Not only do children have a "recorded bias" for ice cream, Prakhar Goel, they also like it! But no sign of this in the post . . .

dearieme writes:

I went to schools that "streamed" children according to ability, so twins tended to end up in the same class.

Les writes:

I agree that "making new friends" does not justify separating the twins.

But it seems to me that the teacher's difficulty in telling them apart is ample justification for separating the twins into different classes.

Robin Hanson writes:

An excellent post!

Grant writes:


I think separating twins in schools is a special case of a wider-ranging problem in schooling. The lack of freedom of association given to children forces them into a social environment very different from the rest of the world. Kids can't choose who their parents are, we could at least let them choose who to associate with at school (or which school to associate with).

Unit writes:

Maybe the teacher is worried that the twins will work on homework and other projects together making it difficult to assess who's good at what. (not that I agree with it).

Tom writes:

I'm a fraternal twin. We were as different as night and day - no danger of the teacher confusing us.

I definitely would have preferred the administration to ask us what we wanted to do, though - I was quiet, and my brother was a clown. It would have given me the opportunity to get one more distraction out of my class.

Loved this article, Bryan.

reason writes:

I thought the reason for seperating twins was to let them develop their own sense of identity. Why do think the reason is the one you gave - it doesn't make much sense to me - you could use the same argument to justify randomly mixing the classes every year?

reason writes:

I just thought - maybe there is a hidden agenda! Having identical twins in seperate classes is a great way to test the ability of teachers (hence the non-sensical official reason)!

guthrie writes:

I went to Jr. High with a set of maternal triplets who were separated into different classes... and they would switch with each other all the time! They'd get each other in trouble, take each other's tests, and mess with the other kids in the classes, all for a laugh. Only a handful of us could tell them apart, and we laughed with them. I suppose they made the friends they wanted and didn't care about the rest.

I think the whole 'socialization' argument is bogus. I loved your suggestion, Bryan, b/c it's very respectful of the kids themselves (shocker!). The real reason to separate twins, etc., is to keep from confusing lazy teachers who won't pay enough attention to these kids to tell them apart.

Boonton writes:

Hearing guthrie's story I'm reminded of the Freakonomics piece on Japanese Sumo wrestlers. If you didn't read it, he found that Sumo wrestlers were probably throwing fights....but not to get money but to 'spread the success around' so no one would get kicked out for loosing too many matches.

Imagine you have twins and one excels and can easily score an A while the other struggles. Might the twins swap places now and then so they can both average a B?

To me this seems like an interesting question of marginal utility. The smart twin may find the marginal benefit of going from a B to an A is very might even be negative. If it means that his or her sibling will get a lot more attention from parents because they are a C or D it might make more sense to give up the A in order to help the other twin out a bit.

Different classes would make this game harder to pull off. Dismissing this as "convenience of the teacher", I think, sells the policy short. It might be better in the long run for the twins since the incentives of having them together in a class may end up masking a problem.

guthrie writes:

Actually, Boonton, aside from the humor they got from it, that's exactly what they would do. One guy was better at math, so he would go to all the math classes (or as many as possible), one guy was better at Social Studies, etc. That's why they would act up. If you already know the material, there's no need to pay attention. And, if one teacher were teaching two different courses, or have two of the brothers at different times, they would sometimes switch clothing in the bathroom to keep the charade up.

I think at one point the teachers' caught wind of the scheme and tried to stop it, but I don't think they were too successful... the guys just figured out a way around it.

If they were all in the same classes, there would be no question that they were all getting the same information and being tested on the same material. My guess is that they would all opt to be together, if given the choice.

And I suppose my dismissiveness stems from my disdain for the teachers that I saw who didn't take the time to simply pay attention. My own bias I guess. I had respect for those who, I felt, would actually make an effort to understand us.

Boonton writes:

Well now that the subjects been mentioned how much do twins and triplets 'swap lives' after childhood? College? Work? Marriages? How about when one older twin dies, might the other one keep him alive on paper to collect his social security? Has anyone ever looked at that?

Stephen Jones writes:

I've taught identical twins. You don't mix them up after the first couple of days, any more than their parents do.

Snark writes:

Logic suggests that separating twins at school enhances their individual development. However, research and family experience supporting the opposite view is equally convincing. Accordingly, many behavioral experts believe that adopting a flexible policy involving the decision to separate twins at school on a case-by-case basis is optimal.

But consider that, in addition to the benefit of making new friends, identical twins could parlay separation at school into a situation of comparative advantage. For example, they could substitute for one another in classes where each excels academically, resulting in a Pareto improvement in their overall GPA.

Identical twins have successfully perpetrated this type of fraud on their dates. Why not on teachers?

Jonathan WR writes:

My fraternal boy-girl twins were together throughout preschool, and are now in separate classes in kindergarten. Apparently by 1st grade the school "requires" them to be separated.

They are thriving and enjoying being in separate classes -- it gives them each their own stories to tell us, and each other, after the school day. Also they are quite different and get along very well.

Also my daughter is exceptionally good at reading (my son is above average, but far below his sister), and I think my son enjoys not having that comparison always hanging over him.

So far I think it makes sense to have them separate, on the theory of developing own identity. But it does annoy me that the school seems to have such a strong policy about it.

Tom West writes:

For twins, the marginal benefit of trying to making new friends unusually small - and cliquishness is their optimal response.

It depends upon whether you think making new friends is a skill that is most successfully instilled early. If you believe that, then you might well decide that it is in the best interest of the children to split them up, since together there wouldn't be the motivation necessary to learn the skill, and failing to learn the skill might cause unhappiness for the rest of their life.

If we were going to poll the twins, I'd tend to poll them at age 25, when they might have a better idea about the long term consequences (both beneficial and negative) of being split.

Zac writes:

If you think the purpose of school is purely education, then maybe Bryan is right here. If you think part of function of school is "socialization," then separating the twins has some obvious utility, i.e. you are forcing them to interact with other people. The ability to interact with other people is a valuable skill and part of the reason why most people don't homeschool their children. I don't think the fact that they're twins really makes a difference - you can say the same for any siblings.

Even if Bryan is right and the "socialization" aspect means little and the children will gain nothing from being split, which is feasible to me (I'm no expert on child psychology, but I figure in some cases its good and in some not so much), the "ask the children what they want" argument is not convincing in the least. Actually it is kind of ridiculous. Ask the kids if they want to wake up in the morning, go to bed at night, eat vegetables, brush their teeth, go to school.. I'm as libertarian as they come, but when dealing with young children I think there is a convincing case for paternalism.

Nichlemn writes:

Perhaps a reason to target twins and not the "anti-new friends bias" in general is the cost of uncovering information. Suppose some research has concluded that children deviate strongly from their optimal behavior with regards to socialisation, but this effect is especially pronounced for twins. On the other hand, the gains from separation tend to be modest. A cost-benefit analysis might not justify trying to separate *every* group of friends, but when they are particularly easy to identify, such is in the case of twins, it may be the case.

Comments for this entry have been closed
Return to top