Conall Boyle on university admittance: (3) who gains?

Posted on April 18, 2011 by Yoram Gat

This is the third and last part of this article. The first two parts are 1 and 2.

In a democratic society university admittance policy would be set according to the informed decision of the members of the society – possibly through a representation by an allotted decision-making chamber. The decision makers would have to consider what would be the advantages and disadvantages of possible admittance policies and attempt to design a system that would create maximum benefit for the maximum number of people. (Indeed, in a democratic society, all aspects of university policy, such as the procedure for setting the curriculum, should also be designed so as to maximize the benefit for society as a whole.)

Two effects of the admittance policy that merit consideration are its impact on slot availability and its impact on the ideological stance of the members of the public regarding the benefits of university education. Both of those considerations indicate that a lottery-based admittance policy has clear advantages over the achievement-based policy. While I think that the long term objective for the university system should be to provide quality education to all who seek it, the advantages of the lottery-based admittance system make it both a reasonable system for societies that cannot afford to provide education to all, and make it a good tool for creating a shared interest in reaching this desirable goal.

1. Slot availability

As Conall Boyle emphasizes, the possibility of employing a lottery emerges when a resource is scarce. If the number of applicants to a certain university course is smaller than the available number of slots, then neither a lottery nor any other filtering method is needed. Why, then, are the university slots scarce? Does this scarcity represent the best interests of society? On the face of it, it seems that the natural response to high demand for university slots would be to attempt generate more slots. Would it be difficult to do so?

Oddly, Boyle never focuses on this matter. His most direct reference to this issue is on p. 122:

In the old days, getting onto a university course always required some minimum qualification such as Matriculation. In some countries, most notably Italy, all qualified candidates were entitled to enroll on any course at any university. This led to huge numbers of students, sometimes in their thousands, crowding on to popular courses. Ability to pay fees and support one-self was another means of limiting numbers which applied in my own student days in Ireland in the 1960’s. The ‘affordability’ hurdle was removed for many by the introduction of student grants and subsidised or free tuition. This in turn creates more demand for places. For courses such as medicine and engineering, where laboratory or workshop availability restricts numbers, then some means of rationing the available places is needed.

This seems a very unsatisfactory treatment of a fundamental issue. Expanding facilities is unlikely to be a serious long term obstacle for a modern society, and in any case, there are many theoretical courses for which demand outstrips supply as well (business administration and law, for example). A different explanation would be that there is a shortage of qualified teachers, and that expanding the number of slots would inevitably cause a deterioration in the quality of instruction, resulting in a deterioration in the quality of graduates. Again, this seems like an unlikely long-term issue. Instructors can be trained and there is no reason to think that the pool of all potentially qualifiable instructors has been exhausted.

It seems much more likely that any shortage in the total number of university slots is artificial in the sense that with appropriate investment more slots can be generated, resulting in a a greater number of graduates. The question then is why that investment is not made. Again, a standard answer would be that the investment would be too high, and again, this seems unlikely: the resources needed for reasonably good education are not particularly high ($10,000 a year in a good public university in the U.S., $20,000 at top private U.S. universities).

A possibility that needs to be examined is that the scarcity is maintained since its existence serves some interest. One obvious example of interests being served are those of the select few who do win one of the few university slots. They enjoy social and financial benefits associated with the fact that their status as graduates is relatively rare.

The use of a competitive admittance policy makes it possible for a segment of the population to predict with a certain degree of probability that they will be the beneficiaries of the advantages associated with holding the scarce degrees. Thus, it builds an inherent conflict of interests between that powerful segment of the population and the interests of the majority of the population. Admittance by lottery, on the other hand, makes predicting the beneficiaries impossible, thus dissipating the conflict of interests and making it more likely that the benefits to the public at large would determine public policy.

2. Ideological stance

One of the effects of the admittance policy is its impact on the views of university graduates as well as by other members of the public on the position of graduates in society. A competitive process tends to generate a view in which university slots are seen as a prize for merit. Seen in this way, graduates do not see themselves as having benefited from a deliberate gift provided to them by society, but as a natural elite which is entitled to any benefits associated with its unique status. When this is the case, there is little reason to expect that the benefits to society of the university system would be distributed widely. The graduates would exchange the benefits of their skills for as high rewards as they would manage to obtain – thus extracting for themselves as large a part of the value of the benefits they provide.

Admittance by lottery, on the other hand, makes it clear that the graduates have no right to benefit disproportionately from their diploma. They will naturally be seen by the members of society as merely being lucky and would be expected to serve the community that put them in a privileged position. This view would probably be largely shared by the graduates themselves, leading to a relationship that is not based on a conflict of interests but on cooperation.

Filed under: Books, Distribution by lot, schools |

« Machiavellian Democracy Participatory Budgeting »

27 Responses

keithsutherland, on April 18, 2011 at 8:03 am said:

Yoram: “In a democratic society, all aspects of university policy, such as the procedure for setting the curriculum, should also be designed so as to maximize the benefit for society as a whole.”

That might be true of a social democracy, but is not necessarily the case for a democratic society per se. For example the legislature (allotted or otherwise) might decide that as further education is not compulsory that it is no business of the state to get involved in the private decisions of individual citizens, such as whether to invest in a university degree or a new BMW. If that were the case then the legislature setting the curriculum would be akin to the legislature telling car manufacturers how many cup holders to put in their new models and supermarkets what items they should put on their shelves. The sovereign legislature might even conclude that the collectivist utilitarian assumptions underlying your post are a throwback to failed twentieth-century political dogmas and that it would be better not to repeat past mistakes.

In sum, it is important not to conflate democracy and social democracy.

LikeLike
Harald Korneliussen, on April 18, 2011 at 8:34 am said:

> a system that would create maximum benefit for the maximum number of people.

Pet peeve: This is ill-defined. It can be interpreted as either to maximize the sum of benefits, or to maximize the minimum benefit (the situation of the one left worst off by the action).

LikeLike
Harald Korneliussen, on April 18, 2011 at 9:49 am said:

I share your concern with slot 1. I have long suspected that it’s mostly for prestige reasons there are so few slots for would-be physicians.

Here in Norway, there is a shortage of physicians. At the same time, you need virtually perfect marks in high school to be accepted (grades are supposed to be normally assigned, so that rather than mastering a set curriculum to get an A, you need to be in the top 3% of students. To have a chance at med school, you need to be in the top 3% of all students in about 3/4 of courses, and no worse than the top 7% for the rest.)

Now of course, physician is a profession where you hardly want the candidates to be underqualified. But the question is, should it be that extreme? The main consequence of it is, after all, that more people just study medicine abroad, where demands are lower. I suspect there’s a good deal of status protectionism involved.

LikeLike
peterstone, on April 18, 2011 at 2:24 pm said:

I must agree, at least in part, with Keith here. The question of whether an AC should set educational policies, and the question of what goals it should pursue in setting those policies, are at least partially separable. I say “partially” because surely the same values that lead us to make use of ACs will apply to problems of education. Those values will constrain how the AC may reasonably set educational policy, but surely they don’t dictate it.

I agree that educational admission by lottery may well serve the purposes you describe. But it’s not obvious to me that those purposes are right. Universal university admission might achieve some desirable purposes. But at least one thing it does right now is serve a gatekeeping function. Firms restrict their searches for new hires to university graduates. They do this because they think graduating from a university signals something important about the graduate. (You can dispute this if you like, but I think there’s no denying that most firms do assume this.) If everyone who wanted to attend university could, then the signal would no longer be there, and firms would have to do something else. Perhaps they’d look for a masters or a doctoral degree as a signal instead. There is some evidence of this “arms race” occurring. There are many jobs today which arguably don’t require a high school degree, and yet which high school dropouts would have a hard time getting.

The second claim raises similar issues. If we don’t want people who get admitted to university to think they’re different from everyone else, surely we also have to believe those those that get admitted AREN”T different from everyone else. That requires the really strong assumption that the admissions policies track absolutely nothing. And that’s going to be a very hard sell even among strong egalitarians.

LikeLike
Yoram Gat, on April 18, 2011 at 5:54 pm said:

Keith:

> Yoram: “In a democratic society, all aspects of university policy, such as the procedure for setting the curriculum, should also be designed so as to maximize the benefit for society as a whole.”

> That might be true of a social democracy, but is not necessarily the case for a democratic society per se.

Harald:

>> a system that would create maximum benefit for the maximum number of people.

> Pet peeve: This is ill-defined. It can be interpreted as either to maximize the sum of benefits, or to maximize the minimum benefit (the situation of the one left worst off by the action).

Peter:

> The question of whether an AC should set educational policies, and the question of what goals it should pursue in setting those policies, are at least partially separable.

I grant those objections – I was somewhat careless in my phrasing. On the other hand, I think the distinction being made (between a democratic procedure and maximizing social benefit) is much less clear than might appear. As Harald points out, “maximizing benefits” is inherently ambiguous. The reasonable operative interpretation, I think, is to presume that benefits are being maximized when the policy decision is taken in a democratic way. This is analogous to presuming that a person maximizes the benefits to herself when she makes decisions herself rather than having them dictated by others.

My argument, in any case, is not that a democratic society must arrange its educational system in a particular way, but that democratic decision making would likely promote certain goals. It is conceivable that a democratic decision making group would voluntarily attempt to concentrate all the benefits of education in the hands of a small minority, but I find this unlikely.

LikeLike
keithsutherland, on April 18, 2011 at 6:41 pm said:

“It is conceivable that a democratic decision making group would voluntarily attempt to concentrate all the benefits of education in the hands of a small minority, but I find this unlikely.”

Well, maybe, but the recent history of the expansion of higher education in the UK might cast doubt on that assumption. When I went to university (1969-72) it was around 15% of the age cohort and most people found university students (and hippies) rather odd. The expansion in the following decades was entirely initiated by politicians, who were not in any way responding to a groundswell of demand (this has never been a dominant political issue in the UK). Conservative PM John Major’s push to double graduate numbers (by merging universities and polytechnics) was motivated by his own personal resentment at not having a degree. After that the push for 50% of the age cohort came from Tony Blair’s Labour Party. Having given up on socialism and economic equality they were searching around for other areas to express the party’s commitment to egalitarianism and education was one of them. Most commentators would argue, with the benefit of hindsight, that the rapid expansion of higher education was a total disaster, on account of the arms race that Peter referred to, the need to generate public sector (non-) jobs for all the graduates and the resultant need to important large number of immigrants to do the real jobs that the new breed of graduates were too “educated” to do.

LikeLike
Yoram Gat, on April 18, 2011 at 9:29 pm said:

Harald,

> I share your concern with slot 1. I have long suspected that it’s mostly for prestige reasons there are so few slots for would-be physicians.

Have a look at the situation in the U.S.

> Now of course, physician is a profession where you hardly want the candidates to be underqualified.

I don’t think there is any reason to believe that raising the admittance score threshold does anything to produce better doctors. Certainly a shortage of doctors does public health significant harm.

LikeLike
Yoram Gat, on April 18, 2011 at 9:38 pm said:

Peter,

> If everyone who wanted to attend university could, then the signal would no longer be there, and firms would have to do something else.

I proposed education on demand, not graduation on demand. Graduation should, of course, involve being qualified.

That addresses your second point as well – being officially certified as having a certain level of education should require measuring qualifications. The measure, however, should be absolute rather than relative. For me to be certified as literate does not require me getting a higher literacy score than you.

LikeLike
Yoram Gat, on April 18, 2011 at 9:51 pm said:

Keith,

> Well, maybe, but the recent history of the expansion of higher education in the UK might cast doubt on that assumption.

I am afraid that based on experience I am unwilling to take for granted your description of historical events. As usual, your historical account is full of unsubstantiated unlikely claims, and experience has shown they often simply untrue.

Be that as it may, however, let me point out the obvious: sharing the benefits of education does not necessarily mean that everyone should be a university graduate. A person can benefit from the existence of doctors, say, without being one themselves. However, if the doctors’ fees put their services beyond the reach of the average person, then the benefits of the doctors’ education flow to the rich only.

LikeLike
keithsutherland, on April 19, 2011 at 7:11 am said:

“your historical account is full of unsubstantiated unlikely claims”

By all means research it for yourself, along with my other “unlikely” claims:

1) “Poor” people in the UK have been as likely to fly abroad for foreign holidays as more affluent ones, who are as likely to go camping in a tent. This is one of the reasons for the very high level of long-term credit card debt.

2) The reason Blair chose not to tax aviation fuel had nothing to do with pressure from the airlines and everything to do with (1) above.

When it comes to democracy, it’s extremely unwise to second-guess the outcome. I’d even go so far as to claim that there is nonexistent popular support for the abstract substantive that is the prime title of this blog and will gladly unpack that argument if you wish.

LikeLike
Harald Korneliussen, on April 19, 2011 at 7:40 am said:

> I don’t think there is any reason to believe that raising the admittance score threshold does anything to produce better doctors.

Really? Because the converse for doctors in Norway would be teachers. Really few people want to become teachers, it’s gone from a comparatively high-status profession to a low-status one. In particular elementary and middle school teachers.

Now a consequence of that (and grade tresholds being set by supply and demand), a lot of elementary and middle school teachers had the absolute minimum math scores – barely above failing. I’m pretty sure that did in fact lead to worse teachers.

So in principle, I think grade tresholds can have an effect on skills. However, whether there’s any meaningful difference between a doctor with a grade average of 5.0 and one with 5.7 is another question – I think that is so small as to be neglible (there is a lot of randomness involved in grading).

LikeLike
Conall, on April 19, 2011 at 11:07 am said:

Harald,

I am disappointed to learn that Norway no longer uses lottery to ration or re-allocate slots on medical or other high demand courses. (see p146 in my book for more on this).

Do you know why or when the lottery was dropped?

(and thanks for pointing out the absurdity of using higher and higher grades as a rationing device)

LikeLike
Conall, on April 19, 2011 at 11:41 am said:

Yoram, hold your horses. In line with my preferred idea of ‘incremental improvement’ (see Deming for more on this) I took the current uni admission system as it exists, and looked for simple ‘patches’ which would make it better. Adding a lottery, Dutch-style, would do that.

It is grand to think how a truly democratic uni system would operate its admissions. We could go deep into the medical profession’s crafty ‘market-capture’ by limiting numbers, raising prestige etc. Milton Friedman was good on this!

But to me the biggest problem of all is not the twisted admission system — it’s what happens to these graduates enter the employment market. This is when your degree becomes a ‘positional good’ par excellence. If it says Harvard or Oxford you glide in to the best jobs. There is plenty of UK evidence that the elite is heavily drawn from Oxbridge.

So tinkering with uni admissions processes by using a lottery will improve the fairness and reward for valid merit a bit. But the Big One is the employment market – hiring, firing and promoting. My next book was going to be about using the lottery in these processes, but perhaps others might like to pick up the baton. (Good work is coming from the Catania, Sicily boys on this)

LikeLike
Harald Korneliussen, on April 19, 2011 at 1:44 pm said:

Conall,

There is a period at the end of medical studies, I don’t know the English term for it, where doctors are bascially “interns”. These openings are drawn by lot: If you get an internship number of 1, you get to choose first between the 400 or so openings. If you get number 400, expect that the only internship left will be in inner Troms or something (very cold, pitch black half of the year, virtually no people). You can also get an internship number higher than the number of available slots, in which case you have to wait half a year to try again.

As far as I know, entrance to the study itself has always been based on grade point averages, but the internship system is still in place, and lottery-based.

LikeLike
Yoram Gat, on April 19, 2011 at 7:01 pm said:

> Now a consequence of that (and grade thresholds being set by supply and demand), a lot of elementary and middle school teachers had the absolute minimum math scores – barely above failing. I’m pretty sure that did in fact lead to worse teachers.

First, I would not take for granted that you have worse teachers now than in the past. It may very well be that it is the lower status of the teachers that allows more criticism and complaints.

Second, again, I am not claiming that no qualifications should be needed in order to obtain a certain certification. I am claiming that the standard should be absolute rather than competitive. Of course, if you raised the score threshold for teachers, you may find that you need to pay them more in order to have enough of them. This, it seems, may be the real source of the problem you are touching on: an attempt to save money by underpaying the teachers.

Third, the Dutch data that Conall presents show that even if we measure the quality of the candidate by the peculiar standard of chance of graduation and speed of graduation, the association levels off at a certain point and may in fact be trending down at the top of the distribution.

Finally, as I argued in the post, even if for the sake of the argument we do assume that higher scoring candidates do make somewhat better graduates in a meaningful sense, a score-based admittance policy has negative effects (artificial scarcity and cornering of the benefits of education) that, in my mind, overwhelm any advantages that may be gained.

LikeLike
peterstone, on April 20, 2011 at 2:27 pm said:

How interesting. I guess there are two important questions here. First, to what extent to higher qualifications from candidates (however we define them) signal improved ability to perform? Second, to what extent can a candidate improve her ability to perform by obtaining higher qualifications? If the latter answer is yes, then the former should be as well, but not the other way around. (It could be that the qualifications signal something about the candidate that the candidate cannot change no matter what she does.) If both answers were yes, then a flat bar of qualifications (making it an admission, not a selection, procedure) might lead to lower overall levels of ability, as candidates did just the minimum to qualify. It’s at least a possibility one would have to consider whenever one decreases the competitiveness of a procedure.

LikeLike
Yoram Gat, on April 20, 2011 at 9:06 pm said:

> I took the current uni admission system as it exists, and looked for simple ‘patches’ which would make it better. Adding a lottery, Dutch-style, would do that.

As you see, I support this idea (although the details may need to be looked at). My point from the beginning was more about the argumentation than about your conclusion.

> But the Big One is the employment market – hiring, firing and promoting.

Like in education, I think that instead of trying to make distribution of a scarce resource fair, it is better to eliminate the scarcity. Elimination of the great disparities between the “good” jobs and the “bad” jobs is more realistic and useful than using lotteries for employment decisions.

LikeLike
Yoram Gat, on April 21, 2011 at 12:10 am said:

> If both answers were yes, then a flat bar of qualifications (making it an admission, not a selection, procedure) might lead to lower overall levels of ability, as candidates did just the minimum to qualify.

Assuming that the qualifications are relevant to the university material, starting out with lower qualifications will simply make for more difficult time at the university so what will the candidate gain by slacking?

In general, I think experience teaches that this kind of “competition encourages excellence” force is much weaker than the “competition encourages anti-social behavior” force.

LikeLike
Mark, on April 22, 2011 at 7:18 pm said:

This article makes the assumption that graduation rates matter most automatically translates into improved income as if the ability to pass standardised tests is the sole criteria of success in the modern world.

Most elite univeristiy institutions look at academic qualitifications and extra-curicular qualifications together. The first cut is based on academic grades, i.e. is this person smart enough and disciplined enough to actually graduate from the course? The second cut is based on leadership potential? Was he class president, did he play for his high school football team and was he captain? etc.

Most elite univesities are in the buisness of manufacturing future leaders and testing leadership potential is a lot more subjective than testing book smarts. That said elite universities tend to do well in separating the well adjusted from the autistic.

The Ivy league schools certainly have a selfish interest is selecting the best as these Universities survive on massive endowments accumulated through the years by their alumni. Had they left part of their selection to blind chance it would not only reduce the value of the degrees from these Universities but would also have a negative knock on effect in terms of a reduced endowment which would mean less money to spend on retaining the best professors and faculty staff.

This of course assumes that we are more interested in promoting excellence in education rather than equalitarian social engineering.

LikeLike
keithsutherland, on April 22, 2011 at 8:00 pm said:

“This of course assumes that we are more interested in promoting excellence in education rather than equalitarian social engineering.”

Agreed! (although I think that we are in a small minority on this site).

LikeLike
Yoram Gat, on April 23, 2011 at 2:29 am said:

> This article makes the assumption that graduation rates matter most automatically translates into improved income as if the ability to pass standardised tests is the sole criteria of success in the modern world.

That is certainly not an assumption that I make. My assumption is the one made by Conall – that university admittance policy (like any public policy) should be designed so as to maximize the benefit for society.

> This of course assumes that we are more interested in promoting excellence in education rather than equalitarian social engineering.

This is simply standard issue sloganeering. What do you mean by “excellence is education”, and why would society be interested in promoting it? What do you mean by “social engineering”? How is it different from other types of public policy making?

LikeLike
Mark, on April 23, 2011 at 4:08 pm said:

Yoram I don’t agree with the idea that selection for University should be even in part subject to the laws of blind chance. I mentioned that private institutions have a pretty good track record of picking out the wheat from the chaff after all they depend upon their alumni to fund their Universities.

By social engineering I mean the idea that society can be made more equal by everyone having a college degree.

Yoram I would think that wanting an excellent education system is pretty self-evident as it adds value.

LikeLike
Yoram Gat, on April 23, 2011 at 6:46 pm said:

> I don’t agree with the idea that selection for University should be even in part subject to the laws of blind chance.

Why? Please engage with the arguments made in the article instead of making unsubstantiated statements.

> I mentioned that private institutions have a pretty good track record of picking out the wheat from the chaff after all they depend upon their alumni to fund their Universities.

First, I find your use of the word “chaff” to describe a set of human beings offensive, if all too typical of dominant dogmatic thought in Western society. Second, I don’t doubt that the admission policies of private universities are designed to benefit the university officers and, indirectly, their customers – i.e., the students in those universities. The point is that they don’t serve most people in our society.

> By social engineering I mean the idea that society can be made more equal by everyone having a college degree.

This would be a very odd definition of “social engineering”. I assume you mean this to be an example of social engineering rather than a definition.

But that doesn’t answer my question – what makes the policy you suggest (which, by the way, is not a policy I suggest) different from other policies? Why can this policy be called “social engineering” while other policies (such as the one that exists, or the one you support) are not “social engineering”?

> I would think that wanting an excellent education system is pretty self-evident as it adds value.

I don’t doubt that having an excellent education is very good for the educated person. (We may find defining “excellent education” to be a non-trivial task, but let’s leave that for now.)

The question, however, is different. The question is how to design the education system so that it provides “value” to as many members of society, rather than only to a lucky few.

LikeLike
peterstone, on April 24, 2011 at 1:03 pm said:

Mark (I don’t believe we’ve met), I would agree with you that it’s pretty self-evident that a quality education adds value to someone’s life, however that’s interpreted. (It could be in terms of the market value of that person’s abilities, but that’s not the only way to define it.) But it’s an additional claim to say that society should give its best education to those who already possess the most ability. One could, with perfect coherence, argue that the people with the most ability will do fine wherever they go, and so society should give the best education to those who need the most help. Or one could argue that the most value added will occur if the best education goes to those in the middle of ability–those at the top will do fine anywhere they go, and those at the bottom might not be able to take full advantage of the best opportunities. I’m not defending any one of these. I’m just saying that what you’re calling “self-evident” does require a careful argument.

As for the ability of top universities to select the best students (I’ve taught at two of them), I recommend to you an article by Barry Schwartz from the CHRONICLE OF HIGHER EDUCATION–

Click to access Chronicle%20of%20Higher%20Education%202-25-05.pdf

Schwartz says, you say your criteria are really good? Prove it. Let’s admit a random sample of applicants to make up a small part of the class, and treat them as a control group. If your admission criteria really work, then the students admitted using those criteria will do better than the random sample. It’s an obvious way to evaluate whether admissions officers know what the heck they are talking about.

Interestingly, Schwartz apparently has written on the subject since then–

http://www.huffingtonpost.com/barry-schwartz/why-selective-colleges–a_b_177909.html

LikeLike
Conall, on April 25, 2011 at 10:49 am said:

Thanks, Peter, for bringing this discussion back to the factual reasons for choosing by lot. Schwartz chaired an Enquiry into Uni admissions, and made the same point, most forcibly to a UK Parliamentary Committee – people, even very clever academics have poor skills in judging ‘charcter’. If for no other reason this is why we need a lottery over the adequately qualified applicants. A lottery sanitizes the decision process by eliminating the weakest link — frail and inevitably biased human judgement.

Sanitise uni entry of its bad, biased human intervention. Use a lottery!

LikeLike
Yoram Gat, on May 5, 2011 at 4:32 am said:

Conall,

I just came across, quite by chance, something that you may find interesting: Democracy By Lot — A College Experiment.

LikeLike
Irish higher education minister laments the cruelty of random selection | Equality by lot, on September 1, 2023 at 10:46 am said:

[…] more on the convoluted elitist logic behind such statements, see my three part review of Connal Boyle’s book Lotteries for […]

LikeLike