Why women fail to take an equal share of top posts in academia

Here’s a nice piece in today’s Irish Times, showing clearly the need for lottery selection in jobs.

It asks the question: Why women (despite being over 50% of the faculty staff) fail to take an equal share of top posts in academia?

You can read the article in full (and for free!) at

http://www.irishtimes.com/opinion/why-women-fail-to-take-an-equal-share-of-top-posts-in-academia-1.2108539.

In the comments section we the usual reaction to any suggestion of ‘affirmative action’

“Here we go, more feminist claptrap. All academic posts must be based on merit.”

To which a wise commentator replies: “You think merit has much to do with academic appointments? There’s usually an assessor’s box called suitability which you can fill with subliminal prejudices: plays golf, politics, sounds like me etc”.

While waiting for grand schemes of sortitionist democracy to be implemented, wouldn’t it be nice to have a bit of genuine equality based on lottery selection over the qualified candidates?

Advertisements

38 Responses

  1. The extent to which the meritocratic dogma is entrenched in the public’s mind is remarkable. With modernity all the traditional justifications for inequality have become obsolete and meritocracy now has to carry the burden of inequality on its own – electoral politics, economic disparity, workplace hierarchy etc. are all justified as being meritocratic.

    Like

  2. > equality based on lottery selection over the qualified candidates?

    But wouldn’t the bias just manifest itself by affecting the pool of “qualified candidates”?

    Like

  3. The qualifications should be relevant, objective and stated in advance. Any bias can then be contested.

    Like

  4. I think in most situation is impossible to specify relevant qualifications that can be evaluated objectively. One might as well claim that it is possible to find a relevant and objectively evaluated criterion according to which the single best candidate can be chosen making the lottery unnecessary.

    Like

  5. Yoram,

    >I think in most situations is impossible to specify relevant qualifications that can be evaluated objectively.

    So should the lottery pool for the role of (say) head of department include:

    1. All tenured faculty?
    2. All permanent faculty?
    3. All staff members — including graduate teaching assistants?
    4. All members of the university?
    5. All citizens?

    And what about (seemingly) relevant qualifications such as PhD, length of service, publication history, peer review? Should these all be excluded as there is no way of evaluating them objectively? And would your dismissal of “meritocratic dogma” apply universally to all appointments that require particular skills?

    Like

  6. I’d love to know what you guys think about Meritocracy:
    Should the job go to the person who is an all-round good egg; is that ‘merit’.
    Or is it as in Young’s 1958 book, the one with the highest IQ?

    Like

  7. My point is that having a lottery of “qualified candidates” is not clearly a situation with less potential bias than a system that claims to select the single best candidate. It seems that fixing a situation of “bias” (non-merit based selection) will have to rely on a different approach.

    I recall that when I reviewed Conall’s book about lotteries in admissions I concluded that the solution to the problem of admissions is to eliminate the scarcity of slots in educational institutes, rather than appeal to any method of filtering.

    Regarding what constitutes merit: that would depend highly on the context, I would presume. In any case, whatever my personal opinion regarding merit is, in a democratic system it is the informed, considered opinion of the population that should matter.

    Like

  8. Yoram

    >Regarding what constitutes merit: that would depend highly on the context . . . in a democratic system it is the informed, considered opinion of the population that should matter.

    If the context is choosing the head of a university department, who constitutes the “population” — the members of that department or all citizens? If the former, then that approximates to the principle of peer review (as the population in the context of this particular decision is the other members of the department).

    Like

  9. If we want to arrive at an informed and considered decision then before discussing what the proper procedure for selecting a head of a university is, we need to decide what the functions of such a position would be. Does it need to exist at all? Should its functions be carried out by a committee rather than by a single person? In other words, rather than trying to find a quick fix for the oligarchical nature of a particular nominating procedure, we need to re-assess the entire hierarchical structure.

    But, again, in a democratic society, no matter what conclusions we personally arrive at, eventually the decision regarding the proper structure and procedures have to be arrived at by the representatives of the entire population.

    Like

  10. Yoram,

    >eventually the decision regarding the proper structure and procedures have to be arrived at by the representatives of the entire population.

    I believe that the legal status of most universities is that of a private corporation, and the traditional structure of governance is collegiate. Are you suggesting that decisions regarding the governance of such bodies (including the tiny number that receive nothing from the public purse) should be the will* of a statistical sample of all citizens? And would this argument apply to all private corporations and civil society groups? And why stop there, why not include the regulation of family life, as the small size of individual families will make them unrepresentative, so individual families are in clear breach of your democratic principle?

    * will (rather than judgment) is the appropriate term for a full-mandate body selected by sortition.

    Like

  11. The really cool effect of using a lottery is to ‘sanitize’ it from the effects of human judgement, which in the case of personnel selection has been shown to be woefully inadequate! Khaneman on ‘Thinking’ is very good on this.

    You guys seem to want to re-introduce endless committees of people reflecting, discussing, evaluating, deciding criteria, procedures, to pass judgments, to hold interviews etc. etc.

    No, no! Stick to what is known, preferably validated by some empirical research. Some nice work by Dawson P M & Dobson S (2002)Managerial Efficiency and Human Capital: An Application to English Association Football, Managerial and Decision Economics (2002), 23, 471-486 .

    60% of managerial talent is how much money at his disposal. Previous experience helps, but that’s it. BTW there are no super-managers; even the highly praised Sir Alec Furguson was ‘average’. Similar research is available for US football, basketball etc.

    Now was that difficult? A very small number of objective criteria will identify the ‘relevantly qualified’ applicants. Human judgment only adds noise (or more likely bias). Hence a lottery.

    Like

  12. > ‘sanitize’ it from the effects of human judgement

    My point is that it is not clear that it is any easier to sanitize the selection of the pool of qualified candidates than to sanitize the selection of the supposedly best candidate.

    I do agree that much of supposed merit is nothing but bias. However, much of the same can be said about qualifications.

    Like

  13. It seems likely that where there is any discrimination going on, it’s mostly in the subjective part of the evaluation.

    Sure, it’s possible that the easily measurable criteria themselves are tilted towards some groups, e.g. by valuing CV-items one group is more likely to have, without there being a good reason for it. But this, at least, is out in the open. It can be criticized, it can be challenged.

    For the less measurable, more subjective criteria, that’s harder to do. Sure, you can point to the outcomes: “You’re hiring less women than you should if you only look at easily measurable characteristics!” But this isn’t really a very good argument. After all, you’ve in effect already conceded that the more subjective qualities matter, and who’s to say those are evenly distributed?

    When you’re hiring, you are discriminating. You are supposed to be discriminating. Lottery can help you make sure you’re only discriminating on the characteristics you explicitly want to.

    But sure, if you want to discriminate based on characteristics such as race and sex, you’ll find a way. It’ll just be a little harder, and a little harder to conceal.

    Like

  14. I do think we need specific terms for the very different uses of the lottery:

    1. Dowlen defines sortition as the random selection of citizens for office. His focus (along with Boyle and Stone) is on impartiality (sanitizing effect) but agrees with Socrates and Vintermann that you wouldn’t want to draw up a short-list for pilots, flute-players or senior academics by bean: “When you’re hiring, you are discriminating. You are supposed to be discriminating.”

    2. In my view the selection of a statistically-representative microcosm should be called stochastion.

    3. Perhaps advocates of the total social lottery (Yoram, and to a lesser extent Campbell and Terry) should take a cue from Barbara Goodwin and refer to their procedure as aleation. As this blog is called Equality by Lot, this has to be the home for this sort of proposal, but Goodwin is adamant that it is an entirely utopian project.

    If we have distinct terms then it makes it easier not to talk past each other. Burnheim’s project relies on sortition but assumes that the selection from the restricted lottery pool (if large enough) will instantiate stochastion. However he completely rejects aleation. Proposals for citizen juries rely primarily on stochastion and (to a degree) sortition and also reject aleation. Boyle and Vintermann rely on sortition, but reject stochastion and aleation (at least for the executive appointment process).

    Like

  15. Harald (vinterman),

    > Lottery can help you make sure you’re only discriminating on the characteristics you explicitly want to.

    How? Why would the qualification criteria for the lottery pool be more easily measurable than the selection criteria for the “best” candidate?

    Like

  16. “le mieux est l’ennemi du bien.” as Voltaire put it.

    Seeking the ‘best’ leads to consideration of secondary and irrelevant characteristics. A lottery prevents the hunt for such false merit.

    Like

  17. Fine, that means that a lottery would could eliminate irrelevant criteria, but that doesn’t mean the relevant criteria would be less open for bias than the irrelevant ones.

    It may be that the irrelevant criteria are the ones that are easy to quantify and measure objectively while the relevant ones are those that are more subjective.

    Like

  18. *** Conallboyle says (23 February) : « Seeking the ‘best’ leads to consideration of secondary and irrelevant characteristics. A lottery prevents the hunt for such false merit. » Right, and these secondary and irrelevant characteristics, being quite arbitrary, are more open to bias.
    *** Yoram Gat answers : « It may be that the irrelevant criteria are the ones that are easy to quantify and measure objectively while the relevant ones are those that are more subjective. » I agree, in some cases it may occur. In these cases the choice will be partly subjective – it will be an election. Therefore sometimes the best procedure will be objective selection x lottery, sometimes it will be objective selection x election. The choice of the procedure belongs to an upper authority – the choice of procedures for public universities will belong in a dêmokratia to the sovereign populus through a mini-populus.
    *** Note that sometimes the subjective parameter is an objective criterion, for instance the trust – subjective feeling, but objectively necessary for a military chief, and maybe some other chiefs.

    Like

  19. *** Yoram Gat writes (February 20, 2015) : « The extent to which the meritocratic dogma is entrenched in the public’s mind is remarkable. With modernity all the traditional justifications for inequality have become obsolete and meritocracy now has to carry the burden of inequality on its own – electoral politics, economic disparity, workplace hierarchy etc. are all justified as being meritocratic. ».
    *** Yoram’s sentence has much truth, and we can remind the French « Declaration of the Rights of Man and of the Citizen » (1789), article 6, even it was only about public positions « All citizens, being equal in the eyes of the law, are equally eligible to all dignities and to all public positions and occupations, according to their abilities, and without distinction except that of their virtues and talents. »
    *** The principle of meritocracy did exist in traditional societies, but with strong competitors. Homer tells us the anger of Achilles, « the best of Achaean warriors » (meritocratic idea) against the king Agamemnon, who lacks any specific talent or virtue, but is as heir the king of Mycenae.
    *** We must remind that the modern decline of « the traditional justifications for inequality » did not only stimulated the ideology of meritocracy, it stimulated likewise the ideologies partitioning the human kind along race, sex, class origin ….
    *** And we must consider which is becoming maybe the strongest kind of legitimizing ideology : the systemic ones.
    *** In many modern societies a good part of economic inequality is linked to inheritance, and inheritance is not meritocratic. For anecdotal reasons in France we hear much today about Mrs. Bettencourt, the « richest person in France » – and she is an heiress. How is justified her incredible wealth? « The best economic system is the capitalist one. We must let it function freely. With too high inheritance taxes the big capitalists will think too much about how passing down their wealth, instead of about increasing it as standard capitalists actors must do. » The wealth of Mrs. Bettencourt is justified not by any singular quality of hers but as a necessary part of a system which is «the best of all possible systems » (Inequality in the society is justified as was evil in the world along Leibnitz metaphysics.)
    *** About the political system, right, some arguments about « virtues and talents » are used against dêmokratia, but only in a negative way : too many common citizens are dull (« moroons »), too many have bad drives, the politicians and high judges who are apparently in the top of the polyarchic systems are generally better. But it is not a thorough meritocratic legitimacy. Nobody argues seriously that polyarchies are aristocracies of talent and virtue ! The powers of our politicians and high judges are legitimized by the political system, as this system is « the best of all possible political systems ».
    *** The ancient dêmokratia did not imply thorough rejection of the meritocratic principle. It was dismissed, specifically, about sovereignty : because the relevant « virtues and talents » are common sense and common decency (sôphrosunê and dikaiosunê, Plato, Protagoras, 323a, in the mouth of Protagoras, democrat theorist), and these are spread over the whole people, as is said in a famous myth of Protagoras (and even if we may think the dividing is not perfectly equal, any attempt of selection will be quite arbitrary).
    *** The Second Athenian Democracy did elect generals and some treasurers with reference to the meritocratic principle (talents and virtues). But it was not the one organizing principle. A modern dêmokratia, by definition, will not use this principle for sovereignty, but it can use it partly in other fields.
    *** I object to the word « dogma » in Yoram’s sentence, given the weight of systemic legitimacies in our world. Supporters of modern dêmokratia may defend this model and give it a systemic legitimization.

    Like

  20. Andre,

    >The Second Athenian Democracy did elect generals and some treasurers with reference to the meritocratic principle (talents and virtues). But it was not the one organizing principle. A modern dêmokratia, by definition, will not use this principle for sovereignty, but it can use it partly in other fields.

    The second demokratia reserved direct sovereignty for all citizens in the assembly, and to a representative sample in the legislative courts. Everything else (apart from minor magistracies)* was meritocratic, along the lines specified by Pericles in his funeral oration. The Athenians would certainly not have viewed this as “meritocratic dogma”, merely a sensible and democratic way of running the state. The meritocratic element is the only part of the demokratia that has survived into modernity, so rejecting the principle tout court is an overreaction.

    * The council was the obvious exception, but the sheer size of it meant that sôphrosunê and dikaiosunê would predominate. But the council was not a sovereign body, it was a collective magistracy.

    Like

  21. Yoram: “Why would the qualification criteria for the lottery pool be more easily measurable than the selection criteria for the “best” candidate?”

    The pool selection criteria are a subset of the criteria for selection of the “best” candidate. You pick those criteria because they are easily measurable, that’s why they are easily measurable.

    You do not deny that there are other criteria that may be just as important, or even more important. It’s just that if we can’t measure them in a fair and accurate way, it’s fairer not to try to filter on them at all. For the important attributes that you can’t measure beforehand, you go for diversity instead – maybe you’ll at least be able to filter afterward.

    Let’s say you are a coach assembling a basketball team out of 200 candidates, and you have only very little, mostly subjective information about how good they are. You are concerned about getting a good team, but you are also concerned about being fair, and you are concerned about subjective biases.

    So, do you pick the 5 tallest people? Do you go with a combination of height and your subjective criteria, risking that you subjective criteria systematically and unfairly discriminates against some people? Or do you set some height threshold, and pick randomly from the people above it?

    It seems obvious to me that the latter approach is both more fair and more practical. Especially if you have the option to change team members later on, based on their actual performance.

    I’m assuming that there are actually right answers here, though. If you’re searching for a specialist, then hopefully you’re looking for someone who can deliver directly measurable outcomes. If what they deliver can only be judged subjectively so that some will say “they did exactly what I wanted them to!” and others say “They were horrible!” then by all means use lottery all the way.

    Like

  22. vintermann: you talk of selecting a basketball team. Did you realise that the NBA – National Basketball Association in the US uses a lottery to pick the new players coming out of college to join the major league teams? (see NBA lottery on Wiki)

    OK, the aim here is only the ‘egalitarian’ one of ensuring all teams stay in the game, that one team cannot dominate with money-power.

    Like

  23. Harald,

    In the context of academic positions, it is quite common to use “objective” selection criteria, like number of publications or number of publications weighted by impact factor, because they are “objective” and therefore “fair”. (I am writing “objective” because, of course, these criteria are not objective, they simply substitute subjectivity during the publication process for subjectivity during the appointment process.)

    So would you say that picking the person with highest number of publications is bad but having a lottery among all the people whose number of publication is above some threshold is good? Supposedly neither procedure is biased, because both rely on objective criteria. But if number of publications is biased one way or another, wouldn’t this bias affect the composition of the pool of qualifying candidates?

    If neither procedure is good, what would be a set of criteria and a procedure that you think is fair for academic appointments?

    I would tend to think that the thorniness of this set of question indicates that the problem is more fundamental, rather than in the appointment process alone. It seems that the entire hierarchical academic structure needs to be examined critically.

    Or to pick on your basketball example, why should this kind of competition for being a member of the basketball team even exist? Why not allow whoever wants to play basketball to have the opportunity to do so? It is the competitive, elitist structure itself that is the problem here.

    Like

  24. The classic meritocratic defense:

    “She did not come close,” Hermle asserted, adding that Pao “lacked the ability to lead others, build consensus and be a team player, which is crucial to a successful career as a venture capital senior investing partner.”

    Like

  25. Yoram,

    >why should this kind of competition for being a member of the basketball team even exist?

    And then you answer your own question:

    >It is the competitive, elitist structure itself that is the problem here.

    Exactly, most athletes, coaches etc. want to win. Given the origin of competitive games in warrior cultures like Greece, competitive sports might well be seen as a form of sublimated warfare (a similar argument can be made for the competition between political parties). Whether competitiveness is an intrinsic element of “human nature” or a form of learned behaviour is debatable (I imagine that you would adhere to the latter view). But whatever the cause, your case for sortition would appear to depend on a fundamental realignment of existing behaviour and motivations. My preference would be to accept human beings as they are and to devise ways of ensuring that competition does not descend into conflict.

    Like

  26. Andre,

    > modern decline of « the traditional justifications for inequality » did not only stimulated the ideology of meritocracy, it stimulated likewise the ideologies partitioning the human kind along race, sex, class origin ….

    Discrimination along those lines is traditional. Modern ideology condemns explicit discrimination based on those criteria (but accepts discrimination against the same people as long as it is justified by merit).

    Like

  27. It is really a sidetrack to ask whether competitive games are cool or not, or get into a discussion with Keith Sunderland on the Glory of Merit. It was only an example of a time when you have to pick someone, and hope that they are “better” in some sense, despite having imperfect and possibly systematically biased measures. If you don’t like the example Yoram, I’m sure you could come up with a better one.

    “So would you say that picking the person with highest number of publications is bad but having a lottery among all the people whose number of publication is above some threshold is good?”

    No, I’m saying it is better. It does not eliminate bias, but hopefully it will prevent the bias from having quite as disastrous side effects. You’re still going to have a lot of people who have gamed the selection criteria (e.g. citation) to the fullest extent possible, but now you also have some people who haven’t – or at least haven’t done it quite as bad. Longer term, you reduce the incentive to game the selection criteria.

    And what is your solution? Assume that we’re not just going to abolish the positions – if you want to abolish them, please think of something similar you would not want to abolish.

    Would you pick randomly between people who had any publications at all? In that case you argue the same as me, only with a very low threshold. I’m not saying such a low threshold would be wrong. Would you pick based on subjective, intangible criteria impossible to scrutinze or criticize – and just throw up your hands and say “there’s nothing we can do about that” to the systematic injustices that may follow? Would you use quotas for visible categories that you think are important (gender, race, etc.) and possibly risk aggravating the skew on invisible categories that you didn’t think of?

    Like

  28. By the way, the story around Pao and KPC&B is an utter farce. Talking about injustice in the intriges of players at that level is a joke, in my eyes they’re all hopelessly corrupt. I can’t conjure up sympathy for any of the parties, and if I could I’d gladly replace them all with people drawn from a pool with a very low threshold.

    Like

  29. Harald,

    > It is really a sidetrack to ask whether competitive games are cool or not […] If you don’t like the example Yoram, I’m sure you could come up with a better one.

    I do like the example – I think it demonstrates that changing the selection procedure is the wrong place to look for improvements.

    > it will prevent the bias from having quite as disastrous side effects.

    Possibly. It will depend on the particulars. In the case of gender bias, for example, we are assuming that men are more over-represented at the upper-decile of the publishing scale than they are at the, say, upper-quartile of the scale. That could be the case.

    > And what is your solution?

    My solution would be to try to eliminate the situation where such coveted positions exist. Whatever the benefits of the coveted position are, we should try to share them across all those interested rather than have them all acquired by a single person, however selected.

    Like

  30. > the story around Pao and KPC&B is an utter farce

    I agree. Pao is fine with the general elitist structure, as long as she is not blocked from climbing higher up the ladder. Again, this is the standard meritocratic attitude. The question, as far as this ideology is concerned, is not whether stratification should occur, but whether the criteria for the stratification are just, i.e., merit-based.

    Like

  31. Eliminated coveted positions everywhere, that sounds somewhat ambitious to me!

    If you’re ever going to make it, I think we must first try randomizing elite appointments ever so slightly, and see that the world doesn’t fall apart.

    Which I’m confident it won’t, in fact I think it can have positive effects, since it e.g. lets us focus slightly more on education for utility’s sake rather than education as signalling.

    The real benefit comes when the specialist applicant realizes “I’m now as attractive as I’m going to get for the purposes of getting my dream job”, and instead of focusing on looking good to employers, can focus on actually being good at the task in question.

    Like

  32. Harald,

    >The real benefit comes when the specialist applicant realizes “I’m now as attractive as I’m going to get for the purposes of getting my dream job”, and instead of focusing on looking good to employers, can focus on actually being good at the task in question.

    That’s the best argument I’ve heard for sortition from a merit-based shortlist as it might lead to better behaviour, rather than just achieving an abstract goal like equality (or, on the other hand, it might just lead to lazyness). But eliminating the human element by mechanising the entire appointment process rules out the possibility of aiming for the best (as opposed to the merely adequate), which would not work for some key appointments.

    Like

  33. If I might summarise: It’s a trade-off.
    –A lottery-mediated appointment may result in someone less than the ‘best’ getting the job, but at least it is systemically equal-opportunity.
    –If the conventional interview-and-human-judgement procedure is the only way of finding the ‘best’ appointee, stick with it, despite its incubus of sexism, racism, etc.

    Like

  34. I don’t think this summary is justified.

    Unless we (1) assume that qualification to the lottery pool would be based on “objective” measures, and (2) assume that “objective” measures are not biased, then there is no reason to conclude that “a lottery-mediated appointment […] is systemically equal-opportunity.”

    Even if we do accept those assumptions, then the objection to “bias-free” lottery may not rest upon a quest for the “best” appointee. It may simply be that “objective” measures are largely irrelevant for finding “good” candidates and that any reasonable selection of candidates to the lottery pool would have to involve human judgement.

    Like

  35. I believe that the finer distinctions you want to make between people, the more likely it is that your measure is systematically biased. The objective measure “went to high school” is a crude measure of software developer skill, but gives you less elbow room for indulging your systematic biases than “fits the culture of our firm”.

    The goal is not to eliminate human judgment, but to circumscribe it and get it out in the open where it can be examined and criticized.

    Like

  36. That’s the Onion. It’s satire.

    Like

  37. I assumed this is common knowledge.

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: