Asymmetrical Information

Hmmm.. Probably obtained in the same manor that I am actually an ordained minister in California.

I forget the church, but it entailed filling out an online form and *poof*, I could marry you in California.

Now all I'm trying to do is find a loophole that will allow me to marry two people without their knowledge!

Robb: I can't speak authoritatively for CA, but at least in other states, simply "being ordained" isn't enough. You need to, like, file paperwork with the state, and perhaps give them some money.

On topic, well... that's what accrediting authorities are for. Very libertarian, that. Anyone can puy a piece of paper saying "I got educated", but without a ceertifier putting its reputation behind the paper (or the school itself, in theory, if it had a stellar reputation. I imagine a Harvard degree would still be "worth" something if they told the accreditors to take a hike.) it's just a piece of paper.

Kinda like money, in a way, except not exchangeable for goods and services.

"Unaccredited school" doesn't necessarily equal "diploma mill." I don't know anything about the schools in question, but I know that new schools have to wait something on the order of five years before they can even be considered for accreditation.

In this case, "diploma mill" seems quite accurate. All three of the named schools are "accelerated, distance learning" institutions that don't require that one attend class. They try to spin themselves as geared toward busy professionals who are too busy for traditional education, but this seems closer to Sally Struthers than an executive MBA. Pacific Western says that

The University is oriented to those individuals not seeking licenses or credentials requiring accredited degrees. Our programs are not designed to meet any established requirements by private or professional associations. If a license or a credential is desired, a check should be made of state, federal, association, and credential requirements before applying.

The first question that came to my mind was whether the identified people were doing a good job or not. I understand that is a heretical view in "public service", but in the real world, the CV/resume just gets you the interview. After you are hired, only competence (not necessarily in the job you were hired for, mind, sometimes office-political competence) keeps the checks coming.

If this is being used to investigate the skills of the people granting interview slots, great. They haven't been doing their jobs in checking references. If it will only be used to fire a few people who don't have the correct boxes checked, it's, at best, misguided. Either the employees are doing their jobs adequately, in which case they should be kept in those jobs, or they are incompetent, in which case they should have been fired for incompetence already.

You create a bureaucracy in which people get hired quicker or paid more for some fancy credentials, THEN you fail to install some sort of quality control.

Why is anyone surprised that people find ways to answer the bell without investing the time?

hmmm...yeah...what's the big deal?
Were these people competent in their jobs?
Did they try to fraud anyone when they claimed they
had a degree? (e.g. They claimed it was from Harvard and it was not?)

As for accrediting, it *is* useful in lending reputation, but it does not guarentee quality education, nor does the lack of an accredited degree imply inferior education or ability.

On a side note, the University of Arizona (a fairly well known tier-1 school) operated it's
School of Information Resources and Library Sciences (SIRLS) with out accreditation for a year
or two.

Sigivald:

"I can't speak authoritatively for CA, but at least in other states, simply "being ordained" isn't enough. You need to, like, file paperwork with the state, and perhaps give them some money."

A little off topic.

http://northernway.org/marriagelaws.html

The above-referenced link will tell you the officiating laws by state (and the specifics in NYC and Canada). In New York State, for example, you just have to be ordained. You don't have to file with the state, you don't have to be approved. Any religion (with the exception of Universal Life Church because they were judged to be a tax dodge) is okay. To perform services within New York City, you have to have been sworn in and paid a $30 fee as well as show them your documentation.

FYI -- In New York State, even if the person performing your wedding isn't a minister, but you THINK they are and you only find out later that they aren't...it doesn't matter. Here in New York State if you think you're married and filed the license...you're married.

Sorry for that little random data-dump. Thanks.

Doug & Nate:

Competence in a job, particularly a high-level job, is hard to determine. For one thing, there are any number of factors and people below you on which to blame poor performance (for ex., I would claim that GWB is incompetent, while I'm guessing most people here would blame poor present results on other factors). For another thing, in high level positions, the baseline is amorphous - in some ways you are hoping that a bright person will produce good results that you cannot anticipate. If such results are never produced, but were never predicted, you have no way of measuring (and therefore recognizing) that loss.

Just as education can act as a signaling mechanism (admission: Jane's analysis first drew me to this blog), it is also, in closely related fashion, a proxy for intelligence or determination or other good qualities. If you believe that companies reasonably hire people with such qualities because they anticipate good outcomes based on little more than the fact their workforce has such qualities, and you believe that education is a good proxy for such qualities, then fake degrees are fraud even if the company hadn’t previously noted any incompetence on the part of the fake degree holder.

By way of minor digression, I note that one of the (I think) better justifications for affirmative action in the educational setting is that getting into a top-level school is usually much more difficult than (a) succeeding at that school, and (b) succeeding in that field. The level of achievement it requires to get into, for example, the U Chicago business school doesn’t correlate well with what you need to do well there or to do well in the chosen business field. Therefore, the B-school might as well look at all competent students (assume competent = min. score of minority candidates) and then choose from the pool above that competency standard by whatever crazed reasoning process that they want (including attempts at social justice). This would seem to be in keeping with what the two of you are saying about degrees and competency, but somehow I doubt y’all favor or even accept the legitimacy of affirmative action programs.

The government runs a market in paper that confers special privileges without merit. Should this latest twist suprise us?

The problem with your idea, Tim, is that the signalling mechanism instantly breaks down as soon as employers know about it -- and employers do know. In fact, one would assume that it hurts minorities who have achieved the same standards as the general population more than it helps the others, because the better an employer is, the more likely they are to be aware of affirmative action, and thus aggressively discount the accomplishments of minority applicants.

I have multiple objections to affirmative action in education; it seems to me to be neither effective, nor just, unless we measure effectiveness solely on the metric of how many bodies of a particular skin color are occupying seats in freshman classes, and justice only by the percentage of melanin in one's skin -- for from what I understand, affirmative action increasingly privileges minority children from affluent backgrounds who have had every advantage over less affluent whites and asians. That seems neither fair, nor effective at ending the legacy of slavery.

I think it was PJ O'Rourke who said that it's almost impossible to measure the output of government workers, so you measure the input. He was explaining why bureaucrats are always at their desk early in the morning, but it would also apply to things like number of degrees possessed by a given employee.

This subject came up about 6 weeks ago on Joanne Jacobs website. In that case, about a dozen Georgia teachers were found to have purchased degrees. The state had a "Professional Standards Commission" which was supposed to vet these, but apparently was a bit lax.

In my experience in private industry, relatively few jobs require a particular degree. I have seen some companies use degree requirements for outside recruits (to limit the applicant pool) but not require them from in-house applicants. Unlike government, industry tends to go by certifications (MCSE, or CPA, for example), and of course by competency.

Leaving aside the diploma mill problem, I've always wondered about people who were impressed with a (say) Harvard degree. Granted that it's better than no degree at all, but people slide thru Ivy League colleges with C and D averages. Even Harvard!

If you cannot determine the competency of someone you are hiring, you are not competent to hire. If you are requiring a diploma just for some class seperation thing, then what the hell. Hire the guy with the fanciest paper.

Jane:

You misunderstand me, which is probably my fault. Here's another try.

1. Signal Discounting

I’m really just saying that given the relatively rarified pool of acceptable applicants to an “elite” school (including the minority candidates), the correlation between their test scores (for ex.) and their abilities in the field is almost certainly weak.

This doesn’t seem like a very controversial point. I’m all but positive that the correlation between GMAT scores and grades at U Chicago is very weak, because the student pool is restricted to such a narrow band of possible scores (what, 750-800?). I’d bet that the correlation between GMAT scores and performance after school is even weaker.

What then does the fact that your friend got a 750 and you got an 800 mean, Jane? Nothing. Maybe you can parlay the difference into a beer, if you can set up the appropriate bar bet. But, as we said above, we can at best weakly predict a better performance by you after school. To any sophisticated potential employer, you and your friend are essentially the same (I’m ignoring MBA grades, interests, your height advantage, etc. – assume they are the same).

Now if the same weak correlation holds for the point between your friend and the lower (for the purposes of this argument) required minority GMAT score (say 700), then your friend and the minority student are the same to the employer. No discounting needed.

If U Chicago is looking for students who are as able as your friend, they might reasonably only require GMAT scores above 700. After all, they can sort of distinguish between you and the minority student, but they can’t distinguish (essentially) at all between you and your friend, or your friend and the minority student. At that point, they still have too many acceptable applicants. They have to winnow, and it isn’t going to be on the basis of merit. And almost any methodology is going to have hidden biases. They might chose everyone who can dunk (I’m assuming this includes you, Jane), which would skew toward the athletic and the tall. They might choose people with certain school pedigrees, which would skew moneyed. Or they might decide to try and pick a diverse class (as among a number of axes). It’s all whim – the group they are picking from is all relatively meritorious.

2. Addressing Social Justice

Does such diversity achieve or even really address social justice (whatever that means)? I don’t know – maybe not, though I suspect it does at least a little. But, given the really small range of scores allowed in, any criteria will have hidden (or at least perceived) biases.

Maybe it doesn’t help address the pernicious effects of slavery (NB: it’s not really slavery we worry about, it’s the pre-Civil Rights Act bad behavior and its remnants that create the problems to be addressed – that is, the relevant date of “it was so long ago” arguments is 1964 at the earliest). Maybe, as people have argued about the “compassionate conservative” trope, it’s really just a sop for liberal white women. But I like liberal white women (again, Jane, you should really join the DLC), so I’m OK with that. And it seems to be not terribly less fair than any other system.

I acknowledge that the argument depends on the weakness of correlation – I’m sure I’ve seen evidence of this weakness, but I’m unwilling to look for it and cite it. Mostly because I'm a centerist-liberal bastard.

Walter Wallis: "If you cannot determine the competency of someone you are hiring, you are not competent to hire."

Walter - I am a tax accountant and have interviewed a lot of candidates over the years. Absent giving a test, I've yet to figure a way to quantify someone's tax skills in the 20 - 30 minutes I have to interview an applicant. This is true whether I am hiring someone off campus or someone with years of experience. We are very, very, careful in whom we hire, but it's still a crap shoot. I suspect the same is true for many other professions.

The sad thing is that even a "real" degree doesn't mean that much these days, the curriculum has been so dumbed down.

http://www.transterrestrial.com/archives/003700.html

Two comments:

(1) Being an unaccredited school in CA doesn't necessarily equate to a diploma mill. There are tons of unaccredited law schools in CA which allow you to take the CA bar exam, but not other states' bar exams. Not a problem if you plan on practicing in CA.

(2) GMATs, like SATs and LSATs, have never been used as a predictor of grades or success. What they are, especially when coupled with undergraduate GPAs, is an excellent predictor of FINISHING THE SCHOOL you are accepted to, which is really critical to graduate/law schools, given that few people transfer into programs. That's all, and any graduate admissions office will tell you the same.

Rex:

Because my argument depends upon it, I checked, and I'm pretty sure that you are wrong about the use of GMATs, LSATs, and SATs as predictors of grades. That is precisely what the scores are supposed to correlate with - specifically, with the first year grades in a given institution.

If a test were merely used to determine whether you would finish the program, how could it be used to determine whether you were "smart" enough for the "elite" institution?

Anyway, you can cross-check me by googling [test name] + predict + "first year".

What's interesting to me is that this discussion seems relatively uninterested in the ethics of lying.

Lying to obtain something of value from another person (i.e.,jobs) can't be fraud in this case, because the jobs were government jobs and probably aren't worth much.

If the 90's taught us anything, they taught us that lying about something inconsequential isn't really lying.

Similarly, lying to obtain something of inconsequential value can't be so bad, certainly not rising to the level of fraud.

For that matter, many persons who already have a job are liars. It's just not fair to discriminate against the not-yet-hired. If the 60's taught us anything, they taught us that discrimination is unacceptable in all its forms.

Telling the truth is of course useful as a sort of hide-in-plain-sight tactic to fool the gullible. It's also advisable when there is high likelihood that one will be caught in a lie by one of those moralisers you can still trip over from time to time.

A tactic. Telling the truth is only a tactic, just like telling a lie. This moral equivalence harms no one, and discriminates against no one . . except the honest.

Come now Jane, how can one make a post on education without a little economic theory? What about the way in which this violates the Spence-Mirlees Single Crossing Property? Marginal disutility of education is seemingly not decreasing in type for these kinds of degrees. :-)

er... That should be "Spence-Mirrlees". I should really quit squinting at the screen and go find my glasses.

GMAT correllates .49 with first year grades. (click link)

Okay, here's how it works in more detail.

Talk to any honest admissions director. There is generally a much larger "qualified" pool of applicants of people who can do graduate level work if they put their minds to it than there are available slots at any particular institution. The GMATs, LSATs, GREs, etc. do a good job at weeding out those folks who are unable to do graduate level work (predictors of who has the ability to do the work), but they are not predictors of graduate level grades or success in life. Similarly, undergraduate GPAs are good predictors of students who are willing to make an effort in their classes. Put the two together and you have an excellent predictor of someone who will finish the graduate level school.

It's up to the academic folks to ensure that standards are kept up; what the school needs from its admissions folks are students who will pay the required tuition for the length of the school. This used to be true of undergraduate admissions as well, until the "open enrollment" movement started (when the pool of applicants decreased nationwide), whereby colleges admitted vast numbers of unqualified students and exposed them to a sort fo Darwinian process, letting the unqualified flunk out or drop out after their frosh year. The survivors usually finished the remaining years.

Graduate schools can't afford to work that way. They have to be precise with their admissions, because (1) the total class sizes are much smaller, (2) they only have the students for 1.5-3 years, and (3) there are very few transfer students at the graduate level to fill any open slots.

I don't know if there is any correlation or how strong it is between GMATs and graduate grades, but that is NOT their purpose NOR how they are used.

"Absent giving a test, I've yet to figure a way to quantify someone's tax skills in the 20 - 30 minutes I have to interview an applicant."

David, either give the test or, more appropriately, monitor the work closely enough during the probationary period to evaluate actual performance. The probationary period, properly used, is an extension of the hiring process.

"I think it was PJ O'Rourke who said that it's almost impossible to measure the output of government workers, so you measure the input."

I immediately thought of a counter-example: the DMV. It's quite easy to measure throughput at the counter, and the resulting affect on waiting time (vs. the employee/applicant ratio). It's also obvious in many states that the people in charge just don't bother to monitor throughput or waiting time. OTOH, in Michigan where an elected official is directly responsible and his or her picture is on the wall in the waiting area, service has never been too bad.

Walter - I don't disagree that the probationary period gives ample time to access someone's skills. There are two problems, at least in my industry, with this hire-someone-to-see-if-they-pan-out method: First, a CPA firm's "product" is the professional skill of its people. We don't want to introduce someone to a client only to have to explain to that same client that the guy who had been working on their account for the past 3 months just wasn't up to the job. Second, it's an awfully expensive approach. The only thing we sell is time, so burning 3 months only to start the process over is something we are loathe to do. Besides, it typically takes 6 months to a year to find the "right" person, so you don't want to start over if you can help it.

Which takes us back to testing the applicant. The problem is I am unaware of any test we could use for this purpose. The kind of tax problems we work on require several hours of research to solve. So, in the case of experienced hires, we rely heavily on the kind of experience the candidate has had with other firms -- the types of clients he or she has worked on, the responsibilities the candidate has been given, etc. -- as a proxy for the candidate's skills. In the case of a new hire, we rely on the strength of the accounting program the candidate is graduating from. If we are familiar with the school, we have a fair idea what kind of skill and effort were required to graduate. If we are unfamiliar with the school, we will seldom even interview the person.

r=.49 is not very much of a correlation. When testing a hypothesis, r had better be at the .95 level or higher (sometimes .90 is used). r=.49 is basically even odds, i.e., there is one chance in two that the relationship is linked.

For the authors of the article that Bob Dobalina linked to say that there is a "strong correlation" indicates that they know little of statistics.

But I'm not confident in the study anyways, because I don't think that graduate grades exhibit a normal curve.

David. Find a tax return that gave one (or more) of your present employees severe headaches. Give it to the candidate as a test. See what they do with it and how long it takes them (you could offer a small reimbursement for their time). You can even 'play' the client and drive them bats by not giving them information and forgetting half of it, requiring the candidate to keep requesting the missing info -- or by giving it all to them in several large unsorted boxes. (Yes, I've prepared taxes, why do you ask?)

On a more serious note, James Joyner also covered this, you might find his comments worth reading.

Dave - go to work for the IRS.

r=.49 is not very much of a correlation. When testing a hypothesis, r had better be at the .95 level or higher (sometimes .90 is used). r=.49 is basically even odds, i.e., there is one chance in two that the relationship is linked.

For the authors of the article that Bob Dobalina linked to say that there is a "strong correlation" indicates that they know little of statistics.

Rex is indeed confused about his statistics.

If you run a regression of individual earnings
on age, age squared, years of schooling and a
test score (add in some other demographic
characteristics if you like) the test score
will be statistically and substantively
significant. This is, at some level, the
whole point of the Bell Curve book, stripped
of all the political interpretation. So test
scores are informative about success. At the
same time, they do not explain all that much
of the variance in an earnings regression.

I've been on a graduate admissions committee.
For what it is worth, we use GRE math scores
(this is an economics department). Bascially,
if you are over 700 out of 800 you are in the
pool and if you are below 700 you are out of
it. Within the 700 and above pool, the level
doesn't matter much - other things do, such
as grades in technical undergraduate classes.

As to inputs and outputs of government workers,
read James Q. Wilson's wonderful book on
*Bureaucracy*. It does a very good job of
explaining the limits to the contracting out
version of privatization.

Jeff Smith

Comments are Closed.