## Thursday, January 20, 2011

### What is Afghanistan’s potential?

The armed services administers a test to all potential enlistees called the Armed Forces Vocational Aptitude Battery (ASVAB). This exam consists of nine subtests in different subjects. As Steve Sailer has explained on numerous occasions, four of these subtests are combined to constitute the Armed Forces Qualification Test (AFQT), a highly g-loaded (i.e., roughly correlated with IQ) exam that determines enlistment eligibility. Additionally, the subtests are combined in different ways to determine eligibility for specific career fields.

For instance, the Air Force has an enlisted career field (AFSC) designated "3E032: Electrical Power Production". Admission to this career field requires meeting thresholds along two different aptitude dimensions: mechanical (M) and electrical (E). These dimensions are calculated from ASVAB subtests using the following formulas:

M = MC + GS + 2 x AS
E = AR + MK + EI + GS

The indicated subtests are:

MC: Mechanical Comprehension
GS: General Science
AS: Auto & Shop Information
AR: Arithmetic Reasoning
MK: Math Knowledge
EI: Electrical Information

The thresholds for these tests are:

M: 56th %-ile of U.S. population
E: 40th %-ile of U.S. population

These percentiles aren't especially high.  We aren’t talking about an engineering degree here, merely the aptitude to work as a technician in a power plant.  No doubt any reader of this blog would qualify for this AFSC. (Full disclosure: according to the ASVAB site, I know next to nothing about automobiles, although I made some shrewd guesses on the shop questions, which are broken out separately.) The electrical dimension is substantially g-loaded, two of its subtests being used to calculate the AFQT score. In theory, the other subjects should be teachable; significantly, however, the Air Force declines to teach these subjects as part of the technical training it provides to members of the AFSC.  Rather it expects them to know the information before they enlist, or to learn it on their own.

Of the two dimensions, E is the more g-loaded, as two of its subtests are also used to calculate the AFQT score.  Note that these percentiles are to some degree cumulative, depending on the degree of covariance between the dimensions.  I tried in vain to find this covariance; the best I could do was this 2006 paper that identified correlations between several of the subtests.  The relevant correlation is between AR and AS:  .35, not especially high.  On the other hand, the GS subtest is a factor in both dimensions, and its correlation with itself is obviously 1.  As a back-of-the-envelope calculation, we will take the midpoint between these two numbers as representing the correlation between the two dimensions:  .672.

A quick Monte Carlo simulation in Matlab tells us that with two dimensions this highly correlated, the overall qualifying percentile is only slightly above that for M:  58th.  I shall refer to this value as ME58.

Armed with this overall percentile, I can now use the inverse Gaussian cumulative density function to calculate that this is about .2 SDs above the population mean, equivalent to an IQ of 103.

Question:  What would this look like in Afghanistan?

Lynn and Vanhanen measured the Afghan IQ as a full SD below the American mean.  If we assume:  first, that the IQ variance among Afghans is the same as among Americans; and second, that the distribution of ME aptitude mirrors the distribution of IQ, then we can see that the aptitude cutoff for Electrical Power Production is 1.2 SDs above the Afghan mean, which corresponds to the 88th percentile.

Now, on the one hand, in a nation of 21 million, the 88th percentile still holds plenty of people capable of running the country’s power plants.  But on the other, this requirement must compete with all the other demands of the kind of technical society that USAID and USACE are determined to foist upon the Afghans.  When you add to the consideration the superior opportunities open to the 88th percentile (e.g., looting USAID contracts), you begin to wonder if there is really enough aptitude in Afghanistan to run the society we want them to have.

Professor Hale said...

Add to that the problem that the local smart kid is more valuable to his tribe as a goat herder than as a generator mechanic, so he gets no education at all and you create a system of vast unusable potentials.

Professor Hale said...
This comment has been removed by the author.
Professor Hale said...

BTW, Impressive math/stats skills. If anything it just goes to prove that IQ is not a pre-req for placement in positions that would benefit from intelligence. Otherwise you would be assigned to be Emperor of Afghanistan. I have my doubts that Karsai could score high enough to run a porta-john.

Φ said...

Add to that the problem that the local smart kid is more valuable to his tribe as a goat herder than as a generator mechanic, so he gets no education at all and you create a system of vast unusable potentials.

This is an excellent point. So good, in fact, that I used it in a set of talking points I wrote for the boss this evening. My impression is that Afghanistan has a considerable merchant class, so they aren't all goatherders, but nor do they make a habit of seeking their own way in an impersonal money economy.

Dexter said...

You may or may not have seen this article, in which an Air Force colonel muses on the futility of trying to create a functioning country when the local population consists of people who cannot even successfully inflate a basketball. He concludes we are "multiplying by zero" in Afghanistan:

http://usacac.army.mil/CAC2/MilitaryReview/Archives/English/MilitaryReview_20110228_art014.pdf

"Afghanistan and the Afghans provide such a limited foundation to build from that "by, with, and through" simply may not be feasible. In many ways, we are multiplying by zero. The Afghans have limited infrastructure; limited agricultural capability; limited to no indigenous industrial capacity; an immature consumer economy; an impotent and incoherent security apparatus; and a fledgling Western-style government overseeing a decentralized, tribally based population. No foundation exists to to build on. The lack of an existing infrastructure prevents the creation of second- and third-order economic effects, construction of a security force, and the development of functioning public transportation and communication services. The United States is investing in a country in which there is literally nothing to invest. Virtually everything the U.S. uses has to be imported because Afghanistan is fundamentally underdeveloped.
What I witnessed in Afghanistan is best summed up in Robert Kaplan's The Ends of the Earth. Kaplan notes that when the United States began the Peace Corps in the 1960s, both Sierra Leone and India required basic agricultural know-how. Thirty years later, India had become a net food exporter and a producer of high technology with no further need of farm assistance. Sierra Leone, on the other hand, remained exactly where it was in the 1960s when the Peace Corps first arrived. The message of Sierra Leone was brutal: The end was nigh in the failed battle, fought valiantly by the liberal West, to equalize cultures around the world. The differences between some cultures and others (regarding the ability to produce exportable material wealth) appeared to be growing rather than diminishing. I could substitute Afghanistan for Sierra Leone. It was difficult to make my interpreter understand this, but he knew it when I asked where the ISAF would get its water, its rental cars, and its Internet service. He knew that whatever we needed would come from somewhere other than Afghanistan."

Kipling is as apropos as ever:

Take up the White Man's burden--
Send forth the best ye breed--
Go bind your sons to exile
To wait in heavy harness,
On fluttered folk and wild--
Half-devil and half-child.

Take up the White Man's burden--
In patience to abide,
To veil the threat of terror
And check the show of pride;
By open speech and simple,