Date sent:      	Wed, 01 Dec 1999 08:12:54 -0500
From:           	Charles Murray 
Subject:        	[h-bd] Re Ogbu and AFQT


[ Double-click this line for list subscription options ]


Just two quick points to respond to a few specifics in one of Scott's
earlier messages.


1. John Ogbu's work is a good example of something that enters the
debate as a sort of icon. Everybody becomes familiar with the basic
premise and a few examples (the Japanese castes, etc.), and hardly
anyone actually confronts the work itself. Scott may have access to
other Ogbu work that I've missed, but the titles that made his
reputation in the 1980s have no serious data. There's no there there.
The references to which I refer are  Ogbu, J. U. (1986). The
consequences of the American caste system. In U. Neisser (Ed.), The
School Achievement of Minority Children (pp. 19-56). Hillsdale, NJ:
Lawrence Erlbaum Associates, and  Ogbu, J. H. (1987). Cultural
influences on plasticity in human development. In J. J. Gallagher & C.

T. Ramey (Eds.), The Malleability of Children (pp. 155-169). Baltimore:
Paul H. Brookes. I stand eagerly ready to be refuted on this point, as
long as I can call on the collegiality of h-bd to observe a ground rule:
not "Can I label something in these titles as data?" but "Seriously now,
would I consider these bits of evidence as anything I myself would be
willing to put under my name in defense of my position?"


2. The AFQT is one of the best paper-and-pencil measures of IQ available
for adolescents and young adults. Its median correlation with full scale
IQ tests such as the WISC and S-B in the NLSY sample is .81, higher than
the median correlation of the WISC and S-B with other cognitive tests.
Herrnstein and I gave over an entire appendix to discussion of the
psychometric characteristics of the AFQT, for anyone who's curious.


Charles Murray






From:           	"Scott MacEachern" 


On 1 Dec 99, at 8:12, Charles Murray wrote:


> Scott may have access to
> other Ogbu work that I've missed, but the titles that made his
> reputation in the 1980s have no serious data. There's no there there. 


Actually, the primary work that I've read has been his 1978 
Academic Press book, _Minority education and caste: the 
American system in cross-cultural perspective_, and that AFAIK is 
the work that he bases much of the later work on. I don't have a 
copy at my desk now, but there seems to me to be quite a 
significant there there -- quite a good book, actually. I would 
recommend it to you. If I may change your question to "Seriously 
now, would I consider that book as anything I myself would be 
willing to put under my name in defense of my position?", the 
answer is 'yes'. 


> 2. The AFQT is one of the best paper-and-pencil measures of IQ available
> for adolescents and young adults. Its median correlation with full scale
> IQ tests such as the WISC and S-B in the NLSY sample is .81, higher than
> the median correlation of the WISC and S-B with other cognitive tests.
> Herrnstein and I gave over an entire appendix to discussion of the
> psychometric characteristics of the AFQT, for anyone who's curious.


Yes, I know -- I read it with the rest of the book. I've also read some 
of the criticisms of that work, _Inequality by design_ for example, 
including a number that included sample questions from the AFQT. 
The issue from my point of view thus isn't whether the AFQT is a 
good proxy measure of IQ test results (and so I apologize for my 
last mistatement) but whether it is a good measure of innate 
intelligence/cognitive ability. It seems to me from reading the latter 
works that you've consistently underestimated the various 
environmental effects that affect AFQT (as you and Dr. Herrnstein 
consistently underestimated environmental effects through the rest 
ofthe book) -- and that it is, as I said, a test of academic 
preparation.


Scott
________________________________________________
Scott MacEachern