Peter Schonemann on
Peter H. Schonemann
For the complete bibliography and abstracts: http://www.psych.purdue.edu/~phs/research.htm
(1978) Schonemann and Steiger, J. H.
On the validity of indeterminate factor scores. Bulletin of the
Psychonomic Society, 12, 287-290.
(1981). Factorial definitions of
intelligence: Dubious legacy of dogma in data analysis.
In I. Borg (Ed.), Multidimensional data representations: When
and why. Ann Arbor: Mathesis Press, 325-374.
(1983). Do IQ tests really measure intelligence?
Commentary, The Behavioral and Brain Sciences, 6, 311-315.
(1987). Jensen's g: Outmoded theories and unconquered
frontiers. In S. and C. Modgil (Eds.), Arthur Jensen: Consensus
and Controversy. Barcombe: Falmer Press, 313-327.
Abstract: An invited chapter to a volume devoted
to Arthur Jensen's work. Originally, the editors intended to have
two contributors per topic, one favorable and the other critical.
As it turned out, only two of the contributors were critical,
the other one being James R. Flynn ("Flynn effect").
The title borrows from Jensen, A.R. (1979) "g: Outmoded theory
or unconquered frontier?' Creative Science and Technology, 11,
16-29. The main story is contained in sec. 4, The great Society.
It gives a concrete and elementary illustration of the factor
indeterminacy issue, complete with two sets of equivalent factor
scores for the same 8x3 score matrix. For a more abstract treatment,
see Schonemann and Steiger (1978).
(1989). New questions about old heritability estimates.
Bulletin of the Psychonomic Society, 27(2), 175-178.
This note summarizes the main results of three recent studies
on the heritability of mental traits:
(1) The inferences Jinks and Fulker (1970) derived from Shields'
(1962) twin data are invalid since the assumptions of the
genetic model are consistently violated by these data. A purely
environmental model fits them better by a factor of 2.
(2) Holzinger's heritability coefficient (h**2) is invalid since
his derivations imply that dizygotic twins share no genes.
(3) In contrast, Nichols' (1965) heritability coefficient (HR)
follows from a strictly additive genetic model.
(4) However, the needed assumptions are consistently violated
by Osborne's (1980) personality data, which produce an
excessive number of inadmisible HRs. A purely environmental model
fits these data better by a factor of 14.
Jointly, these results suggest that heritability estimates of
mental traits in the literature should be view with caution.
(1989). Some new results on the
Spearman hypothesis artifact. Bulletin of the Psychonomic
Society, 27(5), 462-464.
This note constitutes a summary of the main results of two papers
devoted to Jensen's (1980) "Spearman Hypothesis":
(1) Elementary geometry shows that positive ("Spearman")
correlations between the mean black/white difference vector and
the leading eigenvector of correlation matrices are artifacts,
regardless how one interprets Jensens's ambiguous definitions
"Spearman's hypothesis" (Level I versus Level II interpretation).
(2) Empirically, the stronger Level II interpretation (which predicts
positive correlations with both within eigenvectors) also
arises with data that have nothing to do with g, such as SES variables
indicating the number of toys and games.
(3) Mathematically, the Level II interpretation implies not just
approximate but perfect collinearity between the mean difference
and the eigenvectors of all three covariance matirces, if one
assumes multinormality, positivity of both subgroup covariance
matrices, and an equal split into a HI and a LO group, regardless
whether Spearman's factor model holds or not.
(1990). A critique of the Jinks and Fulker reanalysis
of Shields' identical twin data. Abstract. 20th Annual
Meeting Program and Abstracts of Behavior Genetics Association.
Aussois, France, p. 54.
(1990). Environmental versus genetic
variance component models for identical twins: A critique of Jinks
and Fulker's reanalysis of the Shields data. Cahiers
de Psychologie Cognitive / European Bulletin of Cognitive
Psychology, 10, 451-473.
It is shown that the genetic model Jinks and Fulkr (1970) fitted
to the Shields' (1962) twin data is qualitatively inconsistent
systematic trends in these data and, as a result, produces an
inordinately large proportion of negative variance estimates.
In contrast, a purely environmental model yields qualitative predictions
consistent with the Shields data and admissible parameter estimates
throughout. Quantitatively, it fits the Shields data twice as
well as Jinks and Fulker's genetic model. Hence their farreaching
conclusions are not supported by the Shields data. This
reevaluation illustrates that purely descriptive models, even
if they were used with circumspection, remain intrinsically inconclusive
about nature/nurture questions because the possibility can never
be ruled out that other models may fit the same data even better.
(1990). Not beyond a reasonable
doubt. Cahiers de Psychologie Cognitive - European Bulletin
of Cognitive Psychology, 10, 671-674.
(1992). Extension of Guttman's result from g to
PCI. Multivariate Behavioral Research, 27, 219-224.
(1992). Henry F. Kaiser 1927-1992.
Multivariate Behavioral Research, 27(1), 161-163.
(1993). A note on Holzinger's heritability coefficient
h**2. Chinese Journal of Psychology, 35, 61-67.
It is shown that the still widely used heritability estimate
h**2 developed by Holzinger (1937) is not valid because Holzinger's
derivation of it was unsound: The variance component model he
used to derive h**2, together with his claim that it estimates
the genetic variance ratio, imply the counterfactual assertion
that dizygotic twins share no genes. While the competing coefficient,
Nichols' HR does indeed follow from the conventional variance
component model, the necessary conditions it rests on lack empirical
Sch�nemann & Sch�nemann, R. D. (1994). Environmental
versus genetic models for Osborne's personality data on identical
and fraternal twins. Cahiers de Psychologie Cognitive
- Current Psychology of Cognition 13, 141-167.
It is shown that the additive genetic model of Nichols
needed to justify the heritability ratio HR does not fit Osborne's
data very well. A purely environmental model with the same number
of parameters fits these data better by a factor of 14.
Compared with the additive genetic model, these empirical results
suggest that Osborne's personality data contain no genetic
component at all. The responses of the identical twins may be
more similar simply because they are exposed to more similar
environments than fraternal twins. This outcome illustrates the
general principle that conventional variance component models
used to justify heritability estimates are intrinsically inconclusive:
We can never rule out that another, qualitatively quite different
model may fit the same data equally well or, as in the present
case, much better.
(1994). Heritability. In R Sternberg (Ed.),
Encyclopedia of Human Intelligence. New York:
McMillan, p. 528-536.
( 1995) Totems of the IQ Myth: General
Ability (g) and its Heritabilities (h**2, HR). 1995 Meetings
of the American Association of Sciences.
Sch�nemann & Thompson, W.W. (1996)
Hit-rate bias in mental testing. Cahiers de Psychologie
Cognitive-Current Psychology of Cognition, 3-28.
New results are given concerning a form of test bias
which, in the past with few exceptions (esp. Cole, 1973; Hartigan
Wigdor, 1989) seems to have been largely ignored. We call it "hit-rate
bias" because it is defined as the discrepancy between
the hit-rates (= probability that a qualified testee passes the
test) in a low and a high scoring group. Typically, it favors
high-scoring group. In contrast to Cole (1973), our focus in on
binary criteria, such as college graduation. In the first, theoretical
part, we present a (Hit-Rate Bounds) Theorem which underscores
that raising predictor standards is not
equivalent to raising criterion standards, as some believe. Instead,
it typically increases hit-rate bias. We then derive and tabulate
a simple approximation for estimating hit-rates as a function
of validity, base-rate, and admission quota. In the empirical
portion of the paper, we evaluate the extent of hit-rate-bias
in practice by re-analyzing a number of data sets
involving the SAT, the ACT, and the GATB. Finally, we discuss
how the addition of tst score to high school record affects
hit-rate bias in predicting college graduation. We find it increases
(1997). Models and muddles of heritability. Genetica,
One reason for the astonishing persistence of the IQ
myth in the face of overwhelming prior and posterior odds against
be the unbroken chain of excessive heritability claims for 'intelligence',
which IQ tests are supposed to 'measure'. However, if,
as some critics insist, 'intelligence is undefined, and Spearman's
g is beset with numerous problems, not the least of which is
universal rejection of Spearman's model by the data, then how
can the heritability of 'intelligence' exceed that of milk production
of cows and egg production of hens? The thesis of the present
review paper is that the answer to this riddle has two parts:
(a) the technical basis of heritability claims for human behavior
is just as shaky as that of Spearman's g. For example, a once
widely used 'heritability estimate' turns out to be mathematically
invalid, while another such estimate, though mathematically valid,
never fits any data; and (b) valid technical criticisms of flawed
heritability claims typically are met with stubborn editorial
resistence in the main strream journals, which tends to calcify
(1997) Some new results on hit rates and base rates in mental
testing. Chinese Journal of Psychology, 39, 173-192.
(1998) Famous artefacts: Spearman's Hypothesis. Cahiers de
Psychologie Cognitive / Current Psychology of
Cognition, 16, 665-698 (Target Article).
In a number of publications, Jensen has recalled Spearman's
(1927, p.379) observation that the loadings of the first principal
component (PC1) of various 'intelligence tests' tend to correlate
positively with the corresponding Black/White mean
differences ('Spearman's Hypothesis'). Jensen believes this sheds
light on the true nature of g, Level II Ability, test bias, and
Black/White differences. His claims have been warmly welcomed
in some quarters (most recently by Herrnstein and Murray, 1994)
as conclusive confirmation of the Black inferiority myth.
Here it is shown by way of empirical, numerical, geometric, and
algebraic demonstrations tht the positive correlations predicted
by Spearman's hypothesis are psychometric artefacts which also
(a) with measures which have nothing to do with 'general ability',
for example, the number of toys and books a
child has, and, more generally
(b) with any set of moderately correlated random data, once the
sample is split into high and low groups.
Specifically, this interpretation predicts that, if sample sizes
differe substantially, then the correlation will be larger for
the PC1 of
the larger group. This prediction is borne out both in simulated
and in 'real' data sets, including Jensen's.
(1998) The rise and fall of Spearman's Hypothesis. Cahiers
de Psychologie Cognitive, 16, 788-812 (Response).
(1999) The amazing grace of common sense. Cahiers de Psychologie Cognitive/ Current Psychology of Cognition, 18, 232-240.
(2001) Better never than late: Peer review and the preservation of prejudice. Ethical Human Sciences and Services, 3, 7-21.
(2002) Some new results on Spearman's Hypothesis. Chinese Journal of Psychology, 44, 137-150.
(2003) Schonemann and Jaccard, J. Expected mean squares for extended twin designs. A projection approach. Abstract. Chinese Journal of Psychology, 45, 4, 337-343.
(2005) Psychometrics of Intelligence. K. Kemp-Leonard (ed.) Encyclopedia of Social Measurement, 3, 193-201. Abstract.
(2008) Erratum. Some algebraic relations between involutions, convolutions and correlations, with applications to holographic memories. Biological Cybernetics, 98, 355.
(2008) Schonemann and Scargle, J. D. A generalized publication bias model. Chinese Journal of Psychology, 50, 21-29. Abstract. Editorial Comments and Reviews.
(2008) Jerry Hirsch 1922-2008. Scientist, Rebel. Institute for the Study of Scientific Racism.
(2009) Schonemann and Heene, M. Predictive validities: figures of merit or veils of deception? Psychology Science Quarterly, 51, 195-215. Abstract.