B. Brocke et al., THE INTELLIGENCE-STRUCTURE-TEST (I-S-T-70 ) - ANALYZING ITS THEORETICAL BASIS AND PSYCHOMETRIC PROPERTIES, Diagnostica, 44(2), 1998, pp. 84-99
The construction of well-developed and state-of-the-art intelligence t
ests should be based on models of intelligence. The development of Xmt
hauer's I-S-T 70 had no systematic theoretical foundation and relation
to available models. instead, the selection of tasks was primarily gu
ided by intuition and plausibility: The question arises as to which co
mponents and which more general aspects of intelligence are covered by
the I-S-T 70. nir will focus in particular on die internal structure
of the test. Against this background the construction and use of profi
les in the I-S-T 70 will also be discussed. In addition to these quest
ions concerning the validity of the test, which form the focus of this
analysis, we also present results on the most important psychometric
criteria and use them for an evaluation of this instrument. The sample
group consisted of 279 subjects from different occupations. The surve
y was carried out in the former East Germany in 1991, so that the resu
lts do not necessarily apply to the former West Germany. The results s
how chat the internal structure with nine scales postulated by Amthaue
r cannot be confirmed in multivariate analyses: rather, it appears to
be reduced to only two or three components. The usefulness of the I-S-
T 70 intelligence profiles is also questionable. Internet validity - c
alculated via correlation with the Figure Reasoning Test - is moderate
ly high. Structural analyses and a comparison with today's theories of
intelligence structure imply that a fundamental revision af the I-S-T
70 is necessary.