The Stanford-Binet Scales
For many years, the Binet scales were the preferred
tests. They underwent many revisions after Binet’s
work in 1905. Terman’s revision in 1916 was fol-
lowed by the 1937 Revised Stanford-Binet
(Terman & Merrill, 1937). The 1960 version of the
Stanford-Binet (Terman & Merrill, 1960) gave way
to a 1972 test kit with revised norms, followed by
the fourth edition of the scales published in 1986
(Thorndike, Hagen, & Sattler, 1986). The most
recent revision of the scale appeared in 2003, the
Stanford-Binet Fifth Edition, or SB-5 (Roid, 2003).
Description. Like the 1986 version, the SB-5 is
based on a hierarchical model of intelligence. Spe-
cifically, the Stanford-Binet Fifth Edition (SB-5)
assesses five general cognitive factors, and each fac-
tor is tapped by both verbal and non-verbal subtest
activities (Roid & Pomplun, 2005):
- Fluid reasoninginvolves the ability to solve new
problems and is measured by the following
subtests; - Quantitative reasoninginvolves the ability to
solve numerical and word problems as well as
to understand fundamental number concepts; - Visual-spatial processinginvolves the ability to see
relationships among objects, to recognize spa-
tial orientation, and to conduct pattern analysis; - Working memoryinvolves the ability to process
and hold both verbal and non-verbal informa-
tion and then to interpret it; - Knowledgeinvolves the ability to absorb general
information that is accumulated over time
through experience at home, school, work, or
the environment in general.
Each subtest is composed of items at varying
levels of difficulty, from age 2 to adulthood. This
revision uses an adaptive testing procedure called
multistage testing. The examiner first gives the
two routing subtests, the Object-Series Matrices
subtest and the Vocabulary subtest. These are called
routing subtests because the examinee’s perfor-
mance on these two subtests, a non-verbal and a
verbal subtest, respectively, help to determine the
entry point (i.e., which item to start with) for each
remaining subtest. This initial estimate of ability
provides a more appropriate entry or starting
point on subsequent subtests, and it is likely to
result in more efficient testing than relying exclu-
sively on chronological age as a guide for a starting
point. Thus, not all examinees of the same age are
given the same items. Other unique features of the
SB-5 include the infrequent use of time limits on
items, no time bonus points are possible, and the
use of teaching items (practice items) as new tasks
are introduced (Roid & Pomplun, 2005).Standardization. Final standardization of the SB-5
included 4,800 participants, ages 2–96 years. Using
2001 U.S. Census figures, participants were sampled
according to geographic region, community size,
ethnic group, age group, and gender. Socioeconomic
status was considered as well. In addition, the SB-5
was administered to approximately 1,400 individuals
from special populations (e.g., those with mental
retardation, learning disabilities, attention deficit dis-
order, or speech/hearing impairments) to ensure the
clinical utility of the scores.Reliability and Validity. Psychometric analyses
support the reliability and the validity of SB-5
scores (Roid & Pomplun, 2005). Internal consis-
tency reliabilities ranged from .95 to .98 for IQ
scores, and from .90 to .92 for the index scores of
the five factors. Test–retest reliabilities across all age
groups were generally high, in the .80s for factor
scores and in the .90s for IQ scores.
As for the validity of SB-5 composite scores, a
variety of supportive evidence has been obtained
(Roid & Pomplun, 2005). For example, the correla-
tion between SB-5 IQ scores and scores on the 1986
Stanford-Binet Fourth Edition (SB-4) was .90. Cor-
relations between the SB-5 and several Wechsler
scales (described next) were .84 for children and
.82 for adults, and correlations between SB-5 scores
and scores on achievement tests were substantial as
well. Therefore, using the criterion of relationship to
other accepted tests of intelligence and achievement,
the validity of the new Stanford-Binet is supported.
Finally, participants with learning disabilities, mental206 CHAPTER 7