Sl. Bishop et al., THE USE OF IQ ASSESSMENT IN ASTRONAUT SCREENING AND EVALUATION, Aviation, space, and environmental medicine, 67(12), 1996, pp. 1130-1137
Background: The screening and evaluation of astronaut candidates usual
ly includes measures of cognitive and intellectual capacity. Questions
of ceiling effects and discriminability when evaluating individuals a
ssumed to be of above average intelligence should be considered. The c
urrent study examines ceiling effects and discriminability on IQ asses
sment between pilots vs. mission specialists and men vs. women in the
selection of astronaut candidates. Methods: IQ scores on 104 male and
female Shuttle astronaut candidates as measured by the Multidimensiona
l Aptitude Battery (MAB) were analyzed via factor analyses and tests o
f differences between genders, selection status, and job types. Total
group, gender, and job type specificity and sensitivity and overall pe
rcent correctly classified were assessed on selected vs. unselected cl
assification. Results: Exploratory and confirmatory factor analyses fa
iled to replicate the proposed factor configuration prompting reconfig
ured domain scores. Analyses of differences between gender, selection,
and job categories found different results when using reconfigured IQ
scores compared to the original formulation. The 10 subscales perform
ed better than overall domain scores in sensitivity, specificity, and
as significant predictors of individuals selected. Both versions of Ve
rbal, Performance, and Full Scale IQ scores performed similarly regard
ing sensitivity and specificity. The reconfigured scores were better p
redictors of selection. Discussion: Evidence for sensitivity, specific
ity, and classification prediction was better for subscales than overa
ll domain scores. Some consideration to using reconfigured domain scor
es should be given to increase instrument utility.