I__. - International Military Testing Association

tests in a paper-and-pencil mode; these data were used to obtain paper-and-pencil
calibration information. Each of these calibrations was used to estimate the ability of
examinees assigned to the third group, who took the tests on computer. Lengthy test
administration time required splitting the study into two phases. General Science,
Arithmetic Reasoning, Word Knowledge, and Shop Information were addressed in the
first phase. Results from this phase showed: (a) no practical differences in the
estimation of abilities; (b) small, but statistically significant differences in different
tests; and, (c) no significant differences in test reliabilities. The second phase
involved the administration of the Paragraph Comprehension test. Data for this
second phase have been collected and anaIyses are underway.
Cross-Correlation. The purpose of this research was to compare the
measurement precision of CAT-ASVAB and P&P-ASVAB. Subjects were from the ‘Navy
Recruit Training Center, San Diego. Each recruit had taken an operational form of
P&P-ASVAB which was used for enlistment purposes. The total sample was split into
two groups. The first group took CAT-ASVAB Form 1. then CAT-ASVAB Form 2. The
second group took P&P-ASVAB Form 9B, then P&P-ASVAB Form 10B. The second test
for each group was administered about five weeks after the first test. Results indicate
that, despite using substantially fewer items, CAT-ASVAB exhibits significantly
higher alternate form reliability than P&P-ASVAB for most tests, while no P&P-
ASVAB test demonstrates significantly higher reliability than the comparable CAT-
ASVAB test.
Preliminarv Ooerational Check. The purpose of this research was to
demonstrate the communications interface between the ACAP System and USMEPCOM
computer system. The testing procedures were performed jointly by NPRDC and
USMEPCOM personnel at the Seattle MEPS. Data from examinees were loaded onto the
Data Handling Computer at the MEPS, then transferred to the USMEPCOM System-80
minicomputer. Comparison of the data before and after the transfer showed the
procedure was completed with perfect accuracy.
Sco e Eauat’ p Development. The purpose of this research was to equate CAT-
ASVAB with P&P-:SVAB. Equating is essential to insure that the two forms of the
battery are on the same metric, and that the scores are interchangeable. Subjects
were applicants for enlistment at six MEPS (San Diego, Richmond, Seattle, Boston,
Omaha, and Jackson) and their satellite MET sites. These six MEPZYMETS complexes
were selected because, in the aggregate, their applicants are representative of the
nation. The operational measures included P&P-ASVAB Forms lOA, lOB, 11 A, llB, 13A,
and 13B. There were two forms of the CAT-ASVAB (both non-operational). Finally,
P&P-ASVAB Form 8A was used as the non-operational reference battery. Subjects
were randomly assigned to one of three groups. The first group took CAT-ASVAB
Form 1, then the operational P&P-ASVAB. The second group took CAT-ASVAB Form 2,
then the operational P&P-ASVAB. The last group took the reference battery (P&P-
ASVAB Form 8A), then the operational P&P-ASVAB. In each case, the testing was
done on the same day or on successive days. Data collection, editing, and equating
analyses have been completed. New equating procedures have been developed and
applied. Analyses indicated that composite equatings were unnecessary. ‘Provisional
equating tables for operational use in
the subsequent Score Equating Verification
study were developed. The ACAP microcomputer delivery system has performed
satisfactorily, exhibiting fewer problems than anticipated. Finally, the logistics of
testing in the numerous, heterogeneous MEPS/MET sites nationwide has presented no
insurmountable problems.
247