A model for constrained computerized adaptive testing is proposed in w
hich the information in the test at the trait level (theta) estimate i
s maximized subject to a number of possible constraints on the content
of the test. At each item-selection step, a full test is assembled to
have maximum information at the current theta estimate, fixing the it
ems already administered. Then the item with maximum information is se
lected. All test assembly is optimal because a linear programming (LP)
model is used that automatically updates to allow for the attributes
of the items already administered and the new value of the theta estim
ator. The LP model also guarantees that each adaptive test always meet
s the entire set of constraints. A simulation study using a bank of 75
3 items from the Law School Admission Test showed that the theta estim
ator for adaptive tests of realistic lengths did not suffer any loss o
f efficiency from the presence of 431 constraints on the item selectio
n process.