This report summarizes the development and validation of a test of literal comprehension based on a modified cloze technique. This modification of the cloze offers an objective and partially computerized procedure for constructing items in the multiple-choice format. A field test of the multiple-choice cloze (MCC) involved the administration of 36 test forms across three levels to 5,722 students in grades 1-9. A second measure of literal comprehension based on Bormuth's wh-item was adminstered for comparison. An examination of descriptive data indicates that the MCC and wh-item tests are generally consistent both within themselves and across the two test types. Both tests also proved to be highly reliable with a KR-20 range of .94 to .97 for the MCC, and .85 to .96 for the wh-item test. The construct validity of the MCC was supported by generally appropriate correlations with the wh-item, I.Q., California Achievement Test subscores, and a measure of passage independence. With the exception of one level, the factor analyses generated a first factor interpreted as literal comprehension and defined by the cloze, wh-item, and appropriate CAT subscores. A minor second factor included high loadings on I.Q. and inferential subscores of the CAT. (Author)