The internal consistency of the test for Auditory Comprehension of Language (TACL), as it is used in the diagnostic assessment of the language comprehension of black, low socioeconomic preschool children, was investigated in this study. 30 three year old and 30 four year old black Headstart children from Pulaski County, Arkansas were given the TACL after performing satisfactorily on a language screening test. Each child was given the test at his or her Headstart center according to published instructions. Raw scores were collected and analyzed for the test according to the alpha coefficient methods. Results indicated that the test has only a moderate level of internal consistency for use with low socioeconomic, black preschool children, a level less than desirable for satisfactory language assessment. There is no explanation for this level of internal consistency, and some of the factors that may have contributed to the results include the possible sources of error which contribute to lack of reliability in tests: instrument error, subject error, and administrative error. Some possible sources of subject error include variation within subjects in guessing, state of motivation, and language differences. The content of the test as a source of error indicates the need for an item analysis of the TACL. This study also emphasizes the need for assessment devices which measure language performance according to normative data established for the group and data reflecting internal consistency within the structure of the tests. (Author/AM)