In a study of the usefulness of the Rasch model for examining tests for possible bias, 102 native Spanish-speaking and 104 native English-speaking preschool four-year-olds in a remedial education program were administered Spanish and English versions of the Cooperative Preschool Inventory, a standardized measure of school readiness, The Rasch model of analysis was applied to the verbal and motor scales of each version. Results indicated that eight items that fit the model appeared to be improperly functioning items because on four items English-speaking pupils had an advantage over Spanish-speaking pupils, and on four, the advantage was reversed. Several discrepancies were found in the item translations and in the administration and scoring directions of the Spanish and English versions, including more complete examiner information on the English version in the form of correct responses, suggested probes, and possible answers from the examinees. In addition, the directions associated with each item in the Spanish version are given in English, requiring the examiner to translate them into Spanish before directing them to the examinee, and some of the English-to-Spanish translations allow for the change of verb tenses. (MSE)