To explore the appropriateness of the Rasch model for the vertical equating of a multi-level, multi-form achievement test series, both the Rasch model and the traditional Thurstone procedures were applied to the Listening Comprehension subtest scores of the Stanford Achievement Test. Two adjacent levels of these tests were administered in 1981 to a national student sample as part of the National Standardization Program. Pearson product-moment correlations and the conditional p-value procedure (Rentz) were used to further analyze the resulting four scaled scores. These analyses showed the consistency of both the Rasch and Thurstone models when applied to a vertical equating problem. Groups equal in ability were identified from equivalent scores derived through either equating method. There were, however, different patterns to the standardized differences between Thurstone scaled scores and between Rasch scaled scores obtained on the Primary One and Primary Three levels by fourth graders. There was a consistent pattern of departure found in the Thurstone data but not in the Rasch scaled scores. Both approaches have pitfalls in their application to vertical equating problems, but the Rasch model compared well to the traditional model. This has encouraging implications for computerized adaptive testing and customized test development and scoring. (BS)