Efforts of the Department of Education (ED) to simplify the Pell Grant formula by reducing the number of data elements used to calculate awards (i.e., data element reduction) are evaluated. A framework is developed to assess the critical characteristics of individual data elements, to eliminate elements from the formula, and to develop proposals for data element reduction. Individual data items used in the Pell eligibility and award formulae are evaluated on the basis of five measures: budget impact, aggregate distributional impact, sensitivity, reliability, and verifiability. Included is a comparison of two simulations of a reduction in the number of data elements used in the Pell eligibility and award formulae. The two simulations, the standard and the error free simulations, are identical except for the data base used. Both simulations eliminate all but five data elements (adjusted gross income, federal taxes paid, nontaxable income, number in household, and number in postsecondary education). The applicant-based model and data base, and techniques used to adjust the ED applicant data base for the error patterns found in the Pell Stage III data, are described in appendixes. References and a substantial series of data tables are also appended. (SW)