This study compares test battery results through carefully planned analysis of actual Title I evaluation data. The goal of the study is to derive practical recommendations regarding the results of commonly used test batteries which would be helpful at the local level. A criticism of the comparability assumption underlying the Title I evaluation and reporting system (TIERS) is not intended, rather, comparability within the current system with hopes of deriving useful recommendations is investigated. Six data sets from four Midwestern states from 1979 and 1980 were used. Data from reading and math programs, grades two through six, were included. Finding of this study include: (1) very few test batteries are actually given to 80% or more of the students; (2) normal curve equivalents from the Stanford Diagnostic Reading Test 76 and the Stanford Diagnostic Math Test 76 are often erroneously high because of incorrect conversion of Spring test scores; and (3) the 1978 edition of the Gates Mac Gintil Test appears to yield larger gains than the 1964-65 edition. (Author/GK)