This study examined achievement differences between algebra students taught with the non-traditional textbook developed by Saxon and those taught with a "traditional" textbook by Dolciani. Student absences, rate of turning in homework, and ability level were considered, as well as teachers' comments. One Algebra I section in each of seven schools used the Saxon text, while a second section in each school used the Dolciani text, with the same teacher teaching both sections. In an eighth school, the texts were used with two sections of Elementary Algebra students. (However, data from these two classes and from one of the seven Algebra I teachers were not used in the analysis.) An analysis-of-covariance design accounted for students' initial achievement level differences prior to entering Algebra I. The Spring, 1981, California Achievement Test total math score was used as a covariate. The locally-constructed Algebra I Comprehensive Exam assessed course achievement. The mean score of the 98 students using the Saxon text was significantly higher than the mean score of the 67 students using the Dolciani text. The Saxon classes had slightly more absences and turned in homework slightly less frequently than the Dolciani classes. Most teachers preferred the Saxon text. An addendum discusses inter-test correlation and data from the California Achievement Test. (MNS)