Of 10 participating inner-city school teachers, five used a modified basal series, and five used the language experience approach for reading instruction. Monthly training sessions were held, and the Medley and Smith's Observation Scale and Rating-Reading was used to lend objectivity to the monthly observations of each teacher. The following instruments were administered to the pupils: the Otis-Lennon Mental Abilities Test in January, an adaptation of Eunice Askov's Primary Pupil Attitude Inventory in March, and the California Reading Test in May. Analysis of covariance was made with two criterion variables, reading achievement and attitude data. The independent and control variables were sex, intelligence, reading method, teacher experience and professional background, teaching style, and reading time in class. It was concluded that (1) for both achievement and attitude, no significant differences occurred between the two groups when classes were treated as units; (2) achievement differences favored the language experience groups when pupils were treated as units; (3) when Unifon classes were treated as a third group, no significant differences occurred between groups; and (4) when analysis of variance was used, findings showed that girls in the modified basal series were significantly superior in total reading and in IQ. Tables are included. (AW)