PHONETIC, LINGUISTIC, AND SIGHT READERS PRODUCE SIMILAR LEARNING WITH

T h r e e years of computer-stored reading frequencies and celerations yielded average learning pictures for the 94 books in 17 different reading curricula. The median test with both Fisher's exact and with chi-square probabilities determined the significance of differences between the distributions of the average learning produced by the books in each curricular series. The pupils ranged from 5 to 21 years of age and included almost all exceptionalities. Major conclusions are: (a) T h e different curricula are functionally identical: they produce the same learning. (b) The average learning of these exceptional children was the same as that of regular children (x 1.2 per week). (c) Children can learn faster when starting at a low correct frequency (5 per minute), and when starting at high error frequencies (5 to 10 per minute). Strategies for improving learning include: (a) Teachers and children should treat errors as something to be aware of and eliminate, rather than as something to stay away from. (b) Curricula should start with low correct frequencies and high error frequencies, making room for steeper learning of both. Reading experts over the past 140 years have generally described reading i