Word Prediction for Students with Severe Spelling Problems

Abstract

In Study 1, three students with severe spelling problems wrote daily in journals in their classroom, alternating among handwriting (HW), word processing (WP), and word prediction with speech synthesis (PR). No differences among conditions were found for legibility, and only one student consistently spelled more words correctly with PR. One explanation for the lack of impact was that the word prediction software was difficult for students to use and that the modest demands of journal writing did not take advantage of the large vocabulary and complex word prediction capabilities of the software. Study 2 was conducted with the same students and conditions but with a writing task that demanded a larger vocabulary. In this study, improvements in both legibility and spelling were noted for two of the three students. Students wrote two to three times slower with word prediction than with handwriting. Analyses of students' use of word prediction and their spelling errors suggested that the software was difficult to use because of substantial demands on attention and the need to spell the initial letters of words correctly.