ProceduresThis investigation utilized a single subject, multiple probe research design (Tawney & Gast, 1984) across one set of three participants, all of whom used voice output systems to communicate. A multiple probe design was chosen instead of a multiple baseline design to help to minimize the boredom and fatigue that may have occurred with extended measures of the dependent variable at baseline (McReynolds & Kearns, 1983). The same design was utilized with an additional set of two participants, both of whom used non-electronic communication boards.