This study employed a single-subject, multiple-baseline across groups of participants design. In a multiple baseline design there are three or more concurrent baselines and subsequent sequential staggered introductions of the independent variable across baselines. Thus, each tier of participants is introduced to the independent variable in staggered fashion. Each participant, or in this case each group of participants acts as their own control through the manipulation of the independent variable across the tiers. (Gast and Ledford, 2010). While data were collected for each individual student, the unit of analysis focused on the groups’ performance (Kratochwill et al., 2010). Thus, the results were presented both as individual data and group averages. A multiple-baseline design is an appropriate tool for measuring the behavior of a group of participants while still tracking and evaluating individual behaviors.
Due to the use of multiple-choice selection to measure student’s behaviors, high variability in baseline was expected. Since stability of baseline commonly used to guide the decision about entering the treatment phase could not be established, regulated randomization procedures were used to determine when each tier would be introduced to the independent variable. Regulated randomization involves two phases of randomization, namely (1) the random assignment of tiers to the order in which they will enter the treatment phase of the intervention, and (2) the random assignment of the points at which intervention is to be introduced to each tier (Koehler and Levin, 1998). Randomly assigning entrance into treatment can offset validity threats such as history and maturation (Kratochwill et al., 2010). Additionally, Gast (2010) stated that randomized entrance into the treatment phase may be considered if it is anticipated that participants in early tiers may take so long to reach stabilization that later tier participants spend an exceedingly long time in baseline. Entrance points to the treatment phases in the current study were randomly selected from two possible entry points for each tier. Entrance into the treatment phase for the first tier of letter-sound correspondence was randomly selected from one of two possible sessions (Session 6 or 7). Subsequent tiers were likewise randomly selected from one of two possible sessions as follows: tier two (Session 9 or 10), tier three (Session 12 or 13), and tier four (Session 14 or 15) respectively.