Multiple baseline designs: the use of a single-case experimental design in literacy research

Reading Improvement, Winter, 2004 by Richael Barger-Anderson, Joseph W. Domaracki, Nedra Kearney-Vakulick, Richard M. Kubina, Jr.

Single-case experimental designs are becoming more popular and acceptable ways to conduct classroom-based research. Single-case designs can also be beneficial when conducting literacy research. There are a variety of single-case research designs that can accommodate a range of instructional strategies and research questions. This paper provides examples and reasons for using the multiple baseline with literacy research. The effectiveness of multiple baseline designs are also examined in relation to the following experimental design issues: control, replication, causal relationships, internal and external validity and ethical concerns.

**********

Single-case experimental research designs are becoming more popular and acceptable ways to conduct classroom-based research (Birnbrauer, Peterson, Solnick, 1974: Gay & Airasian, 2000: Neuman & McCormick, 1995: Richards, Taylor, Ramasamy, & Richards, 1999). Since the mid sixties, single-case research designs have continued to increase in popularity (Gay, 1987). There is a need to increase awareness about the qualities and attributes of single-case research designs because many misconceptions still exist (Dermer & Hoch, 1999).

Single-case, or single-subject, experimental research design can personalize the data collection process because data is collected on each subject, or participant, and is individually analyzed. The term single-subject is not used because there is only one participant; rather, it refers to the procedure for data collection and the focus of the study as opposed to the number of participants. (Neuman & McCormick, 1995). Additionally, experimental control is established with each participant (Cooper, Heron, & Heward, 1987: Johnston & Pennypacker, 1993).

Wolery and Gast (2000) suggest teachers do not always have access to a large number participants to use group designs therefore single-case designs offer an excellent opportunity for teachers to conduct research in their classroom. Gay and Airasian (2000) agree that single-subject designs may be beneficial for use in the classroom due to the limited size of accessible population. For example, in special education the numbers of students in self-contained or resource rooms are generally small making single-case designs useful. "Approximately one third of all data-based interventions conducted on students with learning disabilities use single-subject designs" (Swanson, Hoskyn, Sachse-Lee, & O'Shaughnessy, as cited in Swanson & Sachse-gee, 2000, p. 144).

A meta-analysis of eighty-five single-subject research design studies involving specific instructional skill domains had reading as a focus (Swanson & Sachse-Lee, 2000). In fact, "most of the intervention studies conducted were in the domain of reading" (p. 149). When examining the meta-analysis Swanson and Sachse-Lee report that a multiple baseline design was used in 62% of the studies. Other findings were that 34% of the studies occurred in a resource classroom and 99% of the materials used were existing materials in the classroom. The meta-analysis indicates that single-case research designs help teachers and researchers examine variables that effect student learning.

There are several types of designs to choose from when utilizing single-case research. The most familiar types include the withdrawal design, reversal design, multiple-baseline designs, and the alternating treatments design (Gay, 1987; Gay & Airasian, 2000: Neuman & McCormick, 1995: Richards et al., 1999). Of these options the multiple baseline design offers a considerable degree of flexibility in classroom research (Swanson & Sachse-Lee, 2000).

Multiple Baseline Designs

Multiple baseline designs provide a means for collecting multiple sets of data in a single-case experimental design (Neuman & McCormick, 1995). Kucera and Axelrod (1995) state multiple baseline designs are "particularly well-suited to literacy research" (p. 47). The authors also maintain that multiple baseline designs can help examine new techniques and strategies that are used and found to be effective in the area of teaching reading. The multiple baseline design is the design of choice when it is not possible for subjects to return to original baseline (Hersen & Barlow, 1984; Gay 1987; Gay & Airasian, 2000; McReynolds & Kearns, 1983). For example, once a student has learned a new strategy for decoding words is not desirable, and in many cases possible, to have the student unlearn the new skill.

There are three types of multiple baseline designs. These include the multiple baseline across behaviors, subjects or participants, and settings (Cooper, Heron, & Heward, 1987; Gay, 1987; Gay & Airasian, 2000: Hersen & Barlow. 1984; Kazdin & Kopel, 1975; McReynold & Kearns, 1983: Neuman & McCormick, 1995). Alberto and Troutman (2003) provide the following examples for potential behaviors of interest to classroom teacher: a multiple baseline across behaviors can study talking out and out of seat behavior. A multiple baseline across subjects could target spelling accuracy for multiple students. And a multiple baseline across settings might examine a student's inappropriate behavior of swearing in recess and the cafeteria. In literacy research, any almost any question involving a dependent variable and independent variable can be asked provided the answer can be obtained by a few participants. For the purposes of illustrating a technical application, the multiple baseline across participants design will be used as example.