MULTIELEMENT DESIGN-MODEL FOR COMPONENT ANALYSIS AND CROSS-SETTING ASSESSMENT OF A TREATMENT PACKAGE

A special school''s natural treatment package for an autistic child''s self-stimulatory behavior was studied. The child''s school day was viewed as a multielement baseline with 4 alternating elements, classroom, play area, small room and large room, each with a different procedure. Conditions were held constant in the classroom and the play area, the normal school settings, so that manipulations for this 1 child did not disrupt the rest of the class. Components of the treatment package were manipulated in 2 experimental settings designated as small and large rooms. The design allowed reversal comparisons of procedures (experimental rooms) over time and daily comparisons of procedures across fall 4 settings (multielement baseline). The behavioral effects were comparable in natural and experimental settings when corresponding procedures were in effect. Daily comparisons of the multielement baseline seemed most desirable when the applied problem required comparison of several procedures and settings, e.g., component and cross-setting analyses), and as a result, when the possibility of time-correlated artifacts was greatest.