Variations of Multiple Baseline Designs and Combination Designs

This chapter describes two variations of the multiple baseline design: changing criterion design, and the "delayed" or "non-concurrent" multiple baseline design. The changing criterion design and non-concurrent multiple baseline design require the attention and caution before using. Though these two designs have some practical advantages, they have serious internal-validity limitations and may be best used in combination with an A-B-A-B design, multiple baseline design, or multiple probe design. The design requires initial baseline observations on a single target behavior. Unlike multiple baseline and multiple probe designs across behaviors, the changing criterion design has the advantage of requiring only one target behavior. The nonconcurrent multiple baseline design requires that the same independent variable be implemented across tiers, that the same dependent variable be repeatedly measured, and that each subsequent tier&s;s baseline condition be longer than preceding tiers. The intervention condition was evaluated within the context of a multiple baseline design across behaviors and participants.