“…By randomly assigning children to groups at the beginning of the experiment, the need for complex statistical manipulations at the end of the experiment is largely eliminated. Even though many people have suggested that the use of experimental/control groups is impossible in early intervention research, such designs have been successfully used by many researchers (see, for example, Andrews, Blumenthal, Johnson, Kahn, Ferguson, Lasater, Malone, & Wallace, 1982;Gordon, 1969;Gray & Klaus, 1970;Ramey & Haskins, 1981;and Williams & Scarr, 1971). Particularly in those cases where the number of children in need of services far exceeds the capacity of the service agency to provide services (an almost ubiquitous occurrence, if one is to believe funding requests to state legislators and federal offices), or in those cases where alternative treatment programs are being considered (e.g., half-day vs. full-day programs), random assignment to treatment/no treatment groups or to alternative treatment groups is both feasible and advantageous.…”