Many children in early intervention programs require the services of a specialist, such as an occupational therapist, a physical therapist, a speech-language pathologist, or an early childhood special educator. Regardless of the specific type of classroom-based setting (self-contained, inclusive, reverse-mainstreamed, etc.), children often receive services from professionals other than the regular teacher. In the past 5 years, the literature has increasingly suggested that the provision of these services should follow certain principles:
1. Therapy and instruction should occur in the child's classroom (e.g., Giangreco, York, & Rainforth, 1989).
2. Other children should be present during therapy and instruction (e.g., Wilcox, Kouri, & Caswell, 1991).
3. Therapy and instruction should be provided as part of ongoing classroom routines and activities (e.g., Miller, 1989).
4. The child's initiations should be used as the primary context for therapy and instruction (e.g., McDonnell & Hardman, 1988; Warren & Kaiser, 1988).
5. Therapy and instruction goals should address behaviors that are immediately useful for the child (e.g., McDonnell & Hardman, 1988; Norris & Hoffman, 1990).
6. A primary role of the external therapist or special educator is as consultant to the child's regular teacher (e.g., Johnson, Pugach, & Hammitte, 1988).
These approaches, together, define "integrated services." The movement toward an integrated and classroom-based model of special services raises the question about the extent to which such services are actually implemented and the factors influencing this decision. Multiple factors will likely determine whether the movement toward integrated services will become a widespread reality.
First, the value orientations of the field, the different disciplines, and individual programs and practitioners are likely to influence this movement. In special education, for example, normalization, generalization, and team collaboration have been predominant themes. The issue of normalization of services is particularly pertinent at a time when the differences and similarities between early childhood education (ECE) and early childhood special education (ECSE) are being debated (Carta, Schwartz, Atwater, & McConnell, 1991; Johnson & Johnson, 1992).
A second factor influencing the extent to which integrated therapy is likely to be used is research on the effectiveness of this model, which suggests that, at the very least, it is no less effective than isolated models. Using a reversal, single-subject design, Giangreco (1986) compared isolated therapy and integrated therapy for instruction in the use of a microswitch. The subject was a 13-year-old girl with multiple disabilities. More trials were correct with integrated therapy, as confirmed by depressed performance during return to baseline of isolated therapy. Cole, Harris, Eland, and Mills (1989) compared in-class and out-of-class direct services for physical and occupational therapy. Both individual and small-group therapy were provided to 61 pre-schoolers with disabilities; services differed only in terms of location. No significant differences were found on standardized measures of motor development, but classroom staff preferred the in-class model. In a language study, 20 children age 20-47 months were randomly assigned to either an indirect therapy condition or an individual pullout condition (Wilcox et al., 1991). Children in both groups received an "interactive modeling" procedure. It was found that their use of target words was the same in both conditions, but the classroom (indirect therapy) group generalized to home settings better. The researchers also found that cognitive ability predicted 46% of the variance in generalized word use for the classroom treatment group, which suggests support for others' findings of aptitude-by-treatment effects (Cole, Dale, & Mills, 1991; Yoder, Kaiser, & Alpert, 1991). …