b. definition of appropriate educational skills and sequences in social and cognitive development, informed by normal developmental literature;

c. measurement of the effects of the interactions between family variables (e.g., family structure, family supports, socioeconomic status), child factors (such as degree of language impairment), and responses to educational interventions (including family-centered, parent training, and other approaches) on outcomes.

d. longitudinal treatment studies, where feasible, built on a clinical model with randomly assigned samples of sufficient size to assess the effectiveness of differing modes of treatment.


Treatment studies should recognize the common components of many comprehensive programs (e.g., standardized curriculum, family training, presence of typically developing peers) and should target and measure, longitudinally when feasible, “active ingredients” and mediating variables that influence the effects of intervention (e.g., communication and interaction opportunities for engagement, levels of interaction and initiation, specific teaching techniques, proportion of time in close proximity of peers). The concomitant development of innovative treatments building on these “active ingredients” should be supported.


In response to amendments in IDEA to make education more outcome oriented, a federal initiative should solicit and fund studies in the following areas, not easily supported under the current review system:

a. the development of instruments for measurement of diagnosis and critical aspects of development, particularly tools for early screening of autistic spectrum disorders and for measurement of response to interventions;

b. the development and application of sophisticated statistical methods of analysis of change and growth, particularly multivariate designs and those applicable to small samples; and

c. the development and dissemination of novel research designs that combine individual and group approaches in ways that minimize biases and maximize the power of small samples.


Competitively funded initiatives in early education in autistic spectrum disorders should require plans and contain sufficient funding for short- and long-term assessment of child outcomes and measures of program efficacy.

The National Academies of Sciences, Engineering, and Medicine
500 Fifth St. N.W. | Washington, D.C. 20001

Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement