The topic of this manuscript is both timely and of great importance. As the utility of the behavioral approach with respect to treatment of autism continues to receive empirical support as well as popular support, issues of costยฑbeneยฎt and early intervention become even more important. The central topic of this manuscript, that is the predictive ability of initial learning for outcome success, is a hotly debated issue and one for which it would be reasonable to conclude that such information will be used both for the positive and for the negative with respect to service delivery. Thus, the standards for such an article need to be high, as the author correctly points out that this study is best described as an initial description of clinical data.
The author correctly points out that this is not a controlled study. We are not given information as to how the particular learning task items were adapted for each individual child. The description of the basic methodology and procedures is clear, but certainly a hallmark of applied behavior analysis is the individualization based on the child's response pattern. Further, there are basic child characteristics for which we do not have information. An example is the range and intensity of items and events that served as reinforcers. Thus, could these dierences be explained if the children who did better on initial learning were also the ones who had the strongest preference and range of potential reinforcers? This argument is similar to the author's issue of IQ. That is, we have a relationship between this early learning and outcome for a small group of children, but we do not know whether there is some third factor that is inยฏuencing the initial early learning rate. Thus, for these reasons and those the author states in the manuscript, conclusions from this study must be drawn with great care and with expression of appropriate caution, which I believe the author is doing.
Along these lines of interpretation, I would also suggest that the introduction needs some work with respect to the critical analysis of the various outcome studies cited. There are, in fact, very few, and, of those mentioned in the ยฎrst CCC 1072ยฑ0847/99/010035ยฑ02$17.50