IV. Issues in grouping items from the Neonatal Behavioral Assessment Scale.

The various analyses of the NBAS that we have described do not permit any simple understanding of the significance of the scales. We have shown that other variables associated with the scale may have more significance than the scales themselves. Predominant state appears to be one of these variables. When the NBAS scores are correlated with behavior only 4 months later, only tiny relationships are evident, and, what is more, many of these relationships do not make obvious sense. In recent times there has been growing evidence that infant behavior shows little consistency or continuity over time (McCall, Eichorn, & Hogarty 1977). These data from the Brazelton NBAS seem to conform that notion. The message we want to end with is that we have still not reached the point where we can adequately understand the behavior of the human newborn. Simple relations between antecedent and consequent behaviors assessed by simple statistical procedures will not solve our problem. Bridging the gap between our global theorizing about newborn behavior and concrete deomonstrations of those theories will require a complex interplay of common sense and statistical sophistication.