ABSTRACT
All variables that had significant associations with minimal brain dysfunction (MBD) symptoms were examined in combination. For each MBD symptom—learning difficulties, hyperkinetic-impulsive behavior, and neurological “soft signs”—all the variables from the five epochs were entered simultaneously into stepwise discriminant function analyses. Although the antecedent variables identified were not efficient predictors of MBD symptoms, moderately increased risks of the magnitude reported can be of considerable practical importance. In spite of their poor predictive power, the discriminant functions are valuable in the identification of variables associated with significantly increased risks of MBD symptoms. The possible mechanism was essentially ruled out by the current study: smoking is related to low birthweight, which in turn is responsible for the MBD symptoms. Rather than the usual comparison group of children with no MBD symptoms, the affected groups were compared with all other children, so the probabilities obtained could be contrasted with the expected 8% values for the entire MBD cohort.
