Absconding from Approved Schools as Learned Behavior

Abstract

The frequency distribution of absconding from approved schools is not a chance distribution. However, it has not proved possible to identify absconders on the basis of their personal char acteristics. This paper investigates whether the greater liability of some boys to abscond could reflect learning. As anticipated, the frequency distribution of abscondings appeared to be a nega tive binomial distribution. However, a learning model is not the only kind which could lead to this distribution and we report methods for determining the process which produced it in this case. Contrary to expectation, there was evidence of unequal initial absconding tendency for our subjects, but no evidence of learning. The statistical methodology may have important appli cations in criminology.