“…It is recognized that examining the significance of interactions through multiple regression is problematic given the large number of biological risk factors being evaluated and the relatively small sample size. Therefore, the interaction effect was computed using a composite rating of overall "biological risk" that was based on a system previously developed by Hommeyer, Holmbeck, Wills, and Coers (1999). The biological risk composite calculated for the present study included the following parameters: 1) lesion level (1=sacral, 2=lumbar, 3=thoracic), 2) number of shunt revisions, 3) presence of seizures (0=no, 1=yes), 4) presence of shunt infection (0=no, 1=yes), 5) presence of Arnold-Chiari II malformation (0=no, 1=yes), and 6) number of oculomotor deficits recorded from the child's medical chart.…”