“…Offenses were given the following seriousness codes: 1=runaway, truancy, curfew violation, ungovernable behavior, alcohol possession, and other status offenses; 2=other misdemeanors; 3=(all misdemeanors) drug law violation, weapons offense, driving intoxicated (third offense), and disturbing the peace; 4=(all misdemeanors) arson, criminal mischief, forgery, criminal trespass, theft with value less than $300, and unauthorized use of a vehicle; 5=other felonies; 6=(all felonies) drug law violation, weapons offense; 7=(all felonies) burglary, theft with value over $300, criminal mischief, arson, and forgery; and 8=murder, manslaughter, assault (all degrees), and sexual assault (all degrees). 8 Social scientists have voiced a concern regarding the possibility of bias in regression coefficients when the sample on which they are based is censored (see for example Berk, 1983). Clearly, the procedure used in the present research, including at the stage of final disposition only those adjudicated delinquent, presents this possibility.…”