Modelling District Court decision-making: Offender identity, judicial attitudes and legal factors

The guiding orientation of the present research was whether two conflicting positions, (1) that sentencing in a criminal case is a complex decision-making process, and (2) that sentencing is a judgment based upon few offence and/or offender details, could be reconciled. The competing claims lend support to the opposing viewpoints on sentencing guidelines, and are explicitly or implicitly utilised by each side in the law and order' debate, but each is deficient in current empirical or theoretical investigation. Archival research was undertaken on assault occasioning bodily harm cases from the Queensland District Court. Study 1 (N = 244) investigated the correlations between 27 offence, offender, victim and sentencer characteristics (e.g., use of a weapon in the assault, offender's prior convictions, victim's gender and sentencer's expertise), five theorised preliminary decisions (e.g., whether to impose a tariff versus individualised penalty, to set a utilitarian versus retributive punishment), and five sentencing decisions (e.g., to record or not record a conviction, to impose a lenient or a harsh penalty). Results favoured the conclusion that sentencing is a complex decision-making process. Numerous correlations between the variables were initially indicated, although the effects on the sentencing decisions decreased when the offence seriousness and the offender's prior criminal history were controlled for. Guided by the results from Study 1, five increasingly complex legal and theoretical models were developed. The models were arranged hierarchically, initially including only legal variables and then expanded to incorporate concepts from Schubert's (1965;1974) Attitudinal Model of Supreme Court decision-making, and from Heise's (1988) Affect Control Theory. These were evaluated in Study 2 (N=393) using structural equation modelling. Model comparisons revealed that while the Simple Legal Model, (the first model presented including the harm caused by the offence, the offender's prior criminal history, and prescribed aggravating features of the attack) explained almost half of the variance in the sentencing decision, the model offered a poor fit to the data. While each of the three succeeding models improved the model fit, it was the fifth model, incorporating legal variables, preliminary decisions, attitudes, and particularly the sentencer's evaluation of the `character' of the offender, which best represented the data, without sacrificing parsimony, in addition to accounting for almost two-thirds of the variance in the sentencing decision (65.3%). It was concluded that sentencing is a complex decision-making process, and that the judgment is based upon few offence and/or offender details. Firstly, a relatively small number of offence and offender characteristics accounted for the greater part of the variance in the sentencing decision. Judges however, appeared to organise the information provided in order to make assessments about the offender, in the manner predicted by Affect Control Theory. The overall result therefore, supported the complex decision-making process subscribed to by judges. Limitations of the research and the implications of these results for the main players in the process, the inclusion of psychological theory in the courtroom, and the community's understanding of the system, are discussed.