Dataset constructed to illustrate the construction of a sentencing severity scale. The method is described in the paper
Wallace, S. and Francis, B. (2024)"Developing a Complete Sentence Severity Scale using Extended Goodman RC models", Journal of Quantitative Criminology. https://doi.org/10.1007/s10940-024-09591-6

The file is a CSV file with a header, with 61066 cases synthesised from a real sentencing dataset using the method described in Jackson, Mitra, Francis and Dale(2022) JRSS Series A (https://doi.org/10.1111/rssa.12876) .The offense codes are supplied in a separate PDF. The method of generating this synthetic file is described in a supplementary file linked to the paper.

An analysis file is also supplied, showing the application of the method in the paper to the synthetic dataset using R or RStudio.

There are five variables:
SENTENCE1 Categorised sentence (29 levels)
OFFENSE Categorised offense categories (111 levels)
PLEA_TYPE Plea made at court (guilty/not guilty)
PREV_CONV Previous court appearances (yes/no)
NUM_OFF Number of offenses at current court appearance (single/multiple)"