Statistical methods for courts, corrections, policing, forensics and sociolegal research, computed in a compiled core with a thin R layer. Where an identity or bound has a machine-checked proof ('Lean'), the package names it. This development version holds the disparity tests: the veil-of-darkness contrast of Grogger and Ridgeway (2006) doi:10.1198/016214506000000168 , the outcome test and its infra-marginality problem, benchmark decomposition of resident disparities, dyadic offending rates, and the bounds a relative risk obeys given an odds ratio; and the first forensic methods: genotype match probabilities with the correction of Balding and Nichols (1994) doi:10.1016/0379-0738(94)90222-4 , the log-likelihood-ratio cost of Brummer and du Preez (2006) doi:10.1016/j.csl.2005.08.001 with its calibrated minimum, mixture likelihood ratios, kinship and paternity indices, database-search corrections, and Tippett and empirical-cross-entropy plots.
Author
Maintainer: Vansh Singh Ruhela vsruhela@proton.me (ORCID)
Authors:
Vansh Singh Ruhela vsruhela@proton.me (ORCID)