Software

LVM: R tools for coarsened data

LVM (working name) is an R package in development for analysing quantities observed through brackets or intervals, such as income recorded in survey categories. It translates my research on coarsening-induced measurement error into tools that help researchers connect these observations to the underlying quantities their theories describe.

The package brings together likelihood-based and Bayesian estimation, tools for reconstructing bracketed variables, and methods for assessing uncertainty and sensitivity to modelling assumptions. It supports applied analysis through model comparison and diagnostics, alongside simulation tools for designing and evaluating statistical methods. The aim is to make the assumptions involved in working with grouped data more transparent and their consequences easier to examine.

The software is under active development, with a public release planned alongside the related research. Researchers interested in using it or discussing applications are welcome to contact me.