hLM

hlm most commonly refers to hierarchical linear modeling, a class of statistical models designed for data that are organized at more than one level. In a typical application, measurements are clustered within higher‑level units, such as students within classrooms or patients within clinics. The model separates variation into fixed effects, which are assumed to be the same across all groups, and random effects, which allow group‑specific deviations. This structure enables analysts to estimate how predictors operate at both levels and to quantify how much of the outcome’s variance lies within and between groups.

The framework extends to hierarchical generalized linear models (HGLMs) to handle non‑normal outcomes such as binary or count data. More complex forms include multiple membership, cross‑classified, and growth models, which can accommodate situations where units belong to multiple higher‑level groups or change over time. Estimation methods typically rely on maximum likelihood, restricted maximum likelihood, or Bayesian approaches, with software implementations that have become widely accessible.

Common applications span education, psychology, sociology, and public health, wherever data exhibit nesting or repeated measures. Examples include assessing how classroom quality influences student achievement, or how patient outcomes vary by clinic while accounting for individual characteristics. While often associated with the capitalized term HLM, the lowercase hlm is used informally as an acronym; the broader concept is widely described as multilevel or hierarchical modeling.