Package: incomeLato 0.1.1

incomeLato: Relative Income Position for 'Latinobarometro' Respondents from a Frozen Imputation Model

Applies a frozen multinomial model (version 'v1.5') that predicts each respondent's relative income position within their country and survey year -- the probabilities of belonging to the bottom 50 percent, the middle 40 percent and the top 10 percent -- to 'Latinobarometro' microdata that the user downloads from the official source. The model was trained on national household surveys and validated against 'LAPOP AmericasBarometer' data for 15 Latin American countries. It uses five predictors (sex, age, education, car ownership, washing machine ownership) plus a country effect, and the package harmonizes them across waves 1995-2024 through a manually curated crosswalk. The package ships only model parameters and variable-name metadata: it contains no survey microdata and no statistics computed from it. Probabilities are meant as continuous weights for aggregate analysis, not for classifying individuals.

Authors:Pedro Barbosa [aut, cre, cph]