Evaluate sensitivity to binary discretisation in Orbit Theory

Assess systematically how the binary discretisation of categorical variables affects the orbit structure and conclusions of Orbit Theory analyses of multivariate longitudinal mortality data.

Background

Orbit Theory requires variables to be encoded as binary fit/unfit states. This encoding collapses distinctions within categories, such as the separate maternal-age groups 25–29 and 30–34, and may therefore influence the resulting state-space geometry, transition structure, and identified clusters.

The paper explicitly notes that the sensitivity of the orbit structure to this discretisation has not been systematically assessed, leaving unresolved how robust the reported structural findings are to alternative codings.