A Vernacular for Linear Latent Growth Models

Abstract

In its most basic form, latent growth modeling (latent curve analysis) allows an assessment of individuals' change in a measured variable X over time. For simple linear models, as with other growth models, parameter estimates associated with the a construct (amount of X at a chosen temporal reference point) and b construct (growth in X per unit time) are not invariant with respect to choice of reference point and time unit. Latent means, variances, and covariances change with different temporal metrics. This article offers a nomenclature for describing linear latent growth models, demonstrates how latent moment parameter estimates vary as a function of changes in the a reference point and in the b growth metric, and presents a set of useful scale-free statistics for describing the results of linear latent growth modeling. Three examples are presented, two with a measured outcome and one with a latent outcome, and implications for applied and methodological researchers are presented.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 28 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.