Statistics > Methodology
[Submitted on 22 Nov 2019 (v1), revised 26 May 2020 (this version, v5), latest version 3 Jan 2022 (v10)]
Title:Estimating knots in bilinear spline growth mixture models with time-invariant covariates in the framework of individual measurement occasions
View PDFAbstract:The linear spline growth mixture model (LSGMM), which incorporates the linear spline change patterns in the growth mixture model framework, is an appropriate tool for analyzing longitudinal data sets that come from a mixture of at least two latent classes when the underlying growth trajectories within each class are nonlinear. For each latent class, it approximates complex patterns by attaching at least two linear pieces. Besides examining within-person changes and between-person differences of trajectories simultaneously, it poses interesting statistical challenges, such as estimating the location of a change point (or knot), grouping individuals into at least two unobserved classes, examining factors that may be associated with those latent groups, and analyzing data with individually-varying measurement occasions. We developed a two-step bilinear spline growth mixture model (BLSGMM) to cluster these linear piecewise individual trajectories as well as to associate time-invariant covariates (TICs) to the latent classes. We present the proposed model by simulation studies and a real-world data analysis. Our simulation studies demonstrate that the proposed BLSGMM-TICs is capable of clustering the nonlinear change patterns well. More importantly, they generally estimate the parameters of interest unbiasedly, precisely, and exhibit appropriate confidence interval coverage. An empirical example using longitudinal math scores shows that the model helps identify latent groups of nonlinear trajectories with different knot locations and associate the classes with baseline characteristics. The corresponding computer code for the proposed model is provided.
Submission history
From: Jin Liu [view email][v1] Fri, 22 Nov 2019 09:17:47 UTC (182 KB)
[v2] Sat, 28 Mar 2020 22:54:51 UTC (550 KB)
[v3] Fri, 1 May 2020 02:44:23 UTC (550 KB)
[v4] Mon, 4 May 2020 22:20:27 UTC (546 KB)
[v5] Tue, 26 May 2020 13:32:35 UTC (562 KB)
[v6] Mon, 10 Aug 2020 01:31:41 UTC (540 KB)
[v7] Sun, 16 Aug 2020 19:07:45 UTC (541 KB)
[v8] Sat, 14 Aug 2021 20:23:35 UTC (241 KB)
[v9] Tue, 17 Aug 2021 01:18:16 UTC (541 KB)
[v10] Mon, 3 Jan 2022 04:47:26 UTC (541 KB)
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.