Scholarlink

<p>Latent profile analysis (LPA) is in the finite mixture model analysis family and identifies subgroups by participants’ responses to continuous variables (i.e., indicators); participants’ probable membership in each subgroup is based on the similarity between the subgroup’s prototypical responses and the person’s unique responses. Compared to latent class analysis (LCA) with categorical data, LPA is a better fit for many variables and theories in behavioral medicine, because LPA can have continuous item, sub-scale, or scale scores as indicators, which can enable identifying and examining subgroups defined by responses representing complex, multidimensional concepts (e.g., orientations, motivations, well-being, ill-being, physical activity engagement) and biomarkers of diseases and disorders. Recently, the use of LPA has increased and as it continues to evolve, it is important researchers know best practice recommendations and explanations for both conducting as well as reading/reviewing LPA models. With this paper we: 1) discuss the strengths and weaknesses of LPA and the questions it is most appropriate to answer, 2) introduce LPA conceptually, 3) illustrate an LPA conducted with exercise psychology variables following current best practice recommendations, and 4) juxtapose resulting models from the LPA approach to a typical approach with the same data. We also share the data and syntax files used to conduct the basic steps of the LPA analyses as supplemental appendix files in addition to including the tables and figures for reporting LPA results following best practices.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Representing subpopulations with latent profile analysis: a non-technical introduction using exercisers’ goal orientation adoption profiles

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年9月9日

查看全文
  • E. Whitney G. Moore,
  • Alessandro Quartiroli

摘要

Latent profile analysis (LPA) is in the finite mixture model analysis family and identifies subgroups by participants’ responses to continuous variables (i.e., indicators); participants’ probable membership in each subgroup is based on the similarity between the subgroup’s prototypical responses and the person’s unique responses. Compared to latent class analysis (LCA) with categorical data, LPA is a better fit for many variables and theories in behavioral medicine, because LPA can have continuous item, sub-scale, or scale scores as indicators, which can enable identifying and examining subgroups defined by responses representing complex, multidimensional concepts (e.g., orientations, motivations, well-being, ill-being, physical activity engagement) and biomarkers of diseases and disorders. Recently, the use of LPA has increased and as it continues to evolve, it is important researchers know best practice recommendations and explanations for both conducting as well as reading/reviewing LPA models. With this paper we: 1) discuss the strengths and weaknesses of LPA and the questions it is most appropriate to answer, 2) introduce LPA conceptually, 3) illustrate an LPA conducted with exercise psychology variables following current best practice recommendations, and 4) juxtapose resulting models from the LPA approach to a typical approach with the same data. We also share the data and syntax files used to conduct the basic steps of the LPA analyses as supplemental appendix files in addition to including the tables and figures for reporting LPA results following best practices.