On Latent Growth Models for Formative Factors

• In Event: Roundtable Session 101
  In Roundtable Session: Extensions and Issues in Growth Modeling

Abstract

Latent growth modeling has become a popular and versatile technique for evaluating longitudinal change. The common single measured outcome model has been extended many ways, most relevant here being multidomain and second order growth models. Whereas the former allows growth function characteristics for multiple outcomes to be modeled simultaneously, the latter models growth in latent outcomes, each having effect indicators repeated over time. But what if one has an outcome that is formative, rather than latent, relative to its indicators? In this case modeling change over time is less straightforward. The current paper provides analytical and applied details for simultaneously modeling growth in formative factors (composites) and their indicators, including a real data example using a General Computer Knowledge questionnaire.

Authors

• Gregory R. Hancock, University of Maryland

• Xiulin Mao, University of Maryland

• Hemant Kher, University of Delaware