Level, Change, and Acceleration: Modeling Correlated Change in Longitudinal Data and Intensive Repeated Measures Designs

Instructor(s):

  • Pascal Deboeck, University of Utah

This workshop will be offered in an online video format.

An ever-increasing number of models are available for the modeling of repeated observations on the same individuals, families, and groups. While theories often express ideas about correlated changes between constructs, matching theories about growth and change to statistical models can be challenging. This workshop will introduce derivatives -- level, change, and acceleration. This framework will provide a basis for articulating theories of change, as well as a framework for understanding a wide variety of change models. A variety of foundational models for longitudinal data will be discussed using this framework. These more common models will then be used to build models for intensive repeated measures designs (e.g., diary data, ecological momentary assessments). This course will provide a conceptual understanding of topics using lecture, reference code through demonstrations, and hands-on practice with sample data.

The course will cover:

  • Software and Conceptual Foundations
    • A brief introduction to key SEM and MLM concepts
    • A brief introduction to R, including packages for Structural Equation Modeling (SEM) and Multilevel Modeling (MLM)
    • A discussion about how change is conceptualized, with an introduction to the idea of derivatives (level, velocity, and acceleration)
  • Longitudinal/Panel Data
    • This section will introduce foundational longitudinal models and the kinds of change that can be tested
    • Growth Curve Models
    • Cross-Lagged Panel Models
    • Latent Difference Scores
  • Intensive Repeated Measures
    • This section will introduce models for intensive data characterized by non-linear changes over time
    • Methods for estimating derivatives (e.g., Generalize Local Linear Approximation, Generalized Orthogonal Local Derivative Estimates)
    • Modeling relations between derivative in regression and MLM
    • Latent Differential Equation Modeling

    Software Requirements: Lectures and exercises will rely on R, so that course tools and code are available to all participants. A concise introduction to R included as part of the course.

    Prerequisites: Prior experience with a graduate-level introductory statistics course (including regression models) and some exposure to longitudinal data is strongly recommended. It is not necessary to have prior exposure to R, SEM, or MLM, although participants with any exposure to one or more of these topics can expect to take more from the course. Introductions to core concepts in R, SEM, and MLM are included as part of the course, but not in sufficient depth to constitute a good background in these topics.

    Registration Fee: Members = $1,500; Non-members = $3,000

    Tags: Longitudinal Data Analysis, Multilevel Modeling, Structural Equation Modeling, Modeling Change, Derivatives, Intensive Longitudinal Designs, Diary Data, Ecological Momentary Assessments

    Course Sections

    Section 1

    Location: Online -- Video format,

    Date(s): June 15 - June 19

    Time: 10:00 AM - 6:00 PM

    Instructor(s):

    • Pascal Deboeck, University of Utah

    Syllabus: